The influence of bilateral internal ma

mmary arteries on short- and long-ter

m outcomes A propensity score matc
hing in accordance with current reco
mmendations

Umberto Benedetto, MD, PhD, Mohamed Amrani, MD, PhD, FETCS, Jullien Gaer, MD, FR
CS(C-Th), ToufanBahrami,MD, Fabio deRobertis, MD, AndreR. Simon,MD, and ShahzadG
.Raja,MRCS,FRCS(C-Th), on behalf of the Harefield Cardiac Outcomes Research Group. T
he Journal of Thoracic and Cardiovascular Surgery c Volume 148, Number 6, 2699-2705
.

PRESENTER: INTERN
SUPERVISOR:

Background
Previous observational cohort studies comparing
the use of bilateral internal mammary arteries (BIM
A) and single internal mammary artery (SIMA) sugge
st a survival advantage from BIMA use.
Ongoing RCT.
Observational cohort studies are liable to methodol
ogic flaws.

Background
Propensity score matching: creates groups of patients
who have similar pretreatment characteristics when
treatment assignment is not random.
Only 4 out of 27 observational studies included in a rece
nt systematic review using PS matching demonstrated a
survival benefit from BIMA.
Surgeons are reluctant to use BIMA routinely because
concerns still exist regarding the detrimental effect of thi
s strategy on operative outcomes, including the increas
ed risk for (deep) sternal wound complications.

Purpose
To validate the safety and efficacy of BIMA grafting a
nd enhance its universal adoption
To evaluate the influence of BIMA over SIMA on sho
rt- and long-term outcomes

Methods
Propensity score match
Patients

First-time isolated CABG, 2 grafts received

SIMA group
Single left internal mammary artery to left anterior descending (LAD) artery
and additional saphenous vein grafts (SVG) for non-LAD targets

BIMA group
BIMA with or without additional SVG

Strict blood glucose control according to a unique intravenous

insulin therapy protocol. The target blood glucose was 100 to
139 mg/dL (5.5-7.7 mmol/L)

Methods
Pretreatment variables: age, female gender, New York Hear
t Association functional class III to IV, prior myocardial infa
rction prior percutaneous coronary intervention (PCI), diab
etes mellitus, current smoking, chronic obstructive pulmon
ary disease, cerebrovascular accident, peripheral vascular d
isease (PVD), history of atrial fibrillation, number of vessels
diseased, left main stem disease, left ventricular ejection fr
action (LVEF) <50%, renal impairment defined as a serum cr
eatinine>200 mmol/L, obesity defined as body mass index>
30, urgent/emergent indication, preoperative use of intra-a
ortic balloon pump, surgery performed by a resident physic
ian, and the use of cardiopulmonary bypass.

Methods
Short-term outcomes: the incidence of superficial stern
al wound infection (SSWI) and deep sternal wound infec
tion (DSWI) as defined by the Centers for Disease Contr
ol and Prevention, postoperative cerebrovascular accide
nt, need for renal replacement therapy, reintubation for
acute respiratory failure, reexploration for bleeding, pos
toperative atrial fibrillation, and operative mortality (wit
hin 30 days).
Long-term outcomes: all-cause late mortality and need f
or repeat revascularization, including PCI and/or redo C
ABG

Results

Results

Results

Results

Results
P=0.04

P=0.004

Results
Total number of grafts per patients was slightly high
er in the BIMA group (2.88 grafts per patient) than i
n the SIMA group (2.74 grafts per patient) (P = .01).

Results

Results

Discussion
SIMA

BIMA

SSWI
Other short term
outcomes
(including DSWI)

Superior
Comparable

Inferior
Comparable

Survival

Inferior

Superior

Revascularization
free survival

Inferior

Superior

Conclusion
When compared with SIMA grafting, BIMA us
e did not increase the incidence of operative
complications, including DSWI, and improve
d long-term overall survival and repeat revas
cularization-free survival.

Thank you for y

our attention!