Limited Blood Transfusion Does Not Impact Survival in Octogenarians Undergoing Cardiac Operations

Limited Blood Transfusion Does Not Impact Survival in Octogenarians Undergoing Cardiac Operations

ADULT CARDIAC Limited Blood Transfusion Does Not Impact Survival in Octogenarians Undergoing Cardiac Operations James J. Yun, MD, PhD, Robert E. Helm...

467KB Sizes 2 Downloads 52 Views

ADULT CARDIAC

Limited Blood Transfusion Does Not Impact Survival in Octogenarians Undergoing Cardiac Operations James J. Yun, MD, PhD, Robert E. Helm, MD, Robert S. Kramer, MD, Bruce J. Leavitt, MD, Stephen D. Surgenor, MD, MS, Anthony W. DiScipio, MD, Lawrence J. Dacey, MS, MD, Yvon R. Baribeau, MD, Louis Russo, MD, Gerald L. Sardella, MD, David C. Charlesworth, MD, Robert A. Clough, MD, Joseph P. DeSimone, MD, Cathy S. Ross, MS, David J. Malenka, MD, and Donald S. Likosky, PhD, for the Northern New England Cardiovascular Disease Study Group Departments of Medicine, Surgery, and Community and Family Medicine, and The Dartmouth Institute for Health Policy and Clinical Practice, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire; Department of Surgery, Portsmouth Regional Hospital, Portsmouth, New Hampshire; Division of Cardiothoracic Surgery, Maine Medical Center, Portland, Maine; Department of Surgery, Fletcher Allen Health Care, Burlington, Vermont; Department of Surgery, Catholic Medical Center, Manchester, New Hampshire; Department of Surgery, Central Maine Medical Center, Lewiston, Maine; Department of Surgery, Concord Hospital, Concord, New Hampshire; Department of Surgery, Eastern Maine Medical Center, Bangor, Maine

Background. We previously reported that transfusion of 1 to 2 units of red blood cells (RBCs) confers a 16% increased hazard of late death after cardiac surgical treatment. We explored whether a similar effect existed among octogenarians. Methods. We enrolled 17,026 consecutive adult patients undergoing cardiac operations from 2001 to 2008 in northern New England. Patients receiving more than 2 units of RBCs or undergoing emergency operations were excluded. Early (to 6 months) and late (to 3 years, among those surviving longer than 6 months) survival was confirmed using the Social Security Death Index. We estimated the relationship between RBCs and survival, and any interaction by age (<80 years versus > 80 years) or procedure. We calculated the adjusted hazard ratio (HR), and plotted adjusted survival curves. Results. Patients receiving RBCs had more comorbidities irrespective of age. Patients 80 years of age or older

underwent transfusion more often than patients younger than 80 years (51% versus 30%; p < 0.001). There was no evidence of an interaction by age or procedure (p > 0.05). Among patients younger than 80 years, RBCs significantly increased a patient’s risk of early death [HR, 2.03; 95% confidence interval [CI], 1.47, 2.80] but not late death 1.21 (95%CI, 0.88, 1.67). RBCs did not increase the risk of early [HR, 1.47; 95% CI, 0.84, 2.56] or late (HR, 0.92 95% CI, 0.50, 1.69) death in patients 80 years or older. Conclusions. Octogenarians receive RBCs more often than do younger patients. Although transfusion of 1 to 2 units of RBCs increases the risk of early death in patients younger than 80 years, this effect was not present among octogenarians. There was no significant effect of RBCs in late death in either age group.

T

bypass grafting (CABG) [7]. Similar findings regarding the impact of RBC transfusion on late survival have been reported [1, 2, 7–9]. Octogenarians represent an increasing proportion of patients undergoing cardiac operations. Because of preoperative anemia, octogenarians receive more RBC transfusions during or after cardiac operations [10,11]. Although favorable survival rates in selected octogenarians have been reported, octogenarians may incur higher levels of resource use [10, 12–15]. Furthermore, the effect of RBC transfusions on mortality in octogenarians is undefined. We undertook a regional prospective study of 17,026 patients undergoing coronary or valve surgical procedures in northern New England to describe the relationship between transfusion of 1 to 2 units of RBCs during

ransfusion of red blood cells (RBCs) during admission for cardiac operations increases postoperative morbidity, early mortality, and late mortality [1–3]. In the perioperative period, RBC transfusion may be associated with a number of adverse sequelae, including immunosuppression, infection, and pulmonary dysfunction [4 – 6], each of which adversely affects early postoperative survival [1, 6]. We previously reported that transfusion of 1 to 2 units of RBCs was associated with a 16% increase in adjusted 5-year mortality after isolated coronary artery

