Keywords
deep sternal wound infection (DSWI) - negative pressure wound therapy - omental flap
- pectoralis major muscle flaps - sternal dehiscence
Background
Deep sternal wound infection (DSWI) following median sternotomy is a life-threatening
complication after this approach for coronary artery bypass grafting (CABG).[1]
[2] Most of the patients who undergo CABG have associated morbidities such as diabetes
mellitus resulting in delayed wound healing. The incidence of sternal wound infections
and dehiscence is reported between 0.2 and 10% among patients who undergo sternotomy
and mortality of 5 to 20% following sternal dehiscence.[3] The use of bilateral internal mammary arteries (BIMA) for revascularization has
been shown to improve graft patency rates and help in reduction of repeated episodes
of angina in comparison to the usage of a single internal mammary artery (IMA).[4]
[5]
[6] The extensive use of BIMAs precludes a higher incidence of DSWI associated with
sternal devascularization.[7]
[8] This is of concern in patients with predisposing factors such as diabetes, obesity,
chronic pulmonary obstructive disease, or diffuse arteriopathy.[9]
[10]
[11]
[12] When DSWI develops, the cardiac surgeons use conventional methods such as wound
debridement, rewiring or plating, and closure with or without negative pressure wound
therapy (NPWT). In contrast, when these patients are referred to plastic surgeons
they are treated with debridement and closure with various flaps.[13] A recent review article by Schiraldi et al advocated flap cover in cases of chronic
DSWI should not be delayed for more than 2 weeks after NPWT.[14] In this study, we aim to review our cases of DSWI with respect to their time of
referral to a plastic surgical unit and study the correlation between the time of
referral to the outcomes.
Aims and Objectives
-
To study the outcome of plastic surgical procedures in cases of DSWI in patients of
post-CABG and their correlation to time of referral to plastic surgical unit.
-
To find the correlation between the early versus late referral to a plastic surgical
unit with the outcome of plastic surgical procedures.
-
To evaluate the indicators of morbidity such as prolonged treatment duration and prolonged
hospital stay associated with the delay in referral to plastic surgery unit.
Null Hypothesis
There is no difference between early referral and late referral in terms of decreased
postoperative stay, morbidity, and mortality.
Patients and Methods
A retrospective cohort study of all patients of DSWI post-median sternotomy who underwent
CABG between January 2005 and June 2018 was performed at a tertiary care center in
India. The data was obtained from the electronic medical records of the patients from
the hospital database that included demographic data, dates of CABG, dates of presentation
of DSWI, and referral to the plastic surgical unit. We defined less than or equal
to 15 days from the presentation as an early referral and more than 15 days as a late
referral. This was an arbitrary value as there is no mention of cutoff value for time
of referral in the literature. Both the groups were compared with respect to the age,
gender distribution, classification as per Jones’ criteria,[15] delay in wound healing, presence of microorganisms in wound swab culture, number
of NPWT changes, duration between presentation to definitive management, duration
between definitive surgery and discharge, choice of a flap, number of flaps, duration
from plastic surgery referral to discharge, early and late postoperative course, and
postoperative complications and mortality. Institutional Ethics Committee permission
was obtained before commencing the study. (IHR_2018_SEP_PS_278)
Wounds were classified as per the Jones’ classification ([Table 1]).
Table 1
Jones and Jurkiewicz et al classification of sternal wound infections
|
Type
|
Depth
|
Description
|
Total cases
|
Early referrals
|
Late referrals
|
|
1A
|
Superficial
|
Skin and subcutaneous tissue dehiscence
|
Not included in the study
|
Not included in the study
|
|
1B
|
Superficial
|
Exposure of sutured deep fascia
|
|
2A
|
Deep
|
Exposed bone, stable wired sternotomy
|
72 cases
|
33 cases
|
39 Cases
|
|
2B
|
Deep
|
Exposed bone, unstable wired sternotomy
|
|
3A
|
Deep
|
Exposed necrotic or fractured bone, unstable, heart exposed
|
|
3B
|
Deep
|
Types 2 or 3 with septicemia
|
Inclusion criteria: All patients who underwent CABG through a median sternotomy approach, who had DSWI
(types 2A, 2B, 3A, 3B), and who underwent a plastic surgery procedure.
Exclusion criteria: Patients having superficial sternal wound infections (types 1A and 1B), or who underwent
any other cardiac procedure other than CABG, and patients lost to follow-up were not
included in the study.
