An Institutional Analysis of Risk Factors in Pectoralis Advancement Flap Reconstruction

The median sternotomy has long been the preferred approach for cardiac surgery since it was first described in 1957.¹ Sternal wound infection is a known complication of this approach. These infections are often difficult to treat as they invariably lead to dehiscence and may result in cardiac exposure.
In the past 45 years, treatment of sternal wound infections has evolved from debridement with open granulation, to rewiring with closed catheter irrigation, and finally to flap reconstruction. Omental flaps, rectus advancement flaps, and pectoralis advancement flaps have all been used to close these defects, usually with either a 1-stage or 2-stage procedure.^2,3 As the treatment of sternal wound infections evolved, it is generally regarded that a 1-stage operation with pectoralis advancement flap carries the lowest morbidity and is associated with the shortest hospital stay.^2–5
Much has been written about the risk factors associated with morbidity and mortality in patients undergoing flap reconstruction for sternal wound infection. Though some of these studies vary in their results, most experts agree that advanced infection with the presence of systemic infection denotes a poor prognosis. Most clinicians agree that early diagnosis with aggressive antibiotic therapy, debridement, and reconstruction are essential to mortality reduction.^2–7

Methods

All patients who underwent bilateral pectoralis flap reconstruction (BPF) from 1996–2004 were identified using the Birmingham, Alabama Veterans Affairs Hospital OR database. A similar database was used at the University of Alabama, Birmingham Hospital to identify patients who underwent the pectoralis advancement flap procedure. Data were obtained through electronic chart review and analyzed with respect to patient demographics, operation, reconstruction, risk factors, and outcome.
The UAB cohort group was comprised all patients at UAB who underwent flap reconstruction for sternal wound infection from 1996 to 2004 (n = 93). All female, organ transplant, pediatric, aortic dissection, and unilateral pectoralis flap patients, as well as those who had undergone their third or fourth sternotomy were excluded. This resulted in 2, all-male cohort groups who had been treated with bilateral pectoralis flap reconstruction for sternal wound infection. All patients’ wounds were from either primary or secondary coronary artery bypass graft surgery (CABG), with or without valve replacement. The same surgeons at both centers performed the cardiac and reconstructive procedures. The UAB group contained 29 patients and the VA group 21 patients.
Age describes age at the time of operation and lab values were within 5 days of reconstruction. Other details regarding operative complications, pre-reconstruction debridement, and cause of death are outlined in Tables 1–4. Mortality was defined as any death that occurred during the same admission as the reconstruction operation or a death that occurred as the direct result of a postoperative infection or complication.
Data were collected with the approval of the BVA institutional review board and the UAB institutional review board. Fisher’s exact test with 2-sided P values was used to determine odds ratios and P values. For value data, an unpaired t-test with a 2-tailed P value was used.

Results

As shown in Tables 1 and 2, the UAB and BVA groups were similar with respect to most demographics and preoperative risk factors, such as sex, age, diabetes, hypertension, and CABG vessel number. Though numerical comparison shows some differences with respect to renal function and nutritional status, these differences are not statistically significant.
Table 2 shows that hospital stays in the BVA group were longer. The BVA group had a longer mean LOS for cardiac (8.9 versus 21.8 days, P = 0.036) and reconstructive procedures (9.5 versus 17.2 days, P = 0.011).
The percentage of complications associated with bilateral pectoralis flap reconstruction was comparable for both groups (27.6% at UAB versus 23.8% at BVA, P = 1.00). These complications are shown in Table 3. Several patients had multiple complications accounting for the listed values.
The number of patients treated with a 2-stage operation or preoperative debridement was comparable between the 2 groups (n = 9 at UAB versus n = 9 at BVA). Patients qualifying for pre-reconstruction debridement include those who underwent wire removal, incision and drainage, soft tissue debridement, and sternectomy. The details of their treatment are shown in Table 4.
The differences in mortality between the 2 groups should also be noted (6.9% in the UAB group compared with 19.0% in the BVA group, P = 0.223). The UAB group had 2 post-reconstruction deaths whereas the BVA group had 4. Sepsis and multiple organ dysfunction syndrome (MODS) were causes of death in 3 out of 4 BVA patients. One BVA patient died from a postoperative pulmonary thromboembolism (PTE). In the UAB group, 1 patient died from acute respiratory failure and another after myocardial infarction (MI).

