CARDIOVASCULAR
Radial Artery Angiographic String Sign: Clinical
Consequences and the Role of Pharmacologic
Therapy
Senri Miwa, MD, PhD, Nimesh Desai, MD, Tadaaki Koyama, MD, PhD, Emily Chan, BS,
Eric A. Cohen, MD, MS, and Stephen E. Fremes, MD, MS, for the Radial Artery Patency
Study Investigators
Divisions of Cardiovascular Surgery and Cardiology, Sunnybrook and Womens College Health Sciences Centre, Toronto, Ontario,
Canada
Background. The radial artery is an increasingly impor-
tant graft for coronary artery bypass surgery. Postopera-
tive angiographic studies have shown that a proportion
of radial grafts become diffusely narrowed but not oc-
cluded, or string signs.
Methods. Four hundred forty patients receiving a radial
artery graft enrolled in a large clinical trial underwent
postoperative angiography at 1 year. Angiograms were
analyzed visually and quantitatively. A complete string
sign was defined as diffuse narrowing along the full
length of the graft, while a partial string sign was defined
as segmental narrowing. Angiographic findings were
correlated with medication compliance and clinical
sequelae.
Results. Thirty-one patients (7.0 %) had radial artery
graft string signs versus 4 patients (0.9%) with a saphe-
nous vein graft string sign (p 0.001). Complete string
signs were present in 28 cases, and the mean diameter
was 0.76 0.14 mm (mean SD), whereas 3 cases had a
I nitial angiographic results using the radial artery as a
graft for coronary artery bypass grafting (CABG) were
unsatisfactory [1, 2]. In 1989, Acar and coworkers [3]
reintroduced the use of the radial artery for CABG, with
encouraging initial findings. Acceptable results were
subsequently reported from several groups, and the
radial artery has become an increasingly important graft
for coronary artery bypass surgery.
While arterial grafting is broadly recommended for
coronary surgery [4], arterial grafts still occasionally fail,
often as a diffuse narrowing seen on angiography as a
string sign. Only a few published reports of radial
grafting have focused on angiographic string sign, and
the conclusions of the studies are limited owing to the
Accepted for publication June 8, 2005.
Presented at the Forty-first Annual Meeting of The Society of Thoracic
Surgeons, Tampa, FL, Jan 24 26, 2005.
Address correspondence to Dr Fremes, Division of Cardiovascular Sur-
gery, Schulich Heart Centre, Sunnybrook and Womens College Health
Sciences Centre, 2075 Bayview Ave, Room H410, Toronto, ON M4N 3M5,
Canada; e-mail: stephen.fremes@sw.ca.
2006 by The Society of Thoracic Surgeons
Published by Elsevier Inc
partial string sign with a diameter of 0.89 0.14 mm.
Fifteen radial arteries showed Thrombolysis in Myocar-
dial Infarction Study (TIMI) 1 flow, 3 cases showed TIMI
2 flow, and 13 cases showed TIMI 3 flow. There was no
difference in incidence of radial string sign between
patients taking nifedipine versus diltiazem postopera-
tively. Multivariate analysis revealed the presence of
radial artery string sign was closely related to the peri-
operative use of alpha-adrenergic agonists and target
vessels stenosis less than 90%. Postoperative symptoms
were associated with radial artery string signs with TIMI
1 flow (p 0.0045).
Conclusions. In the Radial Artery Patency Study, radial
artery string sign was present in 7% of patients. Despite
diffuse narrowing, 52% of grafts had TIMI 2 flow or
better.
(Ann Thorac Surg 2006;81:1129)
2006 by The Society of Thoracic Surgeons
retrospective nature of the investigations or the small
number of patients evaluated [5, 6].
To definitively assess the role of the radial artery as a
coronary bypass conduit, we initiated a randomized trial,
the multicenter Radial Artery Patency Study (RAPS), in
1996 [7] (see Appendix). The primary results were re-
ported in 2004 [8]. Altogether 561 patients were enrolled,
of whom 440 underwent postoperative angiography. Al-
though radial artery patency was substantially superior
to saphenous vein graft patency, there were several
instances of patent radial arteries demonstrating angio-
graphic string sign. The purpose of this study was to
define the etiology and clinical consequences of radial
artery angiographic string signs.
Patients and Methods
Details of the study design have been previously pub-
lished [7]. Patients received both a radial artery and a
saphenous vein graft, randomly anastomosed to two
different coronary territories. The primary study objec-
tive was to determine the angiographic patency of radial
0003-4975/06/$32.00
doi:10.1016/j.athoracsur.2005.06.076
Ann Thorac Surg
2006;81:1129
artery grafts compared with saphenous vein grafts at 8 to
12 months after surgery. Elective, primary isolated coro-
nary bypass surgery patients with graftable triple vessel
disease and preserved left ventricular ejection fraction
were candidates for the study. As entry criteria, the left
circumflex and the right coronary arteries were required
to have proximal lesions with at least 70% stenosis and to
be greater than or equal to 1.5 mm in diameter and of
acceptable quality according to the visual assessment of
the angiogram. Patients with nonpalpable ulnar arteries
or a positive Allens test, an abnormal upper extremity
Doppler study or ultrasonogram, and a history of vascu-
litis or Raynauds syndrome were ineligible. Patients
with bilateral varicose veins or stripping were ineligible.
Patients with any condition that would preclude fol-
low-up research angiography on an ethical basis were
also excluded.
Study Protocol
The patients signed a consent form approved by the
Research Ethics Committee at each participating center.
