Persistent Effect of Red Cell Transfusion on Health-Related Quality of Life After Cardiac Surgery

Article (PDF Available)inThe Annals of thoracic surgery 82(1):13-20 · August 2006with68 Reads
DOI: 10.1016/j.athoracsur.2005.07.075 · Source: PubMed
Abstract
Although red blood cell transfusion has been associated with an increase in early morbid outcomes and reduced long-term survival after cardiac surgery, its relationship to functional quality of life after surgery has not been previously explored. Our objective was to investigate the relationship between perioperative red blood cell and component transfusion and functional health-related quality of life 6 to 12 months after cardiac surgery. Of 12,536 patients undergoing cardiac surgical procedures between May 1995 and January 1999, 7,321 completed a self-administered Duke Activity Status Index (DASI) survey preoperatively and least one follow-up survey at nominally 6 or 12 months postoperatively. The influence of baseline DASI, preoperative risk factors, clinical status, laboratory values, operative events, and postoperative morbidities on follow-up DASI were examined with ordinal regression modeling. After adjustment for preoperative DASI, demographic, cardiac and noncardiac comorbidity, type of surgery, postoperative complications, and interval between follow-up DASI, during which patients continued to improve (p < 0.0001), postoperative functional status after cardiac surgery was incrementally worse the more perioperative red cells (p < 0.0001) and platelets (p = 0.02) that had been transfused. Red blood cell and platelet transfusion have an unintended persistently negative risk-adjusted effect on health-related quality of life after cardiac surgery that extends well beyond initial hospitalization. Reductions in functional recovery paralleled increasing units of red blood cells transfused.
DOI: 10.1016/j.athoracsur.2005.07.075
2006;82:13-20 Ann Thorac Surg
Loop and Eugene H. Blackstone
Colleen Gorman Koch, Farah Khandwala, Liang Li, Fawzy G. Estafanous, Floyd D.
Cardiac Surgery
Persistent Effect of Red Cell Transfusion on Health-Related Quality of Life After
http://ats.ctsnetjournals.org/cgi/content/full/82/1/13
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
Print ISSN: 0003-4975; eISSN: 1552-6259.
Southern Thoracic Surgical Association. Copyright © 2006 by The Society of Thoracic Surgeons.
is the official journal of The Society of Thoracic Surgeons and theThe Annals of Thoracic Surgery
by on May 31, 2013 ats.ctsnetjournals.orgDownloaded from
Persistent Effect of Red Cell Transfusion on Health-
Related Quality of Life After Cardiac Surgery
Colleen Gorman Koch, MD, MS, Farah Khandwala, MS, Liang Li, PhD,
Fawzy G. Estafanous, MD, Floyd D. Loop, MD, and Eugene H. Blackstone, MD
Departments of Cardiothoracic Anesthesia, Quantitative Health Sciences, and Thoracic and Cardiovascular Surgery, and Division
of Anesthesia and Critical Care, The Cleveland Clinic Foundation, Cleveland, Ohio
Background. Although red blood cell transfusion has
been associated with an increase in early morbid outcomes
and reduced long-term survival after cardiac surgery, its
relationship to functional quality of life after surgery has
not been previously explored. Our objective was to inves-
tigate the relationship between perioperative red blood cell
and component transfusion and functional health-related
quality of life 6 to 12 months after cardiac surgery.
Methods. Of 12,536 patients undergoing cardiac surgi-
cal procedures between May 1995 and January 1999, 7,321
completed a self-administered Duke Activity Status In-
dex (DASI) survey preoperatively and least one fol-
low-up survey at nominally 6 or 12 months postopera-
tively. The influence of baseline DASI, preoperative risk
factors, clinical status, laboratory values, operative
events, and postoperative morbidities on follow-up
DASI were examined with ordinal regression modeling.
Results. After adjustment for preoperative DASI, de-
mographic, cardiac and noncardiac comorbidity, type of
surgery, postoperative complications, and interval be-
tween follow-up DASI, during which patients continued
to improve (p < 0.0001), postoperative functional status
after cardiac surgery was incrementally worse the more
perioperative red cells (p < 0.0001) and platelets (p
0.02) that had been transfused.