Accepted for publication June 20, 2012. Presented at the Poster Session of the Forty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –Feb 1, 2012. Address correspondence to Dr Likosky, Department of Cardiac Surgery, University of Michigan Medical School, Ann Arbor, MI 48109; e-mail: [email protected]

© 2012 by The Society of Thoracic Surgeons Published by Elsevier Inc

(Ann Thorac Surg 2012;94:2038 – 45) © 2012 by The Society of Thoracic Surgeons

0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2012.06.059

Ann Thorac Surg 2012;94:2038 – 45

Patients and Methods The Northern New England Cardiovascular Disease Study Group (NNECDSG) is a voluntary research consortium, composed of clinicians, research scientists, and hospital administrators, representing all medical centers in Maine, Vermont, and New Hampshire where cardiac operations are performed. Since 1987, the NNECDSG has maintained a prospective registry of all patients undergoing cardiac operations in the region. The group fosters continuous regional improvement in the quality of care of patients with cardiovascular disease through analysis of outcome data and timely feedback to clinicians [16]. We included all patients undergoing cardiac surgical procedures at all 8 NNECDSG regional medical centers between 2001 and 2008. We excluded patients undergoing emergent operations and those receiving 3 RBC transfusions or more. Our final cohort consisted of 15,715 patients younger than 80 years (11,053 did not receive transfusions and 4,662 did) and 1,311 patients 80 years or older (638 did not receive transfusions and 673 did).

Data Collection Previous publications by the NNECDSG have outlined our data collection methods and definitions [17]. We prospectively collected the preoperative variables displayed in Table 1. Comorbidities analyzed included diabetes, cerebrovascular disease (previous stroke, previous transient ischemic attack, previous carotid artery operations, carotid artery stenosis, or carotid artery bruit), peripheral vascular disease (PVD), treated chronic obstructive pulmonary disease (COPD), renal failure (dialysis or creatinine level ⱖ 2 mg/dL), cardiac history (previous myocardial infarction, congestive heart failure [CHF] during admission or during a previous admission), and cardiac anatomy and function (left main coronary artery stenosis, number of diseased coronary arteries, and preoperative left ventricular ejection fraction [EF]). Cardiothoracic surgeons defined the patient acuity (elective, urgent, emergent) using definitions previously described [17].

2039

Statistical Analysis We used standard statistical methods to compare the characteristics of patients by both age strata and exposure to RBC transfusions, including analysis of variance for continuous variables, the ␹2 test for categorical data, and logistic regression for ordered categories (reported as a p trend). We used the Wilcoxon rank sum test for nonparametric data. The risk of death at predefined times (early ⫽ within 6 months, and late ⫽ up to 3 years) associated with age and RBC transfusion was compared using the Wald test within a Cox proportional hazards model. We report 3-year survival conditional on survival beyond 6 months. Survival was adjusted for covariates previously found to be associated with mortality, including age, sex, EF, case acuity, PVD, diabetes, previous myocardial infarction, renal failure or creatinine levels 2 mg/dL or greater, left main coronary artery disease, COPD, body mass area (BSA), year, preoperative hematocrit value, and procedure type. Time dependence of the HR was assessed by estimating the HR in separate windows of time after follow-up. We tested for interactions between patient age and RBC transfusion in each of the periods. Crude survival curves were estimated using the Kaplan-Meier method. Survival adjusted for differences between patients across age strata and by exposure to RBC transfusions were estimated using the method of Zhang [18], which does not force proportional hazards on the difference between age strata and RBC transfusions. Separate curves were created for each age stratum and RBC transfusion category. Adjusted survival curves were created using Stata software, version 12.0 (StataCorp, College Station, TX).

Protection of Human Subjects Institutional review board approval was obtained at each medical center. Seven of our 8 member centers’ institutional review boards have designated the NNECDGS as a Quality Improvement Registry and therefore patient consent was not required. Written consent was obtained for all patients at the 1 remaining center. The authors had full access to the data and take responsibility for its integrity. All authors have read and agreed to the manuscript as written.

Results

Subject Stratification

Study Population

Patients were stratified and analyzed according to their age (⬍80 years versus ⱖ80 years) as well as their exposure (within age category) to RBC transfusions (none, or 1–2 RBC units).