Patients were randomly assigned to one of two graft
strategies: (1) the radial artery was used to graft the
circumflex territory and a saphenous vein graft was used
for the right coronary system; or (2) the radial artery was
directed to the right coronary territory and a saphenous
vein graft was used for the circumflex system. By proto-
col, the internal thoracic artery was used for revascular-
ization of the left anterior descending artery. Additional
grafts, regardless of conduit used, were optional. The
nondominant arm was used exclusively for radial artery
harvesting. Details of the surgical technique have been
previously published [9]. The radial artery pedicle was
dilated in situ by a slow intraluminal injection of 4 to 5
mL of a dilute solution of verapamil and papaverine (5
mg verapamil and 65 mg of papaverine in 16 mL of
lactated Ringers solution). Sequential grafts were per-
formed in two of 440 radial arteries, one to the right
coronary territory and one to the circumflex territory.
Aortocoronary grafts were performed in 433 of 440 radial
artery grafts. Two radial arteries were used as Y-graft
extensions, both taken off of nonstudy saphenous vein
grafts and anastomosed to the circumflex territory. Five
additional radial grafts were anastomosed to vein graft
hoods proximally.
Postoperative Management
Patients were given aspirin 325 mg daily starting within 6
hours after surgery and continued indefinitely. By proto-
col, patients received intravenous nitroglycerin 1 to 10
ug kg1 min1 during the first 24 hours after surgery in
the intensive care unit. Alpha-adrenergic agonist agents
such as norepinephrine and neosynephrine were
avoided unless other methods to improve hemodynamic
status were ineffective. Oral nifedipine was initiated the
first postoperative day and continued for 6 months post-
operatively. In cases of where patients were intolerant of
nifedipine or institutional policy discouraged nifedipine
use, diltiazem or amlodipine were offered.
MIWA ET AL 113
CARDIOVASCULAR
RADIAL ARTERY STRING SIGN
Follow-Up Angiography
Patients were scheduled to undergo follow-up angiogra-
phy between 8 and 12 months after surgery. The angio-
gram was not performed if any conditions listed as
preoperative exclusions developed in the postoperative
period. Angiograms were generally performed on an
outpatient basis. Nitroglycerin was injected into each
graft prior to filming. At least two orthogonal views of
each graft were obtained, and continued exposure was
used as required to visualize the distal runoff and the size
of the target arteries.
Study Endpoints
The Core Angiographic Committee determined the pres-
ence of string signs as a prespecified secondary angio-
graphic end-point. The Core Angiographic Committee
consisting of four cardiologists experienced with angiog-
raphy, reviewed the postoperative angiograms indepen-
dently in a masked fashion. The films were reviewed by
two of the four Committee members with a third review
in the case of disagreement.
Secondary clinical endpoints included myocardial in-
farction and recurrence of anginal symptoms as deter-
mined by the Seattle Angina Questionnaire. Serial elec-
trocardiograms were obtained perioperatively.
Definition and Analysis of String Sign
String sign was defined a priori as a patent graft that was
diffusely narrowed, with a diameter less than 1 mm
according to visual assessment by the Core Angiography
Committee. For this study, the diameter of the string
graft was directly measured in the proximal, mid-body
and distal portions using quantitative coronary analysis
with the Inturis CVS Software package (Philips Medical
Systems, DA Best, the Netherlands) [10, 11]. Each image
was calibrated by selecting a straight segment of the
catheter. The manual curve selection tool was used to
select the graft portion. In cases where the vessel was not
successfully selected, the selection was modified at the
users discretion. For example, when the graft overlaps
existing sternal wires, clips, and other coronary vessels,
these extraneous objects were excluded from the selec-
tion. Once the curve selection was accepted, the mean
diameter was collected from the total segment analysis.
Empiric determination of interobserver and intraob-
server reliability of quantitative angiographic analysis
showed that our quantitative coronary analysis of radial
artery grafts was highly reliable.
A complete radial artery string sign was defined as
diffuse narrowing along the full length of the graft. In
addition, we identified several grafts with segmental
narrowing of a significant portion of the graft, but not
the whole graft, which we defined as partial string
signs. Quantitative coronary analysis was performed in
the 24 of 31 patients in whom the postoperative angio-
gram was recorded digitally. Quantitative coronary
analysis was also performed in 185 fully patent radials
for comparison.
 114 MIWA ET AL Ann Thorac Surg
CARDIOVASCULAR

RADIAL ARTERY STRING SIGN 2006;81:1129

Table 1. Patency and String Signs in All Study Grafts Quantitative Analysis
Radial Grafts Saphenous Vein Grafts Twenty-eight patients had complete string signs, and 3
patients had partial string signs; the distal portion of the
Total String Sign Total String Sign graft was affected in each partial string sign case. The
n (%) n (%) n (%) n (%) mean diameter of the complete radial artery string sign
TIMI 0 36 (8.2%)a 0 (0%) 60 (13.6%)a 0 (0%) was always less than 1.05 mm (range, 0.46 to 1.04 mm) by
TIMI 1 15 (3.4%) 15 (48%) 2 (0.5%) 0 (0%) quantitative coronary analysis, and the average diameter
TIMI 2 3 (0.7%) 3 (10%) 1 (0.2%) 0 (0%) was 0.76 0.14 mm. By comparison, the average diame-
TIMI 3 386 (87.7%) 13 (42%) 377 (85.7%) 4 (100%) ter of normal patent radial arteries without any string
Total 440 31b 440 4b
sign was 2.42 0.78 mm. The diameter of the radial string
number of sign increased nonsignificantly with TIMI flow classifica-
grafts tion: 0.66 0.12 mm in the TIMI 1 grafts, 0.78 0.16 mm
in TIMI 2 grafts, and 0.88 0.10 mm in TIMI 3 grafts. The
a
Patency of radial grafts were better than saphenous vein grafts (p
0.009). b
The incidence of radial artery string sign was greater than the
average diameter of the string portion of radial grafts
incidence of string sign of saphenous vein grafts (p 0.001). with a partial string sign was 0.89 0.14 mm, and the
average diameter of the nonstring portion was 2.20 0.31
mm. In 9 of the 31 patients, angiographic views of string
Statistical Analysis radial artery grafts were taken before and after the
The comparisons between the incidence of graft occlu- infusion of nitroglycerin. After intragraft nitroglycerin
sion or string sign in the radial and saphenous vein grafts infusion, the graft diameters of these radial string grafts
were performed with McNemars test for paired propor- increased slightly from 0.74 13 mm to 0.80 0.12 mm
tional data. Continuous variables, such as graft diameter, (p 0.0063).