Conclusions. Red blood cell and platelet transfusion
have an unintended persistently negative risk-adjusted
effect on health-related quality of life after cardiac sur-
gery that extends well beyond initial hospitalization.
Reductions in functional recovery paralleled increasing
units of red blood cells transfused.
(Ann Thorac Surg 2006;82:13–20)
© 2006 by The Society of Thoracic Surgeons
T
ransfusion of packed red blood cells (PRBC) has been
associated with increased in-hospital morbidity and
mortality after cardiac surgery [1–5]. Additionally, trans-
fusion of PRBC has been associated with reduced sur-
vival, well beyond the accepted postoperative recovery
phase for cardiac surgical procedures [5]. Our objective
was to examine the impact of perioperative PRBC and
component transfusion on functional health-related
quality of life in a large cohort of patients after recovery
from cardiac surgery.
Patients and Methods
Patient Population and Data Collection
From May 1, 1995, through January 1, 1999, 12,536 patients
underwent isolated coronary artery bypass grafting, iso-
lated valve repair or replacement, or a combination coro-
nary artery bypass grafting and valve procedure at the
Cleveland Clinic Foundation after completing a self-
administered preoperative Duke Activity Status Index
(DASI) survey. Perioperative variables were prospectively
collected concurrently with patient care by dedicated indi-
viduals and entered into the Cleveland Clinic Foundation
Department of Cardiothoracic Anesthesia Registry. The
Cardiovascular Information Registry was used for addi-
tional variable information. The Social Security Death Index
was queried to determine death status during the follow-up
interval. Among these patients, 7,321 completed at least one
of two follow-up surveys scheduled nominally at 6 and 12
months postoperatively. Because our institution is a large
international referral center, 1,272 patients lacked a social
security number and therefore it was not logistically feasi-
ble to obtain mortality information for comparison. Among
nonresponders with social security numbers, 569 died
within the 6-month follow-up interval and 679 died within
the 12-month follow-up interval. Among patients who were
available to respond to the follow-up survey, 79% re-
sponded and 2,695 (21%) of patients failed to respond to the
follow-up DASI survey or refused to participate. If a patient
was unable to complete the questionnaire independently, a
research assistant administered it by reading the exact
words of the survey questions. Postoperative follow-up
DASI surveys were completed by means of telephone
interviews. Institutional review board approval was ob-
tained to perform research using these databases.
Duke Activity Status Index
Duke Activity Status Index is a disease-specific func-
tional quality-of-life questionnaire validated for patients
Accepted for publication July 25, 2005.
Address correspondence to Dr Koch, Department of Cardiothoracic
Anesthesia (G-3), 9500 Euclid Ave, Cleveland, OH 44195; e-mail:
kochc@ccf.org.
© 2006 by The Society of Thoracic Surgeons 0003-4975/06/$32.00
Published by Elsevier Inc doi:10.1016/j.athoracsur.2005.07.075
CARDIOVASCULAR
by on May 31, 2013 ats.ctsnetjournals.orgDownloaded from
with cardiovascular disease [6, 7]. The 12-item instrument
(see Table 1) measures activities of daily living, such as
household tasks, ambulation, personal care, sexual func-
tion, and recreational activities [6]. Weight of each item
was determined by measuring maximal oxygen con-
sumption at the level of activity represented by each
question. In its development, the Duke University inves-
tigators used multiple regression techniques both to
select items most correlated with oxygen uptake at a
given level of activity and to associate each selected item
with a weight reflecting metabolic cost. The DASI score
reflects one dimension of quality of life, physical func-
tioning. Positive responses are summed to produce an
aggregated score consisting of a limited set of numbers
ranging from 0 to 58.2 [6]. Higher values represent better
physical functioning.
