The incidence of bilateral mastectomies for cancer and/or risk reduction is increasing along with the rates of breast reconstructions.1 , 2 Bilateral autologous free flap reconstructions are complex procedures associated with longer operating hours and technical challenges for plastic surgeons, as well as possible increased medical and surgical risks for the patients.3 , 4
A substantial number of candidates for breast reconstruction will present with abdominal scars from prior operative interventions. Almost one-third of pregnancies are delivered via Caesarean section, creating a common abdominal scar in women.5 Other patients may have scars from open or laparoscopic abdominal procedures or due to risk-reducing bilateral salpingo-oophorectomies in BRCA gene mutation carriers.6
As a result, women diagnosed with breast cancer or candidates for prophylactic mastectomy may have 1 or multiple scars on the abdominal wall that could potentially restrict their reconstructive options. In abdominal flap reconstructions, previous abdominal and/or pelvic procedures have been shown to disturb perforator vessels, alter perfusion territories across the scar, and form fibrous scar tissue impeding flap dissection.7 Several studies have evaluated the effect of prior scars and found no difference in complication rates attributable to these prior incisions.7–11 These findings, however, typically rely on retrospective data analyses from mixed series of unilateral (predominantly) and bilateral flaps; therefore, the actual impact of the scars on bilateral reconstructions could be underestimated. Thus, in unilateral reconstructions, there is the advantage of choosing between the ipsilateral or contralateral hemi-abdomen, with the final decision being made intraoperatively depending on perforators caliber, flap perfusion, and tissue properties. In bilateral reconstructions, the impact of abdominal scars may be more significant, as both halves of the abdomen are used and may therefore potentially pose greater risks for the flaps and the donor site.
The aim of this study was to evaluate the impact of prior abdominal surgical scars on complication rates in abdominal flap breast reconstruction, with a special focus on bilateral flap procedures.
PATIENTS AND METHODS
All consecutive patients with abdominally based free flap breast reconstruction performed by the team of plastic surgeons at the Brigham and Women’s Hospital between 2007 and 2014 were eligible for the study; this included deep inferior epigastric (DIEP), superficial inferior epigastric artery (SIEA), and free TRAM flap reconstruction. The surgical techniques employed did not change during the study period and have been previously described elsewhere.7 , 12 In DIEP procedures, a standardized preoperative computed tomography angiography (CTA) assessment has been started in early 2008, and SPY intraoperative laser angiography has been intermittently used since 2012.
The data were prospectively collected into a study-specific database; all relevant characteristics as well as outcome variables were defined and categorized by the senior authors before study initiation. A dedicated physician assistant with expertise in breast reconstruction performed data entry and maintenance ensuring that the database was filled for each patient and flap.
Variables included baseline patient characteristics [age, body mass index (BMI), diabetes mellitus, hypertension, smoking status], oncologic history (mastectomy type, genetic predisposition, prior chemotherapy, or radiation therapy), and reconstructive characteristics (date of procedure, operating surgeon, reconstruction timing and laterality, flap type and weight, perforator characteristics, length of surgery, and length of stay). Data on abdominal scars and prior abdominal/pelvic procedures were also collected, and the scars were further grouped according to the incision/s on the abdominal wall: Pfannenstiel, midline, lower oblique, upper oblique, or laparoscopic.
Complications and reoperations were prospectively registered after postoperative outpatient visits and updated at the end of follow-up, if necessary. For the purpose of the outcome analysis, all complications were divided into flap-related (flap loss, fat necrosis, hematoma, seroma, infection) and donor-site (infection, seroma, hematoma, wound healing problems, hernia, bulging). Major complications requiring return to the operating room were further grouped as recipient-site reoperations and donor-site reoperations. Mastectomy flap revisions were registered separately and excluded from the current analyses. Medical complications (e.g., deep venous thrombosis, pulmonary embolism) were also listed for each patient.