were compared using Students t test with correction for
paired analyses where necessary. Dichotomous variables Impact of Calcium-Channel Blocker Usage
were compared using either the 2 test or Fishers exact Compliance on calcium-channel blocker therapy among
test where indicated. patients who underwent angiography was 95.2% at dis-
Multivariable analysis was performed using logistic charge and 91.1% for the first 6 months after discharge.
regression. A logistic regression model to determine risk The incidence of radial artery string sign was 28 of 419
factors for radial artery occlusion was performed using (6.7%) for patients discharged on a calcium-channel
the following predictor variables: size of target vessel, blocker and 3 of 21 (14.2%) for those who were not (p
severity of proximal target vessel stenosis, age, sex, 0.18). When analyzed according the type of calcium-
smoking status, diabetes, hypertension, hyperlipidemia, channel blocker prescribed at discharge, the incidence of
history of vascular disease, intraoperative use of papav- radial artery string sign was 21 of 292 (7.2%) for nifedi-
erine, perioperative use of alpha-adrenergic agents, cal- pine, 5 of 71 (7.0%) for diltiazem, 2 of 52 (3.9%) for
cium-channel blocker use at discharge, and aspirin use at amlodipine, and 0 of 4 (0%) for felodipine (overall p
discharge. All analyses were performed using StatView 0.78). The incidence of radial artery string sign among
or SAS version 8 (SAS Institute, Cary, North Carolina). patient compliant on calcium-channel blocker therapy

Results
Table 2. Target Vessels of Radial Artery String Signs
Five hundred and sixty-one patients were enrolled in this
Coronary Artery String Signs Total Percenta
13-center clinical trial, and angiography was performed
in 440 patients at a mean 10.9 4.3 months after surgery. Left circumflex artery 17 227 7.4%
In these 440 patents, 36 (8.2%) of the radial arteries were OM1 9 141 6.3%
completely occluded compared with 60 study saphenous OM2 6 54 11.1%
veins (13.6%, p 0.009; Table 1). The main results of this OM3 1 17 5.8%
have been published elsewhere [11]. Among study radial Intermediate branch 1 12 8.3%
and vein grafts, string signs were seen in 31 of 440 radial Other branchesb 0 3 0%
arteries and 4 of 440 saphenous vein grafts (p 0.001; Right coronary artery 14 213 6.5%
Table 1). Seventeen of the radial artery string grafts were PDA 9 126 7.1%
directed to the circumflex system whereas the other 14
RCA 4 78 5.1%
were used to bypass the right coronary artery territory
Posterolateral 1 6 16.6%
(Table 2). Two patients with a radial string sign had a
Other branchesc 0 3 0%
string sign of another graft, one saphenous vein graft and
one left internal thoracic artery. In the 31 cases with a
Number of radial string signs divided by number of radial grafts
b
radial artery string signs, 15 grafts (48%) showed Throm- anastomosed on each coronary area multiplied by 100%. Other
branches include one main circumflex artery, two posterolateral from
bolysis in Myocardial Infarction Study (TIMI) 1 flow, circumflex artery. c
Other branches include two acute marginal
three grafts (10%) showed TIMI 2 flow, and 13 grafts branches and one small branch of RCA.
(42%) showed TIMI 3 flow of the distal native coronary OM obtuse marginal; PDA posterior descending artery; RCA
artery (Table 1). main right coronary artery.
Ann Thorac Surg MIWA ET AL 115

CARDIOVASCULAR
2006;81:1129 RADIAL ARTERY STRING SIGN

Table 3. Predictors of Radial Artery String Signs

Univariate analysis predictors Radial Artery String Sign, n (%) p Value
a
Proximal stenosis of target coronary 0.0009
90% 11/271 (4.1%)
90% 20/169 (11.8%)
Perioperative use of alpha-adrenergic agonistsb 0.003
Alpha-adrenergic agonist, yes 5/16 (31.3%)
Alpha-adrenergic agonist, no 26/398 (6.1%)
Intraoperative use of papaverine 0.8
Papaverine, yes 29/407 (7.1%)
Papaverine, no 2/33 (6.1%)
Taking aspirin at discharge 0.9
Aspirin, yes 29/409 (7.1%)
Aspirin, no 2/31 (6.5%)
Taking calcium channel blocker (CCB) at discharge 0.18
CCB, yes 28/419 (6.7%)
CCB, no 3/21 (14.3%)
Taking lipid-lowering drugs 1.0
Lipid-lowering drugs, yes 22/309 (7.1%)
Lipid-lowering drugs, no 9/131 (6.9%)
History of peripheral vascular disease 0.48
Yes 3/32 (9.38%)
No 28/408 (6.86%)
Smoking 0.83
Yes 22/317 (6.94%)
No 9/123 (7.32%)
Multivariate analysis predictors, odds ratio (95% CI)
Less than 90% proximal stenosis of target coronary 3.1 (1.47.0)
Perioperative use of alpha-adrenergic agonists 8.2 (2.428.6)
a b
By study protocol, proximal stenosis of all target coronary artery was greater than 70%. alpha-adrenergic agonists include norepinephrine and
neosynephrine.