Statistical Analysis
PRELIMINARY ANALYSIS. Tables 2 and 3 compare demograph-
ics, clinical history, symptomatic status, laboratory val-
ues, operative variables, and postoperative morbid
events between those responders who received and did
not receive a PRBC transfusion. Figures 1 and 2 depict the
frequency distribution for PRBC and platelet transfusion.
PROPORTIONAL ODDS REGRESSION. We initially focused on the
association between the first follow-up DASI score at
nominally 6 months after surgery and baseline risk
factors. There were 7,283 patients who completed the
survey between 5 and 9 months postoperatively.
Because the distribution of DASI scores were anoma-
lous with, for example, a large number of observations at
the highest attainable value and gaps in the distribution
of responses, we first grouped the raw DASI scores into
an ordinal scale using a data-driven approach. The
grouping began with an initial ordinal logistic regression
analysis treating each distinct score as its own group
represented by a unique intercept term. Groups were
then formed by consolidating similar intercept estimates
to filter out redundancy. The groups were collapsed
further so that each group had an adequate sample size.
This yielded a five-group scale: 0 to 34.7, 34.7 to 42.7, 42.7
to 45.2, 45.2 to 58.2, and 58.2.
Bootstrap aggregation [8] was used to identify baseline
characteristics, operative details, and postoperative mor-
Table 1. The Duke Activity Status Index (DASI) [6]
Activity: Can You . . . Weight
1. take care of yourself, that is, eating
dressing, bathing or using the toilet?
2.75
2. walk indoor, such as around your house? 1.75
3. walk a block or 2 on level ground? 2.75
4. climb a flight of stairs or walk up a hill? 5.50
5. run a short distance? 8.00
6. do light work around the house like
dusting or washing dishes?
2.70
7. do moderate work around the house like
vacuuming, sweeping floors, or carrying in
groceries?
3.50
8. do heavy work around the house like
scrubbing floors, or lifting or moving
heavy furniture?
8.00
9. do yard work like raking leaves, weeding
or pushing a power mower?
4.50
10. have sexual relations? 5.25
11. participate in moderate recreational
activities like golf, bowling, dancing,
doubles tennis, or throwing a baseball or
football?
6.00
12. participate in strenuous sports like
swimming, singles tennis, football,
basketball or skiing?
7.50
DASI is the sum of positive responses.
Table 2. Continuous Variables by Transfusion Status
a
Variable
Blood Transfusion No Blood Transfusion
N Median (25th, 75th percentile) N Median (25th, 75th percentile)
Demographics
Body surface area (m
2
)
4,195 1.9 (1.8, 2.1) 3,126 2.1 (1.9, 2.2)
Age at surgery (y) 4,195 69 (61, 75) 3,126 62 (53, 70)
Hematopoietic system
Blood volume estimate (L) 4,195 4.7 (4.0, 5.3) 3,126 5.3 (4.8, 5.8)
Red cell mass estimate (L) 4,195 1.8 (1.5, 2.1) 3,126 2.2 (1.9, 2.5)
Hematocrit (%) 4,195 39 (35, 42) 3,126 42 (40, 44)
Preoperative laboratory values
Bilirubin (mg/dL)
b
4,012 0.7 (0.5, 0.9) 3,022 0.7 (0.6, 0.9)
Albumin (mg/dL)
b
4,011 4.1 (3.7, 4.4) 3,021 4.3 (4.0, 4.5)
Blood urea nitrogen (mg/dL) 4,195 18 (14, 24) 3,126 16 (13, 20)
Creatinine (mg/dL) 4,195 1.0 (0.9, 1.3) 3,126 1.0 (0.8, 1.1)
Duke Activity Status Index
Preoperative baseline 4,195 24.2 (15.5, 41.0) 3,126 38.2 (23.5, 58.2)
Follow-up 4,195 42.7 (26.0, 58.2) 3,126 54.2 (38.2, 58.2)
a
p 0.01 for all variables by Wilcoxon rank sum test.
b
287 missing values for bilirubin and 289 albumin values.
