Keywords
obesity - BMI - CABG - intraoperative transfusion - chest-tube output
Obesity is a global health care concern with rising trends across the world and the
United States. In 2008, the World Health Organization found 1.5 billion adults were
overweight with a staggering 200 million men and almost 300 million women being obese.[1] In the United States, 60 to 70% of adults are overweight or obese, and the prevalence
of obesity continues to rise with nearly one in three adults classified as obese.[2]
[3] This patient population has become an area of focus as obesity can affect a wide
variety of clinical parameters in addition to being associated with increased risk
of heart disease, stroke, and type 2 diabetes.[2]
[3] In addition, in cardiac surgery, although there seems to be no effect on mortality,
controversy exists regarding increased morbidity in obese patients undergoing heart
operations.[4]
[5]
[6] Thus, specific variables are being assessed to determine the effect of obesity on
individual clinical parameters in cardiac procedures. In this study, we examined coronary
artery bypass graft (CABG) surgery and how obesity impacted intraoperative transfusion
requirements and postoperative bleeding.
Understanding intraoperative transfusion requirements will streamline blood bank usage
and facilitate patient-specific interventions. Preoperative parameters that predict
usage are varied, with clinical assessment tools already in place for many operative
indications such as trauma patients requiring massive transfusion.[7] Cardiac surgery in particular routinely utilizes red blood cells, and research is
ongoing for evaluating which patients will need blood and how much.[8]
[9]
[10]
Just as with transfusions, a tool to evaluate patients preoperatively for their estimated
blood loss postoperatively will give better patient-specific care. Research evaluating
which factors can predict postoperative blood loss exists[11]
[12]; however, the large proportion of existing literature evaluates blood loss based
on transfusion requirements or looks purely at rate of reoperation for bleeding. In
our study, we diverged from this practice and utilized chest-tube output as a direct
means of measuring blood loss at the surgical site.
A previous study by our group found a decrease in both transfusions and chest-tube
output with increasing obesity in CABG surgery.[13] We wanted to further evaluate this relationship, however, with the goal of our current
study being to utilize a much larger cohort and multivariable analyses. These new
methods allowed us to ascertain the significance of BMI as an independent preoperative
indicator of intraoperative transfusion and postoperative blood loss.
Methods
Institutional Review Board approval was obtained prior to data collection for this
study. Data were collected via a retrospective chart review. Data were collected from
consecutive patients who received isolated CABG surgery (on- or off-pump), performed
by a single surgeon at the University of Kentucky Medical Center between November
2003 and April 2009. A total of 338 charts were reviewed with 290 patients meeting
our inclusion criteria included in the analysis ([Table 1]). Patients undergoing simultaneous valve procedures, aneurysms, or patent foramen
ovale repair were excluded. Patients whose charts were missing specified preoperative
variables including preoperative hematocrit and platelet levels were also excluded.
Table 1
Patient demographics relative to each analyzed preoperative variables, n = 290
|
Variable
|
Incidence
|
|
Age
|
Mean 61.9 ± SD 10.5
|
|
Male
|
83.8% (n = 243)
|
|
Normal hematocrit
|
64.5% (n = 187)
|
|
Low hematocrit
|
17.9% (n = 52)
|
|
High hematocrit
|
17.6% (n = 51)
|
|
Normal platelets
|
91.7% (n = 266)
|
|
Low platelets
|
8.3% (n = 24)
|
|
Normal BMI
|
19.3% (n = 56)
|
|
Overweight BMI
|
43.5% (n = 126)
|
|
Obese BMI
|
37.2% (n = 108)
|
|
On-pump CABG
|
84.8% (n = 246)
|
|
Off-pump CABG
|
15.2% (n = 44)
|
Abbreviations: SD, standard deviation; BMI, body mass index; CABG, coronary artery
bypass graft.
BMI was calculated by dividing the individual patient's weight in kilograms by the
individual patient's height in meters squared. Following American Heart Association
and the World Health Organization guidelines, patients were divided into groups based
on BMI and included normal weight (BMI < 25), overweight (BMI 25 to 29), and obese
(BMI ≥ 30).[1]
[14] Hematocrit was categorized as normal (range of 35.1 to 44.2%), low (< 35%), and
high (> 44.3%), with the normal hematocrit group used as reference. Platelet count
was categorized as normal or low (< 150 × 103 per milliliter) with a normal platelet count used as reference. Packed red blood
cell transfusions were assessed by the total number of transfusions received intra-operatively.
Transfusion triggers included patients with a hematocrit of less than 21% or hemoglobin
less than 7 g/dL or elderly/actively bleeding patients with a hematocrit of less than
24% or a hemoglobin less than 8 g/dL. Final transfusion decisions were made intraoperatively
by the surgeon. Chest-tube output was calculated using the amount in milliliters collected
from the chest-tube draining the surgical site during the immediate 24-hour postoperative
period. Chest-tube placement was standard of care for the procedure, and values were
recorded as total amount regardless of number of chest tubes placed.
