AJR 2002; 178:945-948
© American Roentgen Ray Society
A Closer Look at the Midsternal Stripe Sign
Phillip M. Boiselle1 and
Alberto V. Mansilla2,3
1
Department of Radiology, Harvard Medical School and Beth Israel Deaconess
Medical Center, 330 Brookline Ave., Boston, MA 02215.
2
Department of Radiology, Temple University Hospital, 3401 N. Broad St,
Philadelphia, PA 19106.
3
Present address: Department of Radiology, Florida Hospital, 601 E. Rawlings
St., Orlando, FL 32806.
Received July 12, 2001;
accepted after revision October 10, 2001.
Presented at the annual meeting of the American Roentgen Ray Society,
Seattle, April-May 2001.
Address correspondence to P. M. Boiselle.
Abstract
OBJECTIVE. Reports in the literature offer conflicting data on the
value of the midsternal stripe sign in diagnosing sternal dehiscence. Our
purpose was to determine the frequency with which the midsternal stripe sign
is present in patients with dehiscence compared with a control group without
this complication and to determine whether this sign adds incremental value to
the sign of sternal wire displacement in the diagnosis of dehiscence.
CONCLUSION. A midsternal stripe thicker than 3 mm should raise one's
suspicion of the presence of sternal dehiscence. However, this sign is rarely
observed in patients with this complication and does not add incremental value
to the finding of sternal wire displacement in establishing the diagnosis of
dehiscence.
Introduction
Sternal dehiscence is an uncommon but serious complication of median
sternotomy
[1,2,3,4,5,6,7].
The interest in establishing the radiographic diagnosis of this complication
is long-standing
[8,9,10].
Two major radiographic signs have been described: the midsternal stripe sign
(a midline vertical radiolucency overlying the sternum)
[8,9,10]
and sternal wire displacement
[10] (one or more wires
misaligned with respect to others in the vertical row). Although the latter
sign has been shown to be highly specific for dehiscence
[5,6,7],
reports in the literature offer conflicting assessments regarding the
relevance of the midsternal stripe sign in the diagnosis of dehiscence
[8,9,10].
The purpose of this study was threefold: first, to compare the frequency of
the midsternal stripe sign identified in a large group of patients with proven
dehiscence with that identified in a control group; second, to identify a
threshold value of stripe thickness that could be used to distinguish between
patients with and without dehiscence; and third, to determine whether the
midsternal stripe sign adds any incremental value to the radiographic
diagnosis of dehiscence using the alternative sign of sternal wire
displacement.
Materials and Methods
We used our computerized hospital information system to identify all
patients from January 1993 to April 1999 who had a discharge diagnosis of
sternal dehiscence. We performed a retrospective chart review for each patient
and extracted data concerning the date of the median sternotomy, type of
surgical procedure, and date of diagnosis of sternal dehiscence. On reviewing
the medical records, we included all patients with surgically proven
dehiscence and excluded five patients because the computerized coding of their
discharge diagnosis did not match the written discharge diagnosis in the
medical record (i.e., these patients did not have a true discharge diagnosis
of dehiscence). We subsequently procured all postoperative radiographs for the
patients in the study cohort. Five patients were excluded because their
postoperative chest radiographs were unavailable for retrospective review.
For each patient in the study cohort, a chest radiograph from the date of
diagnosis of sternal dehiscence was compared with the initial postoperative
radiograph obtained after median sternotomy (range of time interval, 3-69
days; mean time interval, 14 days). Each radiograph was retrospectively
reviewed by two radiologists, and findings were determined by consensus.
Radiographs were assessed for the presence of a midsternal stripe, which was
defined as a midline vertical radiolucency overlying the sternum
[8], and sternal wire
displacement, which was defined as one or more wires misaligned in relation to
others in the vertical row [5]
(Fig. 1). To assess for
displacement, we drew a vertical line through the center of the sutures on the
first postoperative radiograph and then used that line as a reference to draw
a vertical line on the subsequent postoperative radiographs. Displacement was
considered present when a change in alignment of more than 3 mm involving one
or more wires was found between the initial and subsequent postoperative
radiographs. To determine the first date that sternal wire displacement was
radiographically apparent, we reviewed additional chest radiographs obtained
during the time interval between the first postoperative radiograph and the
date of diagnosis of dehiscence.