Between 2001 and 2008, a total of 17,026 (15,715 ⬍ 80 years and 1,311 ⱖ 80 years) patients were analyzed. These patients contributed a total of 61,460.1 patient-years of follow-up data (57,598.2 ⬍ 80 years and 3,861.9 ⱖ 80 years). A total of 5,335 patients received 1 to 2 units of RBCs (4,662 ⬍ 80 years and 673 ⱖ 80 years).

Outcome Data We reported both early (6 months) and late (3 years) mortality subsequent to the index admission by linking our registry to the Social Security Administration’s Death Master File using a combination of first name, last name, date of birth, date last known alive, and Social Security number.

Baseline Characteristics DIFFERENCES AMONG PATIENTS YOUNGER THAN 80 YEARS. Patients

aged 80 years or younger receiving 1 to 2 units of RBCs were more often women and more often had a preoperative length of stay 3 days or longer, a BSA less than 2.0,

ADULT CARDIAC

admission and the subsequent risk of early and late mortality.

YUN ET AL TRANSFUSION IN OCTOGENARIANS

2040

YUN ET AL TRANSFUSION IN OCTOGENARIANS

Ann Thorac Surg 2012;94:2038 – 45

Table 1. Preoperative Characteristics Among Patients Undergoing Cardiac Operations by Patient Age and Red Blood Cell Use Age ⬍80 Years Variable

No Transfusion 1⫺2 RBC Units p Value No Transfusion 1–2 RBC Units p Value

ADULT CARDIAC

Number of procedures (%) Preoperative length of stay (d) 0 1 2 ⱖ3 Demographics Female sex (%) Body surface area (m2) ⬍1.70 1.70–1.99 ⱖ2.00 Comorbid disease Diabetes, % yes Vascular disease, % yes COPD, % yes Renal failure or creatinine level ⱖ 2 mg/dL, % yes Hematocrit (%) ⬍36 36–39 40–42 ⱖ43 White blood cell count 12,000/mm3, % yes Cardiac history Congestive heart failure, % yes NY Heart Association III–IV (%) Previous myocardial infarction (%) No ⱕ7 d before procedure ⬎7days before procedure Previous coronary or valve procedure, % yes None CABG PCI Valve Mixture Urgent, % yes CABG ⫽ coronary artery bypass grafting; cells.

Age ⱖ80 Years

11,053 (70)

4,662 (30)

638 (49)

673 (51)

39.6 16.7 11.0 32.7

34.6 16.5 11.3 37.6 np trend

39.0 15.3 10.2 35.5 ⬍0.001

28.7 16.1 9.5 45.7 np trend

⬍0.001

15.5

41.4

⬍0.001

22.3

44.3

⬍0.001

4.1 31.9 64.1

14.4 45.5 40.1 np trend

13.0 48.0 39.0 ⬍0.001

25.7 54.4 19.9 np trend

⬍0.001

29.6 18.8 11.1 1.5

37.0 26.9 14.7 3.4

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

20.7 27.6 9.7 1.9

25.0 30.3 12.8 3.9

0.07 0.28 0.08 0.05

12.0 26.5 29.5 32.1

19.6 34.1 26.2 20.1

5.1

33.8 32.9 20.1 13.2 np trend 5.9

⬍0.001 0.03

3.3

42.2 33.0 16.7 8.1 np trend 5.5

⬍0.001 0.05

13.0 53.1

20.1 59.1

⬍0.001 ⬍0.001

22.8 59.8

31.4 66.4

⬍0.001 0.03

65.8 15.4 18.9

62.1 16.4 21.6 np trend

69.1 13.8 17.1

63.2 16.6 20.2 np trend

0.04

80.7 10.3 3.7 0.7 4.6 67.8

0.72 ⬍0.001

75.3 15.4 1.8 0.9 6.7 58.4

72.0 14.6 2.5 1.4 9.6 62.6

COPD ⫽ chronic obstructive pulmonary disease;

diabetes, PVD, COPD, renal failure or creatinine levels 2 mg/dL or greater (or both), a preoperative hematocrit value less than 36%, CHF, elevated white blood cell count, New York Heart Association class III-IV, previous myocardial infarction of 7 days or longer, previous coronary artery or valve procedures, and urgent procedure status (Table 1).