Discussion

As treatment of post-sternotomy wound infections evolved, much has been written about the risk factors associated with sternal wound reconstruction. Previous studies have shown that poor nutrition, internal mammary artery (IMA) use, female sex, renal failure, and a longer time interval between cardiac operation and reconstruction are associated with increased mortality and morbidity in patients treated with flap coverage.^2–6
Despite variable data on risk factors associate with successful sternal wound reconstruction, most experts agree that early detection, early aggressive treatment, and high quality critical care are essential to successful reconstruction. Jones et al³ published what is perhaps the most comprehensive study on risk factors associated with post reconstruction morbidity and mortality. In their 20 years of experience they found that septicemia (in addition to perioperative MI and intraaortic balloon pump use) is associated with increased mortality. They attribute their reduced mortality rates and shorter hospital LOS to improvements in home antibiotic infusion therapy, trends toward 1-stage reconstruction, and early referral—all of which reduce the risk of sepsis and MODS.³ Others have designed wound classification systems, assigning more aggressive treatment to patients with more advanced infection.^3–5 Though this concept may seem intuitive, it reiterates the importance of early diagnosis and treatment of sternal wound infections.
The present study set out to compare 2 cohort groups matched for demographic and preoperative risk factors. Though not statistically significant, numerical comparison shows that patients treated at the BVA were 3.2 times more likely to die after bilateral pectoralis advancement flap than patients at UAB (P = 0.22). The causes of death may explain the differences in mortality between the 2 groups. The BVA group had 3 of 4 patients die from sepsis-related multiple organ failure. The UAB group had zero sepsis-related deaths. Other studies show a similar association between post reconstruction sepsis and mortality. Jones et al³ report a 25% mortality rate among septic patients versus 3% mortality among those who did not have sepsis.
The BVA group had a significantly longer mean hospital stays for cardiac (8.9 versus 21.8 days, P = 0.036) and reconstructive procedures (9.5 versus 17.2 days,
P = 0.011). Other studies show that patients treated at VA health systems have longer LOS for a variety of illnesses. In 1998, Thomas et al⁸ reported that VA patients have a risk-adjusted, 36% longer LOS when compared with diagnosis-matched patients in the private sector. They site VA system inefficiencies and VA patient population characteristics as explanations for the increased LOS.
One explanation for the BVA group’s increased number of sepsis-related deaths and longer hospital LOS is that BVA patients presented with more advanced sternal wound infections than patients in the UAB group. The VA patient population has a higher prevalence of chronic disease, substance abuse, mental illness, and homelessness compared to the general population. Patients in VA institutions are also known to have lower education levels, poorer social support, and are less likely to have health insurance.^9,10 Despite efforts, such as community-based outpatient clinics (CBOCs), and VA funded home health, VA resources for transportation, nursing homes, and rehabilitation centers remain limited. Patients in the private sector do not face most of these challenges. Consequently, patients treated for sternal wound infection in the private sector are thought to have an earlier referral, diagnosis, treatment, and postoperative discharge.

Conclusion

The explanation for the mortality differences and for the longer hospital LOS is based primarily on what is known about the VA patient population and the VA system in general. There are certainly other potential explanations for the outcome differences in this study. While the same surgeons performed both the cardiac and reconstruction procedures at both centers, there is little data in this study comparing the quality of postoperative care. This study also fails to consider specific patient population, referral network, and healthcare system variables in the BVA and UAB groups. Additionally, this study is underpowered, failing to show statistically significant differences in the mortality rates between these 2 groups. Further evaluation with larger study groups and more detailed patient variables, including culture data and antibiotic therapy, may offer additional insight into the different outcomes for these study groups. These limitations aside, the BVA and UAB groups although matched for a number of patient and treatment variables, clearly have outcome differences. The authors believe that the differences in mortality and longer hospital stays in the BVA group are due largely to the differences in patient population, critical care, and patient referral networks. Early diagnosis with aggressive antibiotic therapy, debridement, and reconstruction remain the mainstays for successful sternal wound treatment.