CI confidence interval.

for 6 months was 26 of 401 (6.5%) versus 5 of 39 (12.8%) stenosis, and 46.6% in patients with both risk factors accord-
for those who were not (p 0.2). ing to the risk-adjusted multivariate model. There was no
relationship between the use of other recommended med-
Predictors of Radial Artery String Signs ical therapies and the occurrence of radial artery string
The presence of a string sign in radial artery grafts was signs (Table 3).
closely related to the severity of the proximal stenosis of the
target coronary artery: 11.8% of the patients with less than Clinical Consequences of Radial Artery String Signs
90% proximal narrowing of the target coronary arteries had Four patients died perioperatively and a further 4 pa-
a radial artery string sign versus 4.1% of the patients with tients died between 30 days and 1 year. None of these
more than 90% proximal narrowing (adjusted odds ratio
[OR] 3.1, 95% confidence interval [CI]: 1.4 to 7.0; Table 3).
The radial artery string sign occurred in 31.3% of patients Table 4. Postoperative Symptoms
who received perioperative alpha-adrenergic agonist infu-
Patent Radial
sion, but only in 6.1% of patients without perioperative Radial Artery Artery
alpha-adrenergic agonists (adjusted OR 8.2, 95% CI: 2.4 to With String Without
28.6; Table 3). There was no meaningful difference in the Sign String Sign
predictors of radial artery string sign whether all patients (n 31) (n 373) p
undergoing angiography or only patients with patent radial n (%) n (%) Value
artery grafts were included in the regression model. The Myocardial infarction 3 (9.7%) 39 (10.4%) 0.9
predicted risk of radial string sign was 2.7% in patients Angina 3 monthsa 1 (3.2%) 16 (4.2%) 0.9
without either risk factor, 21.5% in patients receiving peri-
Angina 12 monthsa 4 (9.7%) 20 (5.4%) 0.12
operative alpha-adrenergic agents, 8.3% in patients with
radial arteries grafted to target vessels with less than 90% a
Angina means any postoperative symptoms suggestive of angina.
 116 MIWA ET AL Ann Thorac Surg
CARDIOVASCULAR

RADIAL ARTERY STRING SIGN 2006;81:1129

Table 5. Postoperative Symptoms in Radial String Sign
Patients
TIMI Scale Recurrent Angina Angina Free Percent
TIMI 1 4 11 26.6%a
TIMI 2 0 3 0%
TIMI 3 0 13 0%
a
p 0.0045.
patients had angiography; there was one postmortem
examination performed that showed both study grafts
patent. Myocardial infarction occurred in 3 of 31 patients
(9.7%) with a radial artery string sign and 39 of 373
patients (10.5%) with a patent radial artery graft (p 0.8).
All three perioperative myocardial infarctions in the
patients with radial artery strings occurred in the inferior
territory; 2 of those patients had radials to the circumflex
territory and 1 to the right coronary artery territory
(Tables 4 and 5). Postoperative symptoms of angina at 3
or 12 months were not associated with the presence or
absence of radial strings overall; however, TIMI flow was
an important determinant of angina at 12 months after
surgery in the 31 radial string patients (Tables 4 and 5).
Angina occurred in 4 of 15 radial artery string patients
with TIMI 1 flow, whereas none of the 16 patients with
TIMI 2 or 3 flow had postoperative angina (p 0.0045;
Tables 4 and 5).
Comment
In a large angiographic trial of radial artery grafting, we
identified 31 cases of radial artery string sign in 440
patients. Previous reports of the radial artery string sign
have been limited with respect to patient numbers and
study design (Table 6) [1118]. The cumulative percent-
age of radial artery strings in these earlier studies was,
however, very similar to the findings of this study. In
keeping with previous studies, string signs occurred
much more frequently in radial artery grafts rather than
saphenous vein grafts, which were typically completely
obstructed rather than diffusely narrowed.
Most cases of radial string signs involved the entire graft,
which we defined as a total string sign. Only 3 of 31 stringed
radial grafts had a partial string sign, and the distal portion
was narrowed in all of these three cases (Fig 1). We did not
include reactive focal narrowing of the radial arterial grafts
at the proximal anastomosis in the definition of radial
string, as these changes likely reflect catheter tip spasm and
do not represent any graft dysfunction. By protocol, all
study grafts were required to receive intragraft injections of
nitroglycerin before filming, to minimize graft spasm re-
lated to the procedure, and prevent identification of false
positive radial artery strings. In 9 patients with radial artery
string signs, radial artery angiograms were performed both
before and after intragraft nitroglycerin. Nitroglycerin di-
lated each of the 9 grafts, although the magnitude of
vasodilatation was minimal. It remains speculative whether
longer nitroglycerin infusions or direct administration of
drugs such as verapamil or adenosine may be associated
with further vasodilatation.