View larger version (16K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1. —Line drawing shows normal appearance of postoperative sternum
and radiographic signs of dehiscence, including displacement (displaced wires
are highlighted in bold print) and midsternal stripe (wires have been removed
from drawing for simplification). It has been proposed that mechanism of
dehiscence involves sternal wires pulling or cutting through sternum rather
than breaking. As sternum separates, some wires (bold print) travel with right
side of sternum whereas others migrate with left side of sternum. Cleft
between two sternal fragments is rarely visualized radiographically as midline
radiolucency or stripe. (Illustration adapted with permission from
[5])
|
|
In images in which a midsternal stripe sign was visualized, we measured the
maximal transverse width of the stripe using a ruler and then recorded values
to the nearest millimeter. Measurements were obtained by consensus agreement
by the two radiologists. We also reviewed serial radiographs between the date
of surgery and the date of diagnosis of dehiscence to determine the first date
that the stripe was present. If the stripe was present on earlier radiographs,
we obtained serial measurements of the stripe to allow us to assess whether
there had been progressive widening.
We retrospectively reviewed a series of postoperative radiographs for a
group of control patients, composed of consecutive median sternotomy patients
with no clinical evidence of sternal dehiscence or mediastinitis. Analysis of
serial postoperative radiographs (mean time interval from preoperative
radiograph to last postoperative radiograph, 8 days) in this group was
performed in the same fashion as for the study cohort. We did not attempt to
have blinded reviewers because of obvious differences in the number of
radiographs in the two groups (the control group had a shorter hospital stay
and fewer radiographs). Statistical analysis was performed using Fisher's
exact test [11]. A p
value of less than 0.05 was considered statistically significant.
Results
The study cohort was composed of 35 patients with proven sternal
dehiscence, including 25 men and 10 women, who ranged in age from 43 to 85
years (mean, 65 years). The patients underwent median sternotomy for coronary
artery bypass graft surgery (n = 24), valvular surgery (n =
3), combined coronary artery bypass graft and valvular surgery (n =
3), aortic dissection repair (n = 2), and cardiac transplantation
(n = 3). The control group was composed of 50 patients, including 29
men and 21 women, who ranged in age from 41 to 83 years (mean, 65 years).
These patients underwent median sternotomy for coronary artery bypass graft
surgery (n = 45), valvular surgery (n = 4), and heart
transplantation (n = 1).
A midsternal stripe (Figs. 2
and 3) was identified in seven
(20%) of 35 patients with sternal dehiscence and in two (4%) of 50 control
group patients. This difference was not statistically significant. The mean
width of the stripe was 15 mm in diameter (median, 4 mm; range, 2-37 mm) for
the study cohort and 2 mm for the control group. The stripe measured less than
3 mm in diameter in both of the control group patients and in three of the
study cohort patients. There were no cases in which the midsternal stripe
became progressively wider on serial chest radiographs.
View larger version (118K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2. —Midsternal stripe sign is visualized in 61-year-old man after
median sternotomy and coronary artery bypass surgery. Coned-down radiograph of
mediastinum from portable chest radiograph reveals midsternal radiolucency
(paired arrows) measuring slightly less than 3 mm at greatest width.
Sternal wires show normal alignment in vertical row with no significant
displacement. (This radiograph of patient in study cohort was not one of
images reviewed for study because it was obtained after sternal rewiring for
dehiscence; however, it was selected for illustrative purposes because of
highly visible stripe and sutures.)
|
|
View larger version (95K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3. —Midsternal stripe sign is visualized in 67-year-old man who
developed sternal dehiscence after median sternotomy for repair of type I
aortic dissection. Coned-down image of lower mediastinum from portable chest
radiograph shows wide midsternal stripe (paired black arrows)
corresponding to cleft between two sternal fragments. Note rightward
displacement of lower sternal wires (right lateral margin demarcated by
open black arrows) with respect to more proximal wire (right lateral
margin demarcated by closed white arrow).
|
|
Sternal wire displacement (Fig.
4) was observed in 26 (74%) of 35 patients with sternal dehiscence
but was not identified in any of the control group patients (p <
0.0001). Sternal wire displacement was evident in six (86%) of seven patients
with sternal dehiscence whose radiographs displayed a midsternal stripe. In
the one case in which a stripe was not accompanied by sternal wire
displacement, the stripe measured only 2 mm in diameter.