STATA/SE (Version 13.1) for MacOS (StataCorp, Tex) was used for all statistical analyses. Pearson’s chi-square or t test were utilized for assessment of differences in patients with and without scars, as well as differences between unilateral and bilateral reconstructions.
All unilateral breast reconstructions were excluded from the statistical analysis, and data on these procedures are provided for the descriptive purpose.
All reconstructed site complications in bilateral procedures were registered per flap, but analyzed and reported on a per patient basis.
In the assessment of flap variables (e.g., flap weight, perforator characteristics), the calculations were performed for all flaps within the subgroup (unilateral versus bilateral or scar group versus no scar group). A 2-tailed P < 0.05 was considered significant in all statistical tests. Fisher’s exact test was used in the analyses of complications due to lower number of outcome events.
Four hundred ninety-three consecutive patients underwent abdominally based breast reconstruction during the study period, with laterality split as follows: unilateral (n = 250; 50.7%), bilateral (n = 243; 49.3%). Overall, DIEP flaps were utilized in 654 (88.9%), fTRAM 43 (5.8%), SIEA 13 (1.8%), TRAM 8 (1.1%) breast reconstructions. Patients in the bilateral group were younger (47.0 versus 50.0 years; P = 0.0001), had a greater BMI (29.1 versus 27.7 kg/m2; P = 0.0012), and more often had comorbidities as obesity (39.9% versus 27.6%; P = 0.014), diabetes (3.7% versus 0.8%; P = 0.028), hypertension (21.9% versus 14.5%; P = 0.032), see Table 1.
The bilateral reconstruction group more frequently had genetic predisposition to breast cancer (41% versus 3.6%; P < 0.001) and significantly more often received immediate breast reconstruction (61.7% versus 47.6%; P = 0.0001). There were also differences in selected flap type (P = 0.022), operating plastic surgeon (P = 0.013), and length of surgery (847.0 versus 460.4 minutes; P < 0.001) between the bilateral and unilateral reconstructions, respectively (Table 1).
Scars and Previous Abdominal/Pelvic Procedures
In total, 148 of 243 (62.2%) patients with bilateral reconstruction had at least 1 prior abdominal (13.2%) or pelvic (31.7%) procedure, or a combination (14.8%; Table 2). Five other patients underwent bilateral prophylactic salpingo- oophorectomy concurrently with breast reconstruction.
The scar locations on the abdominal wall were as follows: Pfannenstiel (n = 73; 30.1%), midline (n = 16; 6.6%), lower oblique (n = 17; 7.0%), upper oblique (n = 5; 2.1%), and laparoscopic (n = 69; 28.4%). Some patients also had more than 1 scar each, most often having a combination of Pfannenstiel and another incision (Table 2; Fig. 1).
When compared with the unilateral group, the proportion of laparoscopic interventions among patients in bilateral group was significantly higher (28.4% versus 12.8%; P < 0.001).
Bilateral Reconstructions: Scar Group Versus No Scar Group
To further evaluate the role of scars in bilateral breast reconstruction procedures, we stratified our patient population into a no prior scar group and a prior scar group. Reconstructive timing in the groups was similar, but there were differences in flap design. In the scar group, fTRAM flap was utilized more frequently when compared with the no prior scar group (Table 3).
The average flap weight was higher in the scar group (739.7 versus 635.6 g; P = 0.0036), although it was in line with the average mastectomy specimen weight within the groups (737.5 and 656.6 g, respectively). Other flap variables such as perforator characteristics, ischemic time, venous anastomoses, and mean vein coupler size did not differ between the groups (Table 3). Although not statistically significant, the mean length of surgery in scar group was higher (971.9 versus 655.8 minutes; P = 0.41).
Complications in Bilateral Reconstructions: Scar Group Versus No Scar Group
Overall, 4 of 238 (1.7%) flap failures (including 1 pedicled TRAM flap) were registered, all occurring in patients from the scar group: 3 with Pfannenstiel scar and 1 patient with prior laparoscopic intervention (Tables 4, 5).