Two separate multivariable regression analyses were done. Logistic regression evaluated
the likelihood of intraoperative transfusion related to age, gender, BMI, preoperative
hematocrit, preoperative platelet count, and procedure type (off-pump vs. on-pump).
Linear regression evaluated the natural logarithm of 24-hour chest-tube output related
to the same variables. In both analyses, the significance level was set at p < 0.05. SPSS™ Version 19 software was used for all statistical analysis (SPSS, Chicago,
IL).
Results
Among these 290 patients, the mean age was 61.9 years (SD of 10.5 years), and the
majority of patients were male (n = 243, 83.8%). The normal hematocrit group accounted for 64.5% of the cohort (n = 187), whereas the low hematocrit group accounted for 17.9% (n = 52), and the high hematocrit group was similarly 17.6% (n = 51). The normal platelet group accounted for 91.7% (n = 266) of the cohort and the low platelet group 8.3% (n = 24). The normal weight BMI group accounted for 19.3% of the cohort (n = 56), the overweight group for 43.5% (n = 126), and the obese for 37.2% (n = 108). On-pump procedures accounted for 84.8% of the cohort (n = 246), whereas 15.2% had off-pump procedures (n = 44).
Eighty-five patients received intraoperative transfusions (29.3%). Preoperative variables
that significantly increased the likelihood of intraoperative transfusions were older
age and low hematocrit, whereas male gender, overweight and obese BMI groups, and
off-pump procedures decreased the likelihood of intraoperative transfusions ([Table 2]). Low platelet count and high hematocrit were not significant contributors to intraoperative
transfusion.
Table 2
Logistic regression analysis of intraoperative transfusion
|
Incidence of intraoperative transfusions (n = 85; 29.3%)
|
|
Variable
|
Odds ratio
|
95% Confidence interval
|
P-value
|
|
Age
|
1.045
|
1.013–1.077
|
0.005
|
|
Male
|
0.138
|
0.062–0.305
|
< 0.001
|
|
Low hematocrit
|
5.473
|
2.591–11.562
|
< 0.001
|
|
High hematocrit
|
0.373
|
0.128–1.093
|
0.072
|
|
Low platelets
|
1.730
|
0.597–5.013
|
0.312
|
|
Overweight BMI
|
0.359
|
0.165–0.783
|
0.010
|
|
Obese BMI
|
0.327
|
0.145–0.739
|
0.007
|
|
Off-pump
|
0.236
|
0.084–0.668
|
0.007
|
Abbreviations: BMI, body mass index; CABG, coronary artery bypass graft.
The median 24-hour chest-tube output amount was 823 mL with an interquartile range
of 631 to 1,070 mL. Preoperative variables that significantly increased 24-hour chest-tube
output were low hematocrit, high hematocrit, and low platelets whereas overweight
and obese BMI significantly decreased 24-hour chest-tube output ([Table 3]). Due to the natural log transformation, the regression coefficients should be interpreted
as the percent change in output per unit change of the predictor variable. Male gender,
age, and off-pump procedure were not significant predictors of 24-hour chest-tube
output.
Table 3
Linear regression analysis of 24-hour chest-tube output
|
24-hour chest-tube output[a]
|
|
Variable
|
Regression coefficient
|
95% Confidence interval
|
P-value
|
|
Age
|
1.044
|
0.999–1.009
|
0.080
|
|
Male
|
0.994
|
0.865–1.141
|
0.930
|
|
Low hematocrit
|
1.152
|
1.016–1.308
|
0.028
|
|
High hematocrit
|
1.153
|
1.008–1.319
|
0.038
|
|
Low platelets
|
1.313
|
1.132–1.523
|
< 0.001
|
|
Overweight BMI
|
0.836
|
0.738–0.947
|
0.005
|
|
Obese BMI
|
0.791
|
0.691–0.905
|
0.001
|
|
Off-pump
|
0.987
|
0.857–1.136
|
0.852
|
a Median output = 823 mL.
Abbreviations: BMI, body mass index; CABG, coronary artery bypass graft.
Our results are in line with several studies that have shown a lower BMI to be related
to increased transfusion requirements.[15]
[16]
[17]
[18] Additional studies have had similar results using BSA or weight as the variable
and noted an increase in transfusion with a lower BSA or lower weight.[10]
[19]
[20]
[21] There is difficulty correlating many of these results with our study for several
reasons. When BMI is used as a variable it is often defined as simply low BMI or as
obese.[15] This does not allow for separate analysis of BMI categories beyond the extreme ranges
of the scale. The advantage to fully stratifying the classifications is that it leads
to a better understanding of more moderate changes in weight. The fact that we saw
significance in our results with overweight and obese BMI points to the idea that
excess body weight is a contributing factor to transfusion requirements and it is
not merely just the extremes of the malnourished and the morbidly obese that cause
variances from normal. One study that did analyze BMI groups was Reeves et al who
examined underweight, normal, overweight, obese, and severely obese patient populations
undergoing CABG.[18] Reeves et al found that the overweight, obese, and morbidly obese had a linearly
decreasing odds ratio relevant to red blood cell transfusion.[18] The difficulty with the Reeves study, however, is a complication in many additional
studies. These studies record transfusions as an occurrence and do not further categorize
transfusions based on broad timing of transfusion (i.e., intraoperative or postoperative).