⬍0.001

⬍0.001 ⬍0.001

82.6 9.9 2.5 0.9 4.1 56.9

PCI ⫽ percutaneous coronary intervention;

RBC ⫽ red blood

DIFFERENCES AMONG PATIENTS 80 YEARS OR OLDER. Patients aged

80 years or older receiving 1 to 2 units of RBCs were more often women and more often had a preoperative length of stay of 3 days or more, a BSA less than 2.0, renal failure or creatinine levels of 2 mg/dL or more, a preoperative hematocrit value less than 36%, elevated white blood cell count, CHF, New York Heart Association class III-IV,

Ann Thorac Surg 2012;94:2038 – 45

YUN ET AL TRANSFUSION IN OCTOGENARIANS

2041

Table 2. Intraoperative Characteristics Among Patients Undergoing Cardiac Operations by Patient Age and Red Blood Cell Use

Variable Number of procedures (%) Type of procedure Isolated CABG Isolated valve CABG ⫹ valve

No RBC Transfusion 11,053 (70) 76.2 15.6 8.2

1–2 Units

Age ⱖ80 Years p Value

4,662 (30)

No RBC Transfusion 638 (49)

70.7 14.8 14.6

55.0 21.5 23.5

1–2 Units 673 (51) 55.4 21.3 23.3

⬍0.001 Cardiopulmonary bypass duration (min) Mean (SD) Median (IQR) Cross-clamp duration (min) Mean (SD) Median (IQR) Intraaortic balloon pump (intraoperative/ postoperative) (% yes) Intraoperative nadir hematocrit (%) ⬍21 21–23 24–25 ⱖ26 Lowest Core Temperature (°C) ⬍33 33–35 ⱕ35 Among patients having CABG or CABG/Valve off-pump procedure Internal mammary artery (% yes) Cardiac anatomy and function Left ventricular ejection fraction (%) ⬍40 40–49 50–59 ⱖ60 Left main coronary artery stenosis (%) ⬍50 50–89 ⱖ90

p Value

0.99

104.6 (40.0) 99.0 (79.0, 124.0)

113.8 (45.3) 106.0 (84.0, 134.0)

⬍0.001 ⬍0.001

112.1 (47.4) 104.0 (83.0, 135.0)

114.1 (38.4) 109.0 (87.0, 136.0)

0.42 0.06

84.1 (285.2) 64.0 (47.0, 85.0) 1.0

86.8 (213.8) 68.0 (50.0, 92.0) 2.4

0.59 ⬍0.001 ⬍0.001

77.6 (37.1) 71.0 (54.0, 95.5) 2.4

84.8 (183.8) 73.0 (56.0, 97.0) 2.5

0.36 0.51 0.94

9.1 24.6 22.0 44.2

33.8 34.9 16.9 14.5 np trend

16.7 34.3 22.2 26.9

39.2 38.0 13.8 9.1 np trend

⬍0.001

0.12 0.71

23.6 54.5 21.9

⬍0.001

9.5

34.8 49.0 16.2 np trend 6.8

⬍0.001 ⬍0.001

10.8

32.2 49.5 18.4 np trend 11.5

94.6

91.9

⬍0.001

89.8

88.0

0.40

12.4 15.0 26.3 46.3

15.9 14.6 24.5 45.0 np trend

16.5 17.8 20.2 45.5

18.3 14.2 25.9 41.6 np trend

0.59

62.1 28.5 9.4 np trend

0.13

71.0 25.2 3.8

70.6 26.0 3.4 np trend

31.3 44.5 24.3

⬍0.001

64.7 29.3 6.0 0.97

(Continued)

ADULT CARDIAC

Age ⬍80 Years

2042

YUN ET AL TRANSFUSION IN OCTOGENARIANS

Ann Thorac Surg 2012;94:2038 – 45

Table 2. Continued Age ⬍ 80 Years

ADULT CARDIAC

Variable Diseased vessels (No.) 1 2 3 Number of distal anastomoses Mean (SD) Median (IQR) CABG ⫽ coronary artery bypass grafting;

No RBC Transfusion 20.8 36.2 43.0

3.3 (1.2) 3.0 (3.0, 4.0)

Age ⱖ 80 Years

1–2 Units

p Value

18.1 36.3 45.6 np trend

⬍0.001

3.3 (1.1) 3.0 (3.0, 4.0)

0.28 0.15

IQR ⫽ interquartile range;

No RBC Transfusion 27.4 28.1 44.5

2.9 (1.2) 3.0 (2.0, 4.0)

RBC ⫽ red blood cell;

1–2 Units

p Value

23.6 30.0 46.4 np trend

0.27

3.1 (1.2) 3.0 (2.0, 4.0)

0.04 0.08

SD ⫽ standard deviation.

previous myocardial infarction of 7 days or longer, and urgent procedure status.

was not significant between the 6-month and 3-year periods (adjusted HR, 1.21; 95% CI, 0.88, 1.67).