View larger version (94K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4. —Sternal wire displacement is revealed in 66-year-old man who
developed sternal dehiscence after coronary artery bypass surgery. Portable
chest radiograph shows marked rightward displacement of first three sternal
wires with respect to lower three sternal wires. Wide radiolucent cleft (right
margin demarcated by arrows, left margin extends to midline adjacent
to lower sternal wires) is visible between sternal fragments; stripe projects
to right of midline because of rightward patient rotation (note asymmetry of
clavicles). Sternal wires have been highlighted with black marker to improve
visualization.
|
|
Sternal wire abnormalities preceded the clinical diagnosis in 16 (46%) of
35 patients. The midsternal stripe sign preceded the clinical diagnosis of
dehiscence in four (11%) of 35 patients but was accompanied by sternal wire
displacement in all these patients. In three of the four patients, sternal
wire abnormalities preceded visualization of the midsternal stripe by 3 days.
In these three patients, the stripe was visible 1 day before the clinical
diagnosis was made and sternal wire displacement was visible 4 days before the
clinical diagnosis. In the remaining patient, sternal wire displacement and
the midsternal stripe sign were first radiographically apparent at the same
time.
Discussion
Sternal dehiscence is an uncommon but serious complication of median
sternotomy, with an estimated prevalence of 1-2%
[1,2,3,4].
Predisposing factors include chronic obstructive pulmonary disease, obesity,
diabetes mellitus, prolonged bypass time during surgery, internal mammary
artery grafting, repeated surgery for mediastinal bleeding, and prolonged
postoperative ventilation [1].
Prompt diagnosis is important for successful management
[1,2,3,4,5,6,7].
Although the diagnosis is often evident from the physical examination
findings, sternal dehiscence may be clinically occult in some patients
[1,2,3,4,5,6,7].
Thus, there has been a longstanding interest in determining whether
radiography could be used in identifying this complication
[8,9,10].
Two major radiographic signs have been described: the midsternal stripe sign
and sternal wire displacement. Our results show that visualization of the
midsternal stripe sign does not contribute to the diagnosis of this
complication. In contrast, sternal wire displacement is a very valuable
radiographic sign of dehiscence and has the potential to play an important
role in the diagnosis of this complication.
The midsternal stripe sign was first described by Escovitz et al.
[8], who detected a midsternal
radiolucency in 12 of 250 postoperative patients during a 2-year period and
found that sternal dehiscence developed in four of these patients. A review by
the same authors of the cases of 100 consecutive patients who had undergone
median sternotomy revealed that sternal dehiscence did not develop in any
patient who did not have a midsternal stripe. These authors concluded that the
midsternal stripe sign may herald sternal dehiscence. Berkow and Demos
[9] refuted this assertion on
the basis of their identification of a midsternal radiolucency in 52 (30%) of
173 patients who underwent median sternotomy, only two of whom developed
dehiscence. Both studies were limited by a small number of patients with
dehiscence (n = 4), which precluded determination of whether a
threshold value of stripe thickness might prove valuable for differentiating
patients with dehiscence from those without this condition. Our study of 35
patients with sternal dehiscence provides the first data regarding the stripe
sign in a large group of patients with this complication.
The relatively low prevalence (4%) of the midsternal stripe sign in the
control group in our series is similar to the results reported by Escovitz et
al. [8] but is significantly
lower than the prevalences reported in other published series
[4,
9,
12]
(Table 1). The reason for this
difference is unclear but may reflect variations in radiographic technique
[10] among various
institutions or in observer thresholds for the detection of thin stripes.
In recognition of the fact that a thin stripe has been reported as a normal
postoperative finding, we sought to identify a threshold value that could be
used to differentiate patients with dehiscence from those without dehiscence.
In this study and other published series
(Table 1), a stripe thicker
than 3 mm was observed only rarely in patients in the control groups. Both
exceptions to this finding came from the series by Escovitz et al.
[8], in which two patients with
stripes of 4 mm and 6 mm, respectively, did not develop dehiscence. Thus, a
stripe thicker than 3 mm is usually an abnormal finding and a stripe thicker
than 6 mm is definitively so. Applying the strict criterion of 6 mm to our
study population, radiographs of less than 10% of patients with dehiscence
would have revealed a stripe sign of meaningful size, and all of these
patients would also have shown sternal wire displacement.
It has been suggested that progressive widening of a midsternal stripe is a
better indication of sternal dehiscence than the mere presence of a stripe
sign [2,
3], although the prevalence of
this finding in a group of patients with proven dehiscence has not been
reported. It makes intuitive sense that progressive widening of the stripe
should correlate with increasing degrees of sternal separation. Because we did
not observe any patients with progressive widening on serial radiographs, we
are unable to directly address this issue. Our data suggest that progressive
widening is an uncommon radiographic manifestation of dehiscence, and thus, it
is possible that our sample size of 35 patients with dehiscence may not have
been sufficient for studying this variation of the midsternal stripe sign.