Without such characterization it is difficult to assess and estimate blood product
requirements during the CABG procedure. This is further proved by the results of our
study in which the variables relating to postoperative bleeding are not the same as
those that predict intraoperative blood product needs. Classifying transfusions based
on time frame can help the surgeon and blood bank to predict the amount of blood product
needed in a timelier manner; specifically, predicting intra-operative blood usage
can better define the product that needs to be in the operating room at the time of
the procedure. Our results show that intraoperative blood transfusions requirements
are decreased for the overweight and obese BMI populations.
Discussion
Our results demonstrating BMI is a significant preoperative indicator of decreased
chest-tube output and thus decreased postoperative bleeding correlates well with other
similar studies. It is important to note that chest-tube output is only a surrogate
for surgical site bleeding as other factors can contribute to drainage amount, including
amount of fluid left in the pleural cavity from irrigation or intraoperative blood
loss for example. There is evidence, however, that chest-tube site location does not
affect output which allows us to subtract that variable from our results as our analysis
did not ascertain location of chest-tubes.[22] As mentioned earlier, historically studies have evaluated postoperative bleeding
based either on required transfusions or reoperation for bleeding. Several studies
evaluating reoperation for bleeding have had similar results to our findings with
a decrease in reoperation in the obese patient population.[4]
[23]
[24]
[25] Interestingly, Alam et al examined postoperative bleeding and reoperation for bleeding
and found obesity was associated with a decrease in both parameters[23]; however, the method by which postoperative bleeding was measured was not included
in the publication. Again the closest study assimilating our results is that by Reeves
et al[18] who examined not only reoperation for bleeding but also blood loss as evaluated
by chest-tube output. Their results showed that overweight, obese, and severely obese
populations appeared to be protected against reoperation and chest-tube output greater
than 1,000 mL. In this output category the odds ratio was linear; however, further
analysis of amounts less than 1,000 mL was not discussed. Yet, utilizing the greater
than 1,000 mL parameter is still appropriate and applicable to our findings as Wynne
et al found that mean total chest-tube output following CABG is roughly 1,300 mL.[26] Correlating these findings with the results of our study, our results further show
that overweight and obese patient populations show a decreased percentage change in
chest-tube output in the 24-hour postoperative period indicating that increasing BMI
is protective against postoperative surgical site blood loss.
Limitations of this study revolve mostly around it being a retrospective study as
opposed to a prospective study. With a retrospective approach, there were several
parameters that were unknown or not well documented and thus were unable to be included
in this analysis. Such parameters include medications such as aspirin and other prescriptions
affecting the coagulation cascade, which have proven to affect the amount of chest-tube
output and transfusion requirements of CABG patients.[27] In addition, preoperative percutaneous intervention and specific operative variables
such as cardiopulmonary bypass circuit type that could affect transfusions and bleeding
were not evaluated.[28]
The results from our study indicate that indeed obesity as measured by BMI has a significant
effect on intraoperative transfusion and postoperative bleeding. Our results are unique
in that we evaluate transfusions requirements based on their relation to the procedure
itself and that we define and stratify obesity based on BMI categories. In addition,
our results are one of the few that utilizes surgical site drainage as a measure of
postoperative bleeding and found that the preoperative variables affecting this parameter
are not the same as those that would require transfusion in the operative time period.
As such, we believe BMI to be an important variable in preoperative patient assessment.
Further research will elucidate how to incorporate BMI into a predictive model that
will give the surgeon a preoperative prediction of patient-specific coagulation. As
a variable, BMI is easy to ascertain and does not require additional costs. The only
potential obstacles to its collection are emergency situations in which these data
points are not able to be collected. With respect, an urgent/emergent situation is
in itself a separate entity and as such has shown to be strongly associated with increased
transfusion rates.[29] In addition to the role BMI will play in predictive tools, further research is also
necessary at the molecular level to understand the etiology behind BMI's effect on
coagulation. Regardless of where this research leads, this study proves that it is
important at the clinical level to take into account a patient's BMI when preparing
for the perioperative sequelae of CABG surgery.
Conclusion
Overweight and obese BMI is a significant independent predictor of decreased intraoperative
transfusion and decreased postoperative blood loss in CABG surgery patients. These
values were significant even when adjusted for low preoperative hematocrit and platelets
and on-pump versus off-pump procedure type.