Procedure
All patients with DSWI were assessed with a three-dimensional computed tomography
of the thorax, the microbiological culture of the wounds, and routine hematological
evaluation and C-reactive protein levels. All patients underwent radical debridement
under general anesthesia and specimen of bone and tissue were sent for aerobic, anaerobic,
and fungal cultures. Any hardware used for sternal stabilization was removed. Nonadherent
dressing with paraffin gauze was applied if the heart was exposed[16]
[17] followed by NPWT. The set pressure of –75 mm Hg was used when the cardia was not
exposed, and –50 mm Hg in exposed cardia.[2] The patients underwent repeat debridement under general anesthesia if deemed necessary
during the change of NPWT foam.
Definitive closure was done after all visible necrotic tissue or slough was cleared.
In certain cases, the sternum was stable with approximable edges, in which case rewiring
was done. As per the wound size, it was decided to do a unilateral pectoralis major
muscle advancement flap, bilateral pectoralis major muscle advancement flaps, omental
flap, or a combination of both flaps. The pectoralis major muscle advancement flap
was done by the technique described previously by the authors.[18] If the defect was large, the omental flap was used as well. After the patient was
referred to the plastic surgery unit, our approach was aggressive in terms of radical
debridement and early flap cover preferably after one to two sessions of NPWT. The
average time between referral to flap was 8 days (range: 5–14).
Statistical Analysis
Data was analyzed using the R statistical software. Both the groups were compared
with respect to the demographics. Quantitative data was expressed in median and interquartile
range (minimum–maximum). Fisher's exact test was applied to find the association.
The p-value < 0.05 was considered as statistically significant.
Results
The data of 72 patients who had DSWI out of 4,214 patients who underwent median sternotomy
for CABG was collected. Out of them, 68 patients underwent definitive plastic surgery
management with flap cover. Two patients died before the definitive closure. Fifty-eight
patients required multistage debridement and delayed closure, while only 12 patients
could undergo single-stage debridement and immediate closure. We had 59 out of 72
patients, wherein BIMA was used, and 85% (n = 50) of these were diabetic. The age distribution was comparable in both the early
and late groups with a slight male preponderance in the late group. As per the Jones’
classification type 2B was the commonest and the distribution of severity of cases
was found to be similar in both the groups. No association was found between the presence
or absence of diabetes as a preexisting comorbidity (p-value = 0.169). No association was found between the usage of BIMA during CABG (p-value = 0.759).
The delay in wound healing was defined as the inability of the wound to heal completely
by the end of 2 weeks from the time of definitive closure. The patients who could
not undergo definitive closure or who died soon after closure were excluded only from
the analysis of delayed wound healing (n = 6). Out of the total patients who were referred early, 41% patients had delayed
wound healing, and in patients who were referred late, 27% patients had delayed wound
healing. There was no difference in outcome of early versus late referral ([Table 2]). Both the early and late referral groups were compared as per the following variables
such as presence of microorganisms, duration from definitive surgery to discharge,
postoperative complications, number of NPWT changes before definitive closure, mortality,
presentation to definitive surgery, duration from plastic referral to discharge, number
of flaps, presence of microorganisms in bone culture, postoperative shoulder movement,
postoperative scar, and the number of cardiothoracic procedures before plastic surgery
referral ([Table 3]). But time of referral has affected the total duration of wound healing as average
duration of treatment of DSWI in early and late groups were 16.58 and 89.36 days,
respectively (excluding the earlier outlier 5,088 days). The number of procedures
done by the cardiothoracic team (resuturing, rewiring, and local debridement with
NPWT in both the groups before referral to the plastic surgery unit) is significant
(p = 0.002) and has delayed overall duration of wound healing. The rest of the variables
that were compared in both the groups did not show significant differences ([Table 3]). In our series, 83% patients reported unrestricted shoulder movement and 85% reported
cosmetically acceptable scars. Statistically significant difference was not found
in the two groups with respect to the mortality and postoperative complications. However,
the odds ratio of 1.39 shows that late referral leads to 30% increased chances of
prolonged hospital stay (> 7 days), though statistically not significant ([Table 4]).
Table 2
Correlation of time of referral to the delay in wound healing
|
Overall
n = 66
|
Early
n = 29
|
Late
n = 37
|
p-Value
0.29
|
|
Delay in wound healing
|
22 (33%)
|
12 (41%)
|
10 (27%)
|
|
Table 3
Effect of early versus late referral on various variables
|
Variable
|
Overall
n = 72
|
Early
n = 33
|
Late
n = 39
|
p-Value
|
|
Abbreviations: CVTS, cardiovascular and thoracic surgery; IQR, interquartile range;
NPWT, negative pressure wound therapy.