The Core Angiographic Committee identified radial
artery string signs for grafts that were diffusely narrowed
and estimated to be less than 1 mm in diameter. By
quantitative coronary analysis, all grafts were in fact less
than 1.05 mm, and less than 40% of the diameter of the
normal fully patent radial graft. The Core Angiography
Committee also determined TIMI flow. Patients with
radial string signs had various distal coronary flow pat-
terns, of whom nearly 50% had TIMI 3 flow. Patients with
radial artery string signs and TIMI 3 flow appeared less
likely to suffer recurrent angina. The numbers of patients
involved are small, but suggest that the flow reserve of
the TIMI 3 string radial artery grafts was adequate.
Multivariate analysis revealed two factors that were
closely related to the occurrence of postoperative radial
string signs. Patients with less than 90% proximal stenosis
in the target coronary artery by visual assessment had an
increased risk of radial string sign. The inclusion criteria for
the trial stipulated that the right coronary and circumflex
territory lesions needed to be greater than 70%, as we
aimed to place the radial grafts in high demand situations
with limited competitive flow from the native coronary
artery. Our study demonstrated that the threshold coronary
stenosis for the radial artery string sign is greater than 70%,
and that directing the radial artery to coronary vessels with
very high grade proximal lesions was associated with a
much lower incidence of radial artery string sign. We
previously identified this same relationship for radial graft
occlusion [8]. While the mechanism by which flow compe-

Table 6. Recent Studies of Radial Artery String Sign

Author [Reference] Year String Sign Radial Grafts Percent Follow-Up

Affonzo da Costa [12] 1996 1 61 1.64% 8.7 months

Manasse [13] 1996 2 55 3.64% 11.1 months
Chen [14] 1996 4 94 4.26% 11.6 weeks
Tatoulis [15] 1998 1 32 3.13% 4.2 months
Royse [16] 2000 14 137 10.22% 13.0 months
Merlo [17] 2003 1 1 6 months
Gaudino [18] 2004 9 128 7.03% 66 months
Desai [11] 2004 31 440 7.0% 10.9 months
Total 63 948 6.65%
Ann Thorac Surg
2006;81:1129
Fig 1. The left photo shows a complete string sign (arrows) with
diffuse narrowing along the full length of the graft. The right photo
shows a partial string sign (arrows) with segmental narrowing in-
volving the distal half of the graft, while the proximal half of this
graft is almost normal in diameter.
tition from the native vessels leads to graft narrowing is
unclear, it has been previously shown that reduced flow in
arterial conduits may lead to low shear stress, inducing
graft dysfunction [19, 20].
The second major risk factor was the use of alpha-
adrenergic agents in the early postoperative period. The
protocol recommended that these agents be avoided. Over-
all, alpha-adrenergic agents were only used in 25 of 561
enrolled patients, 16 of whom underwent angiography and
5 of whom had radial artery string sign. We are conse-
quently unable to accurately estimate the magnitude of risk
(CI: 2.4 to 28.6). Because radial arteries are known to have a
significantly increased force of contraction when exposed to
alpha-adrenergic agents versus other conduits [21, 22], it
seems prudent to minimize the exposure of patients to high
doses of these drugs after radial grafting.
To mitigate perioperative vasospasm. papaverine buff-
ered in lactated ringers was used for radial graft prepara-
tion in 92% of patients, but did not influence the occurrence
of string signs. Other perioperative pharmacologic treat-
ment described in the protocol included nitroglycerin infu-
sion in the intensive care unit and calcium-channel blockers
(usually nifedipine) starting on the first postoperative day.
Other methods of radial preparation have been described;
Buxton and coworkers [23] reported on the use of milri-
none, and Taggart and colleagues [24] described phenoxy-
benzamine. Alternative perioperative preparation of radial
grafts may reduce the incidence of strings or occluded
grafts. Corvera and colleagues [25] have recently shown
that the vasoconstrictive response of the radial artery to
alpha-adrenergic agents can be completely abolished using
phenoxybenzamine.
The optimization of calcium-channel blockade after ra-
dial artery grafting remains controversial. In this trial,
compliance with calcium-channel blocker therapy was
more than 90% at 6 months, and although radial artery
string sign tended to be more common in patients who
were not on calcium-channel blockers, our study was not
designed or powered to assess this relationship. Gaudino
and coworkers [26] have previously shown that calcium-
MIWA ET AL 117
CARDIOVASCULAR
RADIAL ARTERY STRING SIGN
channel blockade with diltiazem does not affect radial graft
patency or clinical and scintigraphic results. In vitro exper-
iments on human radial artery segments have shown a
superior relaxation and resistance to contractive stimuli
with the use of nifedipine or amlodipine, dihydropyridine
calcium-channel antagonists, rather than agents such as
diltiazem, a 1-5 benzothaizapine calcium-channel antago-
nist [27, 28]. Although the protocol specified nifedipine use,
several institutions preferentially discharged patients on a
regimen of diltiazem, and less commonly, amlodipine or
felodipine, allowing for comparison of the different cal-
cium-channel blocking agents. Our study found no differ-
ence in the incidence of radial graft string signs according to
the type of calcium-channel blocker used.
This study has certain limitations. Although this study
is larger than any of the previous reports, a sample of 31
string signs did restrict our ability to perform multiple
analyses. These are 1-year findings; midterm angio-
graphic follow-up of the study population 5 to 6 years
postoperatively is in progress. It is our intention to
restudy the string sign patients in addition to patients
with fully patent grafts.