Intraoperative Care

PATIENTS 80 YEARS OR OLDER.

DIFFERENCES AMONG PATIENTS YOUNGER THAN 80 YEARS. Patients

younger than 80 years receiving 1 to 2 units of RBCs more often underwent CABG in conjunction with a valve operation and had longer cardiopulmonary bypass times and longer median cross-clamp times, received an intraaortic balloon pump, and had a lower nadir hematocrit value on bypass and lower core temperatures (Table 2). Within the cohort of patients undergoing CABG (with or without a concomitant valve operation), patients receiving 1 to 2 units of RBCs were less likely to receive an off-pump procedure and an internal mammary artery graft but had a lower EF and greater number of diseased vessels. DIFFERENCES AMONG PATIENTS 80 YEARS OR OLDER. Patients 80 years or older receiving 1 to 2 units of RBCs more often had a lower nadir hematocrit value on bypass. Within the cohort of patients undergoing CABG (with or with a concomitant valve operation), patients receiving 1 to 2 units of RBCs were more likely to have a higher average number of distal anastomoses.

Short-Term and Long-Term Survival There was no interaction by age or procedure and RBC transfusion in either period (p ⬎ 0.05). Crude and adjusted HRs by period are displayed in Table 3. Corresponding survival curves are shown in Fig 1 (early phase, from operation to 6 months) and Fig 2 (late phase, 6 months to 3 years). Among patients younger than 80 years, the hazard of death within a 3-year period increased to 1.29 (95% CI, 1.07, 1.55) after adjusting for age, sex, EF, surgical case acuity, PVD, diabetes, previous myocardial infarction, renal failure or creatinine level 2 mg/dL or greater, left main coronary artery disease, COPD, BSA, year, preoperative hematocrit value, and type of procedure (Table 3; Figs 1, 2). A similar trend in hazard was found during the 0- to 6-month period (adjusted HR, 2.03; 95 CI%, 1.47, 2.80), although the effect

PATIENTS YOUNGER THAN 80 YEARS.

Among patients 80 years or older, the hazard of death within a 3-year period decreased to 0.76 (95% CI, 0.56, 1.04) after adjustment (Table 3; Figs 1, 2). No significant effect of transfusion on patient survival was seen across periods: 0- to 6-month period (adjusted HR, 1.47; 95% CI, 0.84, 2.56); 6-month to 3-year period (adjusted HR, 0.92; 95% CI, 0.50, 1.69).

Comment Our study has 3 main findings. First, a limited (1- or 2-unit) RBC transfusion during admission for cardiac surgical procedures in our regional cohort was nearly twice as prevalent in octogenarians (51%) as in younger Table 3. Crude and Adjusted Hazard Ratio for Death for the Use of Red Blood Cell Transfusion by Age Group and Time

Variable Operation to 6 months % dead Crude Adjusted 6 months to 3 years % dead Crude Adjusted Overall % dead Crude Adjusted

Age ⱖ80 Age ⬍80 Years Years HR HR 95% CI 95% CI 2.06 2.64 2.03

2.12, 3.28 1.47, 2.80

7.63 1.51 1.47

1.00, 2.26 0.84, 2.56

4.97 1.96 1.21

1.66, 2.31 0.88, 1.67

10.98 1.31 0.92

0.89, 1.92 0.50, 1.69

9.68 1.85 1.29

1.67, 2.05 1.07, 1.55

24.71 1.10 0.76

0.88, 1.37 0.56, 1.04

Adjusted for sex, ejection fraction, number of diseased vessels, left main coronary artery disease, white blood cell count, preoperative myocardial infarction, priority, vascular disease, diabetes, renal failure/creatinine level ⬎ 2 mg/dL, chronic obstructive pulmonary disease, body mass index, preoperative hematocrit value, type of procedure, year. CI ⫽ confidence interval;

HR ⫽ hazard ratio.