14 KOCH ET AL Ann Thorac Surg
QUALITY OF LIFE, TRANSFUSION, AND CABG 2006;82:13–20
CARDIOVASCULAR
by on May 31, 2013 ats.ctsnetjournals.orgDownloaded from
bid events that were associated with follow-up DASI
scores. Two hundred bootstrap data sets were used with
a p value for retention of 0.05. Variables selected more
than 50% of the time were submitted to further logistic
ordinal regression analyses. A score test and a goodness-
of-fit test were used to check the proportional odds
assumption. The influence of baseline DASI; preopera-
tive risk factors; laboratory values; platelet, fresh frozen
plasma, and cryoprecipitate transfusion; operative
events; and postoperative morbidities associated with
follow-up DASI were examined with ordinal regression
modeling.
Table 3. Categorical Variables by Transfusion Status
Variable
Blood Transfusion No Blood Transfusion
p Value
a
N (%) N (%)
Demographics
Females 1,788 42 571 18 0.01
Cardiac comorbidity
Hypertension 2,519 60 1,714 55 0.01
Heart failure 1,246 30 545 17 0.01
Prior cardiac surgery
0 3,087 74 2,629 84 0.01
1 886 21 416 13
1 222 5 81 3
Noncardiac comorbidity
Chronic obstructive pulmonary disease 351 8.4 179 5.7 0.01
Smoking 2,392 57 1,911 61 0.01
Insulin-dependent diabetes 431 10 211 6.8 0.01
Non–insulin-dependent diabetes 550 13 333 11 0.01
Renal disease 58 1.4 6 0.19 0.01
Stroke 304 7.3 150 4.8 0.01
Peripheral vascular disease 549 13 226 7.2 0.01
Prior vascular surgery 181 4.3 76 2.4 0.01
Carotid disease 733 18 249 8.0 0.01
Symptoms
NYHA functional class
b
I 570 14 469 15 0.01
II 1,832 44 1,652 53
III 751 18 417 13
IV 1,040 25 586 19
Emergency surgery 151 3.6 19 0.61 0.01
Preoperative intraaortic balloon pump 116 2.8 26 0.83 0.01
Ventricular function
Left ventricular ejection fraction 50% 2,909 69 2,425 78 0.01
History of myocardial infarction
b
1,986 47 1,251 40 0.01
Procedure
Valve procedure 1,722 41 1,269 41 0.7
CABG 3,143 75 2,076 66 0.01
LITA (% of CABG) 1,930 61 1,589 77 0.01
RITA (% of CABG) 379 12 372 18 0.01
Postoperative morbid events
Total intubation 72 hours 325 7.8 13 0.42 0.01
Tracheostomy 1 0.02 0 0.0 0.99
Focal or global neurologic deficit 87 2.1 13 0.42 0.01
Low cardiac output 67 1.6 3 0.1 0.01
Serious infection 140 3.3 3 0.1 0.01
Renal failure 26 0.62 0 0.0 0.01
a
2
test.
b
4 NYHA missing values, 2 missing values for history of myocardial infarction.
CABG coronary artery bypass grafting; LITA left internal thoracic artery; NYHA New York Association functional classification; RITA
right internal thoracic artery.
15Ann Thorac Surg KOCH ET AL
2006;82:13–20 QUALITY OF LIFE, TRANSFUSION, AND CABG
CARDIOVASCULAR
by on May 31, 2013 ats.ctsnetjournals.orgDownloaded from
RESPONDERS VERSUS NONRESPONDERS. Characteristics of the
nonresponders were generally similar to responders ex-
cept for a small group of 346 patients who had a low
preoperative DASI, more comorbidity, different opera-
tions, more PRBC transfusions, and more postoperative
morbidity. It is likely their follow-up DASI would have
been lower than average, accentuating the findings of our
study.
All statistical analyses were done using SAS 8.2 (SAS
Institute, Cary, NC) and R (www.r-project.org).