In conclusion, radial artery string signs occurred in 7% of
patients in this randomized clinical trial. Risk factors for
radial artery string signs were perioperative alpha-
adrenergic agonists and grafting a coronary vessel with a
proximal lesion less than 90%. The presence of a radial
artery string sign was for the most part well tolerated
clinically; however, patients with radial artery string signs
with TIMI 1 flow were more likely to be symptomatic
postoperatively. Vigilant postoperative surveillance of ra-
dial grafts in patients requiring perioperative alpha-
adrenergic support is warranted as these grafts appear to be
at exceptionally high risk for string sign formation. Al-
though the type of calcium-channel blocker used had little
influence, use of calcium-channel blockers generally may
be associated with a reduced incidence of the angiographic
string sign.
This study was supported by Canadian Institutes of Health
Research Grants MT-13833 and MCT 52681.
References
1. Carpentier A, Guermonprez JL, Doloche A, et al. The aorta-
to-coronary radial bypass graft: a technique avoiding patho-
logical changes in grafts. Ann Thorac Surg 1973;16:11121.
2. Chiu CJ. Why do radial artery grafts for aortocoronary
bypass fail? A reapprasial. Ann Thorac Surg 1976;21:513 8.
3. Acar C, Jebara V, Portoghese M, et al. Revival of the radial
artery for coronary artery bypass grafting using the radial
artery. Ann Thorac Surg 1992;54:652 60.
4. Taggart DP, DAmico R, Altman DG. Effect of arterial revas-
cularization on survival: a systematic review of studies
comparing bilateral and single internal mammary arteries.
Lancet 2001,15;358:870-5.
5. Tatoulis J, Buxton BF, Fuller JA. Patencies of 2127 arterial to
coronary conduits over 15 years. Ann Thorac Surg 2004;77:
93101.
6. Bhan A, Gupta V, Choudhary SK, et al. Radial artery in
CABG: could the early results be comparable to internal
mammary artery graft? Ann Thorac Surg 1999;67:1631 6.
 118 MIWA ET AL
CARDIOVASCULAR
RADIAL ARTERY STRING SIGN
7. Fremes SE. Multicenter radial artery patency study (RAPS).
Study design. Control Clin Trials 2000;21:397 413.
8. Desai ND, Cohen EA, Naylor CD, Fremes SE. Randomized
comparison of radial-artery and saphenous-vein coronary
bypass grafts. N Engl J Med 2004;351:23029.
9. Fremes SE, Christakis GT, Del Rizzo DF, Musiani A, Mallidi H,
Goldman BS. The technique of radial artery bypass grafting
and early clinical results. J Card Surg 1995;10:537 44.
10. Johnson MR, Brayden GP, Ericksen EE, et al. Changes in
cross-sectional area of the coronary lumen in the six months
after angioplasty: a quantitative analysis of the variable
response to percutaneous transluminal angioplasty. Circu-
lation 1986;73:46775.
11. Scoblionko DP, Brown BG, Mitten S, et al. A new digital
electronic caliper for measurement of coronary arterial ste-
nosis: comparison with visual estimates and computer-
assisted measurements. Am J Cardiol 1984;53:689 93.
12. Affonzo da Costa FD, Affonso da Costa I, Poffo R, et al.
Myocardial revascularization with the radial artery: a clinical
and angiographic study. Ann Thorac Surg 1996;62:475 80.
13. Manasse E, Sperti G, Suma H, et al. Use of the radial artery
for myocardial revascularization. Ann Thorac Surg 1996;62:
1076 83.
14. Chen AH, Nakao T, Brodman RF, et al. Early postoperative
angiographic assessment of radial artery grafts used for
coronary artery bypass surgery. J Thorac Cardiovasc Surg
1996;111:1208 12.
15. Tatoulis J. Bilateral radial artery grafts in coronary recon-
struction: technique and early results in 261 patients. Ann
Thorac Surg 1998;66:714 20.
16. Royse AG, Royse CF, Tatoulis J, et al. Postoperative radial
artery angiography for coronary artery bypass surgery. Eur
J Cardiothorac Surg 2000;17:294 304.
17. Merlo M. Terzi A, Tespili M, Ferrazzi P. Reversal of radial
artery string sign at 6 months follow up. Eur J Cardiothorac
Surg 2003;23:432 4.
18. Gaudino M, Tondi P, Serricchio M, et al. Atherosclerotic
involvement of the radial artery in patients with coronary
artery disease and its relation with midterm radial artery
graft patency and endothelial function. J Thorac Cardiovasc
Surg 2003;126:1968 71.
19. Maniar HS, Sundt TM, Barner HB, et al. Effect of target
stenosis and location on radial artery graft patency. J Thorac
Cardiovasc Surg 2002;123:4552.
20. Shimizu T, Ito S, Kikuchi Y, et al. Arterial conduit shear
stress following bypass grafting for intermediate coronary
artery stenosis: a comparative study with saphenous vein
grafts. Eur J Cardiothorac Surg 2004;25:578 84.
21. Chardigny C, Jebara VA, Acar C, et al. Vasoreactivity of the
radial artery. Comparison with the internal mammary and
gastroepiploic arteries with implications for coronary artery
surgery. Circulation 1993;88(Suppl):II11527.
22. He GW, Yang CQ. Characteristics of adrenoceptors in the
human radial artery: clinical implications. J Thorac Cardio-
vasc Surg 1998;115:1136 41.
23. Buxton BF, Raman JS, Ruengsakulrach P, et al. Radial artery
patency and clinical outcomes: five-year interim results of a
randomized trial. J Thorac Cardiovasc Surg 2003;125:136371.
24. Taggart DP, Dipp M, Mussa S, Nye PC. Phenoxybenzamine
prevents spasm in radial artery conduits for coronary artery
bypass grafting. J Thorac Cardiovasc Surg 2000;120:8157.