Ann Thorac Surg 2012;94:2038 – 45

YUN ET AL TRANSFUSION IN OCTOGENARIANS

2043

ADULT CARDIAC

Fig 1. Adjusted survival by red blood cell use: early phase (operation to 6 months). (A) Patients younger than 80 years. (B) Patients 80 years or older. (RBC ⫽ red blood cells.)

patients (30%). Second, exposure to 1 to 2 units of RBCs during admission for cardiac operations increased early mortality (to 6 months) for patients younger than 80 years, even after adjusting for preoperative characteristics. Third, among octogenarians, RBC transfusions did not increase the risk of early or late death. The most surprising finding of our study was the lack of a significant increase in early mortality risk associated with a 1- to 2-unit RBC transfusion in octogenarians. Although this was not expected based on our group’s previous analyses [7], our results confirm those of Carrascal and colleagues [11], who reported that in 246 octogenarians undergoing cardiac operations, transfusion of 1 to 2 units of RBCs did not significantly decrease early (in-hospital) or late (90 months) survival. We postulate 2 mechanisms for this observation. First, agerelated comorbidities such as anemia [19] contribute to mortality in all octogenarians, and age-related causes of death could simply mask the truly adverse impact of RBC transfusions on survival. An example of an age-related morbidity is immune senescence, in which elderly patients suffer multiple immune deficits, including a reducFig 2. Adjusted survival by red blood cell use: late phase (6 months to 3 years). (A) Patients younger than 80 years. (B) Patients 80 years or older. (RBC ⫽ red blood cells.)

tion in absolute T-cell number, depletion of naive T-cell populations, and diminished CD28 expression [20 –22]. Immune senescence, although not quantified in this study, could have predisposed octogenarians to morbidities such as infections and cancer, potentially increasing postoperative mortality. Another age-related alteration in adaptive immunity is the upregulation of certain inflammatory responses from T cells. Sansoni and coworkers [23] demonstrated that in very elderly (⬎ 85 years) individuals, CD8⫹ T cells produce higher levels of the proinflammatory cytokines interleukin 2, tumor necrosis factor-alpha, and interferon-gamma in response to inflammatory stimuli [24]. This age-related phenomenon could result in greater perioperative systemic inflammation after cardiopulmonary bypass in the elderly and in turn impact outcomes after cardiac operations. Finally, it is possible that in the context of severe anemia, RBC transfusion might actually improve survival after cardiac operations. In a retrospective analysis of 78,974 elderly Medicare patients, Wu and associates [25] reported that transfusion in elderly patients with low

2044

YUN ET AL TRANSFUSION IN OCTOGENARIANS

ADULT CARDIAC

hematocrit values (⬍ 24%) who were experiencing acute myocardial infarction was associated with improved early (30-day) mortality [25]. The authors suggested that RBC transfusion, by correcting severe anemia, reduced the mortality of patients with myocardial infarction and hematocrit values less than 24% to those who had hematocrit values of more than 30%. Nevertheless, given the overall weight of evidence supporting a survival disadvantage associated with RBC transfusions [2–9, 26], we do not endorse transfusing 1 to 2 units of RBCs in octogenarians; to the contrary, our study shows that RBC transfusion carries an increased risk of early mortality (albeit nonsignificant) in octogenarians. As such, transfusing even 1 to 2 units of RBCs may lead to some degree of increased mortality risk in this age group. Our study results represent our regional experience in caring for elderly patients undergoing cardiac operations, and further studies including longer follow-up and greater sample size are clearly warranted to confirm our findings. Several limitations are worth noting. First, although our study reflects our regional practice during this period, we cannot discount lingering effects of confounding. Nonetheless, we used methods of risk adjustment to account for baseline differences in case mix, differences across centers, and temporal differences in practice. Second, our NNECDSG database does not include specific information on the frequency of RBC leukoreduction or the age of the individual RBCs transfused, although during the period studied, most centers in this analysis had transitioned to the use of leukoreduced RBCs. Third, although we had sufficient statistical power to study the effect of RBC transfusion on survival among patients younger than 80 years, we enrolled 1,311 patients 80 years or older (we estimate the need to enroll 1,300 patients ⱖ 80 years to detect a difference in survival at 3 years from 85% versus 91%). As such, we acknowledge there may be some lingering risk of a type II error in our findings among octogenarians. Nonetheless, we found the use of RBCs among those 80 years or older to be associated with a higher hazard of death (albeit nonsignificant) in the early postoperative period. In summary, we found that octogenarians undergoing nonemergent cardiac operations frequently undergo transfusion with 1 to 2 units of RBCs. In patients younger than 80 years, a 1- to 2-unit RBC transfusion was associated with increased early mortality. We were surprised to find that among octogenarians, a 1- to 2-unit RBC transfusion was not associated with a significant increased risk of early or late mortality. Nonetheless, 1 to 2 units of RBCs conferred a 47% increased risk of early mortality and an 8% decreased risk of late mortality among octogenarians. We should continue to apply thoughtful restrictive transfusion practices in older patients rather than considering this analysis as license for liberal transfusion practices. These data support the need for continued blood conservation and investigations directed at further clarifying the sustained effects of RBC transfusion on survival among octogenarians.