Results
Patients who received a perioperative PRBC transfusion
had lower follow-up DASI scores at nominally 6 months
compared with those who did not. Among the patients
who received a transfusion, only 31% of patients
achieved the highest follow-up score 58.2, whereas 42%
of patients who did not receive a transfusion did so (Fig
3). Even after adjusting for variables known to be asso-
ciated with functional recovery after cardiac surgery,
PRBC units and platelet transfusion were associated with
significantly reduced postoperative functional recovery
(Table 4). These relationships are evident from the pro-
portion of patients in each of the five follow-up DASI
groups according to PRBC units transfused and platelet
usage (Fig 4). As PRBC transfusion increased, more
patients were in the lowest DASI score group (0 to 34.7)
and fewer achieved the highest DASI score (58.2) group.
The predicted probability of achieving the highest
follow-up DASI by increasing age and transfusion status
is shown in Figure 5. The depiction is based on a male
patient, undergoing coronary artery bypass grafting with
a left internal thoracic artery graft and with a preopera-
tive DASI score of 50.7, a serum creatinine of 0.89 mg/dL,
a bilirubin of 0.90 mg/dL, a hematocrit of 38%, and no
other preoperative comorbidity or postoperative morbid
events. The predicted probability of achieving the high-
est follow-up DASI category decreases with increasing
age and is further decreased with the addition of a blood
transfusion.
Female sex, older age, history of chronic obstructive
pulmonary disease, diabetes, stroke, peripheral vascular
disease, higher preoperative serum creatinine, and oc-
currence of postoperative neurologic events were also
associated with worse postoperative functional recovery.
High baseline DASI score, use of an internal thoracic
artery graft, and having an isolated valve procedure were
associated with better postoperative functional recovery.
Comment
Principal Findings
Although patient-centered outcomes have historically
received little attention [9], they represent important end
points with regard to assessing the success of an opera-
tion. We report that functional recovery after cardiac
surgery is significantly reduced in patients who received
PRBC transfusion perioperatively. Transfusion of PRBC
remained a significant predictor for poor postoperative
functional recovery even after maximal adjustment for
baseline DASI, preoperative clinical status, comorbidi-
ties, hematopoietic system measurements, laboratory
values, operative factors, surgical procedure, and postop-
Fig 1. Frequency histogram of red blood cell (RBC) units transfused.
Fig 2. Frequency histogram of platelet units transfused.
Fig 3. Mirrored histogram displays nominally 6-month follow-up
Duke Activity Status Index scores according to whether red blood
cells had been transfused. (Blue bars no blood transfusion; red
bars red blood cell transfusion.)
16 KOCH ET AL Ann Thorac Surg
QUALITY OF LIFE, TRANSFUSION, AND CABG 2006;82:13–20
CARDIOVASCULAR
by on May 31, 2013 ats.ctsnetjournals.orgDownloaded from
erative morbid events. The model presented in Table 4
suggests that a transfusion of 4 PRBC units and a post-
operative infectious or neurologic morbid event have
equivalent effects on 6-month functional recovery; a
similar comparison can be made between a 2-unit trans-
fusion of PRBC and a preoperative history of chronic
obstructive pulmonary disease.
In addition, there appeared to be a dose–response
relationship, with incrementally poorer functional recov-
ery with each unit of blood transfused. Figure 4 graphi-
cally depicts the increasing probability of achieving the
lowest DASI functional score group (0 to 34.7) as the
number of PRBC units increases. The probability of a
patient achieving the highest DASI functional score
group (58.2) dramatically decreases as the number of
PRBC units transfused increases. The impact of platelet
transfusion follows a similar graphical trend.
Increasing age was also associated with poorer func-
tional health status postoperatively. Older patients’ func-
tional status was further reduced by the addition of PRBC
transfusion. As Figure 5 depicts, the older the patient age,
the less probability of achieving the highest functional
DASI score group (58.2). For patients who then receive
PRBC transfusion, further reductions in the probability of
achieving the highest DASI (58.2) functional score group
are noted.
Other Factors Related to Quality of Life After Cardiac
Surgery
Several studies have examined preoperative predictors of
quality of life after cardiac surgery with a variety of
quality-of-life instruments [10–14]. Our findings of
poorer functional recovery for