Fat necrosis/partial flap necrosis, infection, and seroma occurred in 14 (5.9%), 8 (3.4%), and 5 (2.1%) patients, respectively, and no differences between the groups were identified (Table 4). The incidence of postoperative hematoma at the recipient site was 14 of 238 (5.9%) and found more frequently in patients with scars (8.1% versus 2.2%; P = 0.051).
Recipient-site reoperation rate (i.e., unplanned revisions) was 19 of 238 (7.9%) and patients with scars were affected more often (10.1% versus 4.4%; P = 0.090), although the difference did not reach statistical significance.
In the subgroup analysis, the presence of Pfannenstiel scar was associated with higher risk of hematoma at the recipient site when compared with no scar group (13.7% versus 2.2%; P = 0.006). No other differences were seen when comparing subgroups of patients with Pfannenstiel (n = 73) or laparoscopic incisions (n = 69) with the no scar group (n = 90), Table 4, right.
Mastectomy flap revisions were performed in 16 patients: 11 (7.4%) scar group and 5 (5.6%) no scar group, but were not included in the study analyses (data not shown).
Donor-site complications included delayed wound healing, infection, bulging, and seroma in 11 (4.6%), 1 (0.4%), 7 (2.9%), and 4 (1.7%) patients, respectively, and the donor-site reoperation rate was 5 of 238 (2.1%) with no significant differences between the groups (Table 4, below).
In the subgroup analysis of Pfannenstiel or laparoscopic incisions (Table 4; right), patients with prior laparoscopic procedures had problems with wound healing more often compared with no scar population (8.7% versus 2.2%; P = 0.069), although not statistically significant.
Systemic complications included 3 cases of deep venous thrombosis (1.2%) and 1 case of pulmonary embolism (0.4%).
This is the first study to assess the role of abdominal scars in a cohort of bilateral abdominal flap breast reconstructions. Our findings suggest that surgical outcomes of bilateral reconstructions in patients with abdominal scars are generally comparable with ones in patients without prior surgery; however, some problems have been identified.
Several previous articles have addressed the effects of prior abdominal surgery on the autologous breast reconstruction.7–11 The main limitation of these studies is that their data were analyzed on a per flap basis, where the variables for unilateral and bilateral procedures were collapsed to increase the power of analyses. In the current study, we exclusively focused on bilateral reconstructions and report the outcomes on a per patient basis to adjust for intrinsic complexity of each bilateral procedure.
Parrett et al.7 were the first to evaluate abdominal scars in DIEP breast reconstructions in a cohort of patients with (n = 78) and without (n = 90) scars showing no increase in flap complications, but significantly higher donor-site morbidity in the scar group (24% versus 6.7%; P = 0.003).
A recent article by Wes et al.11 evaluated abdominally based breast reconstructions in patients with (n = 417) and without (n = 395) scars and came to similar conclusions. In this large series of patients, no differences in flap- related outcomes or medical complications were identified, but patients with prior rectus sheath violation (both open and laparoscopic) had problems with donor-site wound healing more frequently (22.7% versus 16.5%, P = 0.03).
Although both studies analyzed unilateral and bilateral flaps, their findings on the donor site were consistent with ours where delayed wound healing after laparoscopic procedures trended to be higher (8.7% versus 2.2%; P = 0.069).
In addition, our results indicate that reconstruction site reoperation trended to occur more often in the scar group (10.1% versus 4.4%; P = 0.090), and the incidence of reconstructed site hematomas was statistically significantly higher in the subgroup with Pfannenstiel scar (13.7% versus 2.2%; P = 0.006). This could be explained by the fact that in patients with larger abdominal incisions, the anatomy of the vessels and respective angiosomes/venosomes could be disturbed and the tissues in the areas of previous dissections have been violated.10 This may be more consequential in bilateral reconstructions when the surgeon is forced to maximize tissue utilization and employ both sides of the abdomen (without the luxury of choosing the better perfused side).
Conversely, some studies found that Pfannenstiel scar was associated with larger perforators and may potentially augment the perfusion of