Ann Thorac Surg 2012;94:2038 – 45

Dr Likosky was supported by a grant from the Agency for Healthcare Research and Quality (1K02HS015663-01A1). This work was partially funded by the Northern New England Cardiovascular Disease Study Group.

References 1. Michalopoulos A, Tzelepis G, Dafni U, Geroulanos S. Determinants of hospital mortality after coronary artery bypass grafting. Chest 1999;115:1598 – 603. 2. Koch CG, Li L, Duncan AI, et al. Transfusion in coronary artery bypass grafting is associated with reduced long-term survival. Ann Thorac Surg 2006;81:1650 –7. 3. Engoren MC, Habib RH, Zacharias A, Schwann TA, Riordan CJ, Durham SJ. Effect of blood transfusion on long-term survival after cardiac operation. Ann Thorac Surg 2002;74: 1180 – 6. 4. Leal-Noval SR, Marquez-Vacaro JA, Garcia-Curiel A, et al. Nosocomial pneumonia in patients undergoing heart surgery. Crit Care Med 2000;28:935– 40. 5. Zacharias A, Habib RH. Factors predisposing to median sternotomy complications. Deep vs superficial infection. Chest 1996;110:1173– 8. 6. Rogers MA, Blumberg N, Saint SK, Kim C, Nallamothu BK, Langa KM. Allogeneic blood transfusions explain increased mortality in women after coronary artery bypass graft surgery. Am Heart J 2006;152:1028 –34. 7. Surgenor SD, Kramer RS, Olmstead EM, et al. The association of perioperative red blood cell transfusions and decreased long-term survival after cardiac surgery. Anesth Analg 2009;108:1741– 6. 8. Jakobsen CJ, Ryhammer PK, Tang M, Andreasen JJ, Mortensen PE. Transfusion of blood during cardiac surgery is associated with higher long-term mortality in low-risk patients. Eur J Cardiothorac Surg 2012;42:114 –20. Epub 2012 Jan 12. 9. van Straten AH, Bekker MW, Soliman Hamad MA, et al. Transfusion of red blood cells: the impact on short-term and long-term survival after coronary artery bypass grafting, a ten-year follow-up. Interact Cardiovasc Thorac Surg 2010;10: 37– 42. 10. Scott BH, Seifert FC, Grimson R, Glass PS. Octogenarians undergoing coronary artery bypass graft surgery: resource utilization, postoperative mortality, and morbidity. J Cardiothorac Vasc Anesth 2005;19:583– 8. 11. Carrascal Y, Maroto L, Rey J, et al. Impact of preoperative anemia on cardiac surgery in octogenarians. Interact Cardiovasc Thorac Surg 2010;10:249 –55. 12. Speziale G, Nasso G, Barattoni MC, et al. Short-term and long-term results of cardiac surgery in elderly and very elderly patients. J Thorac Cardiovasc Surg 2011;141:725–31, 731. e1. Epub 2010 Jun 19. 13. Saxena A, Poh CL, Dinh DT, Smith JA, Shardey GC, Newcomb AE. Early and late outcomes after isolated aortic valve replacement in octogenarians: an Australasian Society of Cardiac and Thoracic Surgeons Cardiac Surgery Database study. Eur J Cardiothorac Surg 2012;41:63– 8. 14. Likosky DS, Dacey LJ, Baribeau YR, et al. Long-term survival of the very elderly undergoing coronary artery bypass grafting. Ann Thorac Surg 2008;85:1233–7. 15. Likosky DS, Sorensen MJ, Dacey LJ, et al. Long-term survival of the very elderly undergoing aortic valve surgery. Circulation 2009;120:S127–33. 16. O’Connor GT, Plume SK, Olmstead EM, et al. A regional prospective study of in-hospital mortality associated with coronary artery bypass grafting. JAMA 1991;266:803–9. 17. O’Connor GT, Plume SK, Olmstead EM, et al. Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Circulation 1992;85:2110 – 8. 18. Zhang X, Loberiza FR, Klein JP, Zhang MJ. A SAS macro for estimation of direct adjusted survival curves based on a