25. Corvera JS, Morris CD, Budde JM, et al. Pretreatment with
phenoxybenzamine attenuates the radial arterys vasocon-
strictor response to alpha-adrenergic stimuli. J Thorac Car-
diovasc Surg 2003;126:1549 54.
26. Gaudino M, Glieca F, Luciani N, Alessandrini F, Possati G.
Clinical and angiographic effects of chronic calcium channel
blocker therapy continued beyond first postoperative year in
patients with radial artery grafts: results of a prospective
randomized investigation. Circulation 2001;104(Suppl 1):64 7.
27. He GW, Yang CQ. Comparative study on calcium channel
antagonists in the human radial artery: clinical implications.
J Thorac Cardiovasc Surg 2000;119:94 100.
Ann Thorac Surg
2006;81:1129
28. Bond BR, Zellner JL, Dorman BH, et al. Differential effects of
calcium channel antagonists in the amelioration of radial
artery vasospasm. Ann Thorac Surg 2000;69:1035 40.
Appendix
Radial Artery Patency Study Group
The members of the Radial Artery Patency Study Group are as
follows (all institutions are in Canada unless otherwise speci-
fied): Executive CommitteeS. E. Fremes, E. A. Cohen, C. D.
Naylor, N. D. Desai, R. Feder-Elituv; Manuscript Committee
N. D. Desai, E. A. Cohen, C. D. Naylor, S. E. Fremes; Steering
CommitteeS. E. Fremes, E. A. Cohen, C. D. Naylor, M. Carrier,
G. Cote, D. Doyle, O. Gleaton, R. Masters, L. Higginson, L.
Errett, K. Watson, S. Lichtenstein, R. Carere, M. L. Myers, D.
Almond; Participating CardiologistsD. Almond (Victoria Hos-
pital, London, Ontario), C. Buller (University of British Colum-
bia, Vancouver), F. Charbonneau (McGill University, Montreal),
E. A. Cohen (University of Toronto, Toronto), C. Constance
(McGill University, Montreal), G. Cote (Montreal Heart Insti-
tute, Montreal), J. Ducas (Health Sciences Centre, Winnipeg,
Manitoba), O. Gleeton (Hopital Laval, Sainte-Foy, Quebec), L.
Higginson (University of Ottawa Heart Institute, Ottawa), L.
Schwartz (University of Toronto, Toronto), W. Tymchak (Uni-
versity of Alberta Hospital, Edmonton), R. Watson (University of
Toronto, Toronto), G. Devlin (Waikato Hospital, Hamilton, New
Zealand); Data CommitteeN. D. Desai, H. R. Mallidi, R.
Feder-Elituv (all at University of Toronto, Toronto); Statisti-
ciansJ. P. Szalai, M. Katik, K. Sykora, A. Kiss (all at University
of Toronto, Toronto); Angiographic CommitteeE. A. Cohen, J.
Dubbin, S. Radhakrishnan, A. Adelman (deceased), L. Schwartz
(all at the University of Toronto, Toronto); Clinical End-Points
CommitteeZ. Sasson (University of Toronto, Toronto), P.
Dorian (University of Toronto, Toronto), K. Teoh (McMaster
University, Hamilton, Ontario); Electrocardiogram Commit-
teeG. Newton, Z. Wullfart, R. Myers, E. Crystal (all at the
University of Toronto, Toronto); Data and Safety Monitoring
CommitteeS. Brister, C. Morgan, S. Logan (all at the Univer-
sity of Toronto, Toronto); Investigators (the number of patients
recruited is in parentheses): Hopital Laval, Sainte-Foy, Quebec:
D. Doyle (2), D. Desaulniers (2), R. Baillot (1), G. Raymond (6), M.
Lemieux (6), P. Cartier (deceased) (2); Institute de Cardiologie de
Montreal, Montreal: R. Cartier (2), M. Carrier (6), Y. Leclerc (1);
London Health Sciences CenterUniversity Campus, London,
Ontario: A. Menkis (4), D. Boyd (24), R. Novick (2); London
Health Sciences CenterVictoria Campus, London, Ontario:
M. L. Myers (20); Montreal General Hospital, Montreal: D.
Shum-Tim (1), J. F. Morin (48); Sunnybrook and Womens College
Health Sciences Centre, Toronto: B. Goldman (14), C. Cutrara (32),
G. Bhatnagar (39), S. E. Fremes (108), G. T. Christakis (43), L.
Abouzhar (16); Health Sciences Centre, Winnipeg, Manitoba: D.
Del Rizzo (10); St. Michaels Hospital, Toronto: D. Bonneau (6), D.
Latter (23), L. Errett (11); Toronto General Hospital, Toronto: C.
Peniston (4), H. Scully (1), R. Weisel (22), R. J. Cusimano (1), S.
Brister (3), T. Ralph-Edwards (1), T. Yau (9); University of Alberta
Hospital, Edmonton: E. Gelfand (8), P. Penkoske (2); University of
Ottawa Heart Institute, Ottawa: F. Rubens (26); Vancouver Hospital
and Health Sciences Centre, Vancouver, B. C.: G. Fradet (25), L.
Burr (14), D. Thompson (2); Waikato Hospital, Hamilton, New
Zealand: R. Ullal (14); Site CoordinatorsM. Aleggretti, A. M.
Powel, H. Brochu, R. Feder-Elituv, R. Fox, L. Lepicq, G. Keuen, C.
Jessina, S. Finlay, E. Reeves, A. MacDonald, M. El-Tawil, L. Paul,
M. A. James, L. Verreault, B. Weller, C. Nacario, J. Wilson, D.
Penny, F. Denis, A. Munoz, L. Montebruno.