However, we emphasize that our study cohort with proven dehiscence is by far
the largest group to date in which this sign has been studied. A contributing
factor to the lack of observance of this finding in our patients may be the
fact that, in most patients, the stripe sign was first observed
radiographically near the time of clinical diagnosis. Had the clinical
diagnosis been further delayed, it is possible that a progressively widening
stripe may have been visualized in the setting of worsening sternal
separation.
In contrast to the stripe sign, sternal wire displacement is a highly
valuable sign for detecting dehiscence. The high frequency of sternal wire
displacement reported in association with dehiscence fits well with the
proposed mechanism of this complication, in which sternal sutures are thought
to pull or cut through the sternum rather than breaking
[3,4,5,
10]. As the sternum separates,
some sutures travel with the right side of the sternum and others migrate to
the left side, producing a radiographic appearance that has been referred to
as "wandering wires"
[5] (Figs.
1,
3, and
4). Because the proposed
mechanism of dehiscence involves wires pulling or cutting through the sternum
rather than breaking, sternal wire fracture is not thought to play any major
role in the development of dehiscence
[3,4,5].
In a study that assessed the frequency of various sternal wire abnormalities
in patients with dehiscence, Boiselle et al.
[5] reported sternal wire
fracture in only a small number of patients with sternal dehiscence, all of
whom also showed other abnormalities such as sternal wire displacement. Thus,
the identification of a sternal wire fracture should prompt one to look
carefully for sternal wire displacement.
In summary, the midsternal stripe sign is an infrequent finding among
patients with dehiscence and does not add incremental value to the finding of
sternal wire displacement in establishing the diagnosis of this complication.
When a midline radiolucent stripe wider than 3 mm is identified in the
postoperative setting, one should carefully assess the sternal wires for
evidence of displacement; displaced wires usually accompany the stripe sign in
the setting of dehiscence.
Acknowledgments
We thank Charles White for his thoughtful review of this manuscript,
Michael Larson for assistance with photography, Max Rosen for assistance with
statistical analysis, and Patricia Wyatt for administrative assistance.
References
-
Bryan AJ, Lamarra M, Angelini GD, et al. Median sternotomy wound
dehiscence: a retrospective case control study of risk factors and outcome.
J R Coll Surg Edinb
1992;37:305
-308[Medline]
-
Fraser RS, Müller NL, Colman N, Paré PD, eds.
Complications of therapeutic, biopsy, and monitoring procedures. In:
Fraser and Paré's diagnosis of diseases of the
chest, 4th ed. Philadelphia: Sauders, 2001:2660
-2661
-
Thorsen MK, Goodman LR. Extracardiac complications of cardiac
surgery. Semin Roentgenol
1988;23:32
-48[Medline]
-
Carter AR, Sostman HD, Curtis AM, Swett HA. Thoracic complications
after cardiac surgery. AJR
1983;140:475
-481[Abstract/Free Full Text]
-
Boiselle PM, Mansilla AV, Fisher MS, McLoud TC. Wandering wires:
frequency of sternal wire abnormalities in patients with sternal dehiscence.
AJR
1999;173:777
-780[Abstract/Free Full Text]
-
Hayward RH, Knight WL, Baisden CE, Reiter CG. Sternal dehiscence:
early detection by radiography. J Thorac Cardiovasc
Surg 1994;108:616
-619[Abstract/Free Full Text]
-
Boiselle PM, Mansilla AV, White CS, Fisher MS. Sternal dehiscence
in patients with and without mediastinitis. J Thorac
Imaging 2001;16:106
-110[Medline]
-
Escovitz ES, Okulski TA, Lapayowker MS. The midsternal stripe: a
sign of dehiscence following median sternotomy.
Radiology
1976;121:521
-524[Abstract]
-
Berkow AE, Demos TC. The midsternal stripe and its relationship to
postoperative dehiscence. Radiology
1976; 121:525
-
Ziter FMH. Major thoracic dehiscence: radiographic considerations.
Radiology
1977;122:587
-590[Abstract]
-
Dawson-Saunders B, Trapp RG. Basic & clinical
biostatistics, 2nd ed. Norwalk, CT: Appleton & Lange,
1994: 152
-
Katzberg RW, Whitehouse GH, deWeese JA. The early radiographic
findings after cardiopulmonary bypass surgery. Cardiovasc
Radiol 1978;1:205
-215[Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