19. 20. 21. 22.

stratified Cox regression model. Comput Methods Programs Biomed 2007;88:95–101. Patel KV, Guralnik JM. Prognostic implications of anemia in older adults. Haematologica 2009;94:1–2. Mariani E, Facchini A. Characterization of NK cells in the elderly. In: G Pawelec, ed. Advances in Cell Aging and Gerontology. Philadelphia, PA: Elsevier; 2003:133–57. McLeod J. T cell aging and immune surveillance. In: Advances in cell aging and gerontology, Volume 13, G Pawelec, ed. Philadelphia, PA: Elsevier; 2003:159 –72. Effros R. Loss of CD28 expression on T lymphocytes: a marker of replicative senescence. Devel Comp Immunol 1997;12:471– 8.

YUN ET AL TRANSFUSION IN OCTOGENARIANS

2045

23. Sansoni P, Vescovini R, Fagnoni F, Biasini C, et al. The immune system in extreme longevity. Exp Gerontol 2008;43:61–5. 24. Zanni F, Vescovini R, Biasini C, et al. Marked increase with age of type 1 cytokines within memory and effector/cytotoxic CD8⫹ T cells in humans: a contribution to understand the relationship between inflammation and immunosenescence. Exp Gerontol 2003;38:981–7. 25. Wu WC, Rathore SS, Wang Y, Radford MJ, Krumholz HM. Blood transfusion in elderly patients with acute myocardial infarction. N Engl J Med 2001;345:1230 – 6. 26. Alfirevic A, Xu M, Johnston D, Figueroa P, Koch CG. Transfusion increases the risk for vasoplegia after cardiac operations. Ann Thorac Surg 2011;92:812–9.

INVITED COMMENTARY In this issue of The Annals of Thoracic Surgery, the article by Yun and colleagues [1] examined a large group of patients who underwent cardiac surgery from the Northern New England Cardiovascular Disease Study Group’s prospective registry. Patients represented nonemergent cases who received 2 units or less of red blood cells (RBCs). Risk-adjusted early death was higher for transfused patients younger than 80 years but not for patients 80 years or older. Older patients received blood more frequently, almost twice as often, yet RBC transfusion did not increase their risk for death. Taken at face value, these findings are surprising, even counterintuitive. If one were to choose a high-risk group most susceptible to adverse complications associated with RBC transfusion, or any intervention, it would likely be an elderly population. The authors consider that age-related morbidity might mask the signal, if present, for RBC transfusion-related morbidity and reduced survival. Of note, the group of patients 80 years or older had a higher proportion of patients with preoperative hematocrit values less than 36% and a higher proportion with a nadir hematocrits less than 21% and 21% to 23%. There was no nadir hematocrit that triggered transfusion. For the group of patients 80 years or older, could RBC transfusion fix the problem of chronic anemia and its negative consequences? Evidence suggests that patients with exposure to RBCs and anemia both have morbidity risk.

© 2012 by The Society of Thoracic Surgeons Published by Elsevier Inc

That these findings are surprising is also evidence that we do not understand the optimal circumstances for RBC transfusion, nor do we understand the mechanisms by which transfusion increases mortality. Issues with RBC transfusion and anemia constitute work-in-progress as the field moves toward more patient-centered anemia and blood management. There is no broad acceptance of evidence-based guidelines to address variability in clinical practice patterns. This article provides interesting perspective as the critically ill elderly patient is becoming a more common demographic in surgical theaters and RBC continues to be a limited resource. Colleen G. Koch, MD Department of Cardiothoracic Anesthesia Cleveland Clinic 9500 Euclid Ave Cleveland, OH 44195 e-mail: [email protected]

Reference 1. Yun JJ, Helm RE, Kramer RS, et al. Limited blood transfusion does not impact survival in octogenarians undergoing cardiac operations. Ann Thorac Surg 2012;94:2038 – 45.

0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2012.07.023

ADULT CARDIAC

Ann Thorac Surg 2012;94:2038 – 45