Ann Thorac Surg
2006;81:1129
DISCUSSION
DR BRIAN BUXTON (Melbourne, Victoria, Australia): Doctor
Miwa, thank you very much for an excellent presentation and I
commend you on the clarity of your results. In the same way that
you have had a problem with the radial artery string sign, our
dilemma is whether to regard this as a failure or success. A
stenosis greater than 90% is one factor in determining compet-
itive flow. Unfortunately, there is no simple method of assessing
competitive flow, and therefore, a high-grade native vessel
stenosis is only a surrogate of collateral or competitive flow and
is not always reliable. For instance, it is possible to have a total
occlusion of the circumflex with a large collateral and yet have a
highly competitive situation.
Regarding the vasodilators, I noticed they were only used in a
very small number of patients. Would you consider bending
your rule regarding the use of norepinephrine if the patient
presented with extreme vasodilatation, which can occur after
bypass surgery or after using a vasodilator?
DR MIWA: Thank you, Dr Buxton. About the vasoconstrictors,
from our protocol we defined vasoconstrictor agents as pure
alpha-adrenergic agonists, that means norepinephrine or neo-
synephrine. Sometimes we have to use these vasoconstrictors,
but if we can avoid it, for example, by using an intra-aortic
balloon pump or other methods to support the patients condi-
tion, we do. Of course, sometimes we have to use these vaso-
constrictors, but this paper strongly shows the relationship
between vasoconstrictors and radial artery string sign.
So if we have to use vasoconstrictors in the perioperative
period, the most important thing is that we pay special attention
to the medical care of the patient after surgery and we have to
follow them more closely. I think that is the main suggestion
from these results.
DR BUXTON: I have one follow-up question regarding the
string sign. What do you think will happen at the 5-year
angiography? Do you think the flow reserve, the lumen diame-
ter, and the patency will increase, or is it too early to tell?
DR MIWA: Can you ask again?
DR BUXTON: I will ask Dr Desai to comment on that.
DR TOM ORSZULAK (Rochester, MN): I enjoyed your talk. I
think radial arteries have a place in coronary revascularization.
At our institution, based on published laboratory data by Dave
Cable, we found that nitrates alone prevented a lot of radial
artery spasm. You seem to avoid using nitrates in people that
have radial artery grafts. We have generally put everyone on
nitrates for a short term. I am not sure I know the answer to how
long nitrates should be given in radial arteries, but have you
considered using nitrates for some period of time to prevent
some of the spasm or string sign?
DR MIWA: In this study we strongly recommended the use of
nitrates after surgery. I do not have any convincing evidence
regarding how long we have to use nitrates after surgery.
DR FRANK SELLKE (Boston, MA): But you put everybody on
some agent, a calcium-channel blocker or a nitrate?
MIWA ET AL 119
CARDIOVASCULAR
RADIAL ARTERY STRING SIGN
DR MIWA: Yes. In this study more than 90% of the patients
received a calcium-channel blocker and, of course, aspirin upon
discharge. Regarding the calcium-channel blockers, more than
90% of the patients took a calcium-channel blocker for at least 3
to 6 months.
DR SULAIMAN HASAN (Charleston, WV): I want to congratulate
you on your wonderful study and very well presented. My question
is, do you consider discontinuing things like angiotensin-
converting enzyme (ACE) inhibitors 1 or 2 days before so that you
might avoid the use of vasoconstrictors postoperatively?
DR MIWA: In our results, ACE inhibitors, beta blockers, or
other antithrombotic medicines did not have any significant
relationship with the formation of radial artery string sign. So
our data do not show that ACE inhibitor use really affected the
incidence of radial artery string sign.
DR HASAN: My question was that if you were to stop an ACE
inhibitor 1 or 2 days before, your chance of having to use, say,
Levophed or norepinephrine might be less after the operation.
Do you do that or not? Do you stop the ACE inhibitors?
DR MIWA: In this multicenter study, we did not have any data
regarding preoperative cessation of the ACE inhibitors or other
ways to avoid vasoconstrictors. It was not in the protocol. Our
clinical practice, generally, is to stop ACE inhibitors the night
before the operation.
DR DESAI: I am one of Dr Miwas coauthors. I can address a
couple of these questions. For Dr Buxtons question about the
long-term survival of the string radial grafts, it is our impression
that the grafts that have TIMI I flow will likely occlude over time.
Those with good flow reserve are less likely to do so.
Regarding Dr Orszulaks and Dr Sellkes questions about
nitrates, as stipulated in the protocol, all of our patients received
intravenous nitroglycerin infusion for the first night postopera-
tively, provided the patients hemodynamics could tolerate it.
Use of oral nitrates was not required or common after surgery.
DR ALISTAIR G. ROYSE (Melbourne, Australia): I have a
10-year experience with the radial artery, and our published and
to-be published data has 90 to 92% patency, which is the same as
yours. I would use vasoconstrictors in at least 60% of my
patients. I remind you that the duration of action of these
vasoconstrictors such as noradrenaline or metaraminol is mea-
sured in minutesnot months or years. The notion that a
transient use of a short-term acting agent causes permanent
irreversible spasm is not logical and not supported by pharma-
cological evidence. I am unconvinced that there is sufficient
robustness in your data to make the association between peri-
operative use of vasoconstrictors and graft failure so confidently.
DR MIWA: This study included follow-up of both asymptomatic
patients and symptomatic patients. From these data, we found a
strong relationship between vasoconstrictors, that is, alpha-
adrenergic agonists, and the radial artery string sign. I think this
makes biological sense and that the relationship between vaso-
constrictors and radial string sign is indeed real.