|
|
Presence of organisms
|
|
|
|
0.19
|
|
None
|
15 (21%)
|
10 (30%)
|
5 (13%)
|
|
One
|
43 (60%)
|
17 (52%)
|
26 (67%)
|
|
Multiple
|
14 (19%)
|
6 (18%)
|
8 (21%)
|
|
Number of NPWT changes
|
|
|
|
0.33
|
|
0
|
16 (23%)
|
8 (24%)
|
8 (21%)
|
|
1
|
25 (35%)
|
12 (36%)
|
13 (34%)
|
|
2
|
23 (32%)
|
8 (24%)
|
15 (39%)
|
|
3
|
5 (7%)
|
4 (12%)
|
1 (3%)
|
|
4
|
0
|
0
|
0
|
|
5
|
1 (1%)
|
0
|
1 (3%)
|
|
6
|
1 (1%)
|
1 (3%)
|
0
|
|
Bone C/S
|
|
|
|
0.63
|
|
Bone culture: Negative
|
43 (60%)
|
21 (64%)
|
22 (56%)
|
|
Bone culture: Positive
|
29 (40%)
|
12 (36%)
|
17 (44%)
|
|
Number of flaps
|
|
|
|
> 0.95
|
|
Single
|
67 (96%)
|
30 (97%)
|
37 (95%)
|
|
Multiple
|
3 (4%)
|
1 (3%)
|
2 (5%)
|
|
Post-op complications
|
|
|
|
0.15
|
|
None
|
44 (61%)
|
17 (52%)
|
27 (69%)
|
|
Any complication
|
28 (39%)
|
16 (48%)
|
12 (31%)
|
|
Mortality
|
5 (7%)
|
3 (9%)
|
2 (5%)
|
0.66
|
|
Shoulder movement
|
|
|
|
0.64
|
|
Good
|
60 (83%)
|
26 (79%)
|
34 (87%)
|
|
Restricted
|
6 (8%)
|
3 (9%)
|
3 (8%)
|
|
Death before wound healing/before closure
|
6 (8%)
|
4 (12%)
|
2 (5%)
|
|
Scar problems
|
|
|
|
0.49
|
|
No
|
61 (85%)
|
26 (79%)
|
35 (90%)
|
|
Yes
|
5 (7%)
|
3 (9%)
|
2 (5%)
|
|
Death before wound healing/before closure
|
6 (8%)
|
4 (12%)
|
2 (5%)
|
|
Definitive surgery to discharge (days)
Median (IQR)
|
8 (7–13)
|
8 (6–14)
|
9 (7–13)
|
0.14
|
|
Presentation to definitive surgery (days)
Median (IQR)
|
8 (5–14)
|
9 (4–15)
|
8 (5–11)
|
0.61
|
|
Plastic referral to discharge (days)
Median (IQR)
|
18 (14–25)
|
19 (11–24)
|
17 (14–25)
|
0.85
|
|
Number of CVTS procedures before plastic surgery referral
|
|
|
|
0.002
|
|
Yes
|
34 (47%)
|
9 (27%)
|
25 (64%)
|
|
No
|
38 (53%)
|
24 (73%)
|
14 (36%)
|
Table 4
Effect of early versus late referral on key outcomes
|
Key outcomes
|
Odds ratio (95% CI)
|
p-Value
|
|
Abbreviation: CI, confidence interval.
|
|
Death
|
0.54 (0.08–3.45)
|
0.52
|
|
Post-op complications
|
0.66 (0.21–2.08)
|
0.48
|
|
Prolonged stay (> 7 days)
|
1.39 (0.24–8.17)
|
0.71
|
Discussion
The approach of cardiac surgeons and plastic surgeons to DSWI differs drastically.[19] The traditional approach of cardiac surgeons has been sternal debridement, rewiring,
and closed drainage as opposed to the plastic surgeons’ approach that is more radical
with aggressive debridement and definitive closure with flaps. According to Brandt
and Alvarez,[19] the first-line treatment of DSWI showed better outcomes through a plastic surgical
strategy than those achieved by the traditional approach of cardiac surgeons. In our
study, both the early and late referral groups had similar outcomes with respect to
the total treatment duration after their referral to the plastic surgery unit. This
is due to the fact that in both groups, the treatment offered by plastic surgeons
was radical debridement and early flap cover.
A similar study comparing immediate versus delayed one-stage sternal debridement and
pectoralis major muscle flaps was done by Cabbabe and Cabbabe.20 Delayed group consisted of 14.8% patients who underwent multiple procedures by the
cardiothoracic team for sternal infection and were referred to plastic surgery for
sternal dehiscence and persistent infection after 22 to 63 days of an established
DSWI diagnosis. The immediate group had 85.2% patients who were seen by a plastic
surgeon within 4 days of the diagnosis. The plastic surgical management was a one-step
debridement and bilateral pectoral muscle flap reconstruction. In our series, it was
not possible to close the wound immediately. We had 59 patients wherein BIMA were
used and 85% patients (n = 50) out of these were diabetic. They had extensive sternal loss due to devascularization
of sternum as both vessels were harvested. So multistage procedures were required
in these cases. In our study, the early referral group underwent procedures by the
cardiothoracic team in 27% of cases before plastic surgery referral, whereas in the
late referral group it was 64%. This was found to be statistically significant (p = 0.002). The delay in referral to plastic surgery unit after onset of symptoms increased
the total duration of wound healing thereby increasing the treatment duration. The
average duration of treatment of DSWI, that is, the onset of symptoms to definitive
flap surgery in the early group was 16.58 days and in the late group it was 89.36
days. After referral to plastic surgery, the patients were discharged in a median
of 8 days in the early group and 9 days in the late groups, with no statistically
significant difference in both the groups. This shows that even after late referral
because of aggressive debridement and early definitive flap cover, comparative results
could be achieved.
In the series by Ringelman et al,[21] the pectoralis major was used as either a transposition flap or a reverse turnover
flap that needed extensive dissection and caused denervation, leading to cosmetically
unacceptable bulges, impairing the normal shoulder functioning and required other
modality for complete wound coverage. Fifty-one percent of patients complained of
persistent pain, 33% had shoulder weakness, and 25% had unacceptable cosmetic result.
Ascherman et al[22] and Scully et al[23] have reported great results after usage of pectoralis major flaps and omental flap
at 2 and 3 years, respectively. In our series, 83% patients reported unrestricted
shoulder movement and 85% reported cosmetically acceptable scars. The striking difference
in the study by Ringelman et al and all the other studies including ours was that
they have used a transposition or a turnover flap, whereas we have used advancement
and double breasting flaps,[18]based on the thoracoacromial pedicle. Since the plane of the muscle is same in the
advancement flap, there are no bulges associated, resulting in better functionally
and cosmesis in our series.
Two patients in our study had spontaneous right ventricular rupture before they could
undergo definitive closure and they succumbed to it despite best efforts. After the
first episode, we started using paraffin impregnated gauze dressing as an interface
between the NPWT foam and the exposed heart to reduce the negative pressure.[24] The reasons that could have caused serious bleeding during NPWT are infectious erosion[16]
[25]
[26] or a displacement of the heart toward or in between the sternal edges[27] and fibrous adherence of the right ventricle (RV) to the sternum.[28] Both the patients who had RV rupture had extensive bony erosion of sternum and were
found to have Klebsiella infection with positive bone cultures.
Even if the patients are referred late to a plastic surgical unit, aggressive and
radical debridement, judicious use of NPWT, and early definitive flap cover are the
key to managing this life-threatening condition. Hence, we have proposed aggressive
treatment protocol for management of DSWI.[29] In patients with positive bone culture, the intravenous antibiotics are continued
for 6 weeks.[2]
The patients who were referred late to us were treated by cardiac team per the traditional
conservative approach (resuturing, rewiring and NPWT). This was a significant factor
in both the groups in our series. The plastic surgical approach was more aggressive
and radical with removal of all the hardware and nonviable bone. Delay in referral
to plastic surgery unit after onset of symptoms of DSWI increased the total duration
of wound healing, thereby increasing the overall morbidity. With our aggressive approach
of early flap cover, we did not find any significant difference in the other variables
such as presence of microorganisms, duration from definitive flap surgery to discharge,
postoperative complications, number of NPWT changes before definitive closure, mortality,
duration from plastic referral to discharge, number of flaps, presence of microorganisms
in bone culture, postoperative shoulder movement, and postoperative scars in both
early and late referral groups. Late presentation had equally better outcome contrary
to general belief that mortality will be higher in late referral group. Hence, we
conclude that despite late referral, with good technique and methodological approach,
we were able to achieve similar outcomes as early referral. However, early reference
could lead to highly significant reduction in total duration of hospital stay and
wound healing.
Conclusion
As seen in the study, procedures done by cardiothoracic team for DSWI (p = 0.002) only prolonged total recovery time and did not give additional benefits
in terms of early wound healing. Although the statistical difference between outcomes
of early and late cases managed by plastic surgeons after referral to them is not
significant in terms of mortality or post-flap cover functionality, delayed wound
healing, and postoperative complications, yet an early reference could lead to highly
significant reduction in total duration of hospital stay, wound healing, and costs.
Early referral of post-CABG DSWI to plastic surgeons by the cardiothoracic surgeons
is highly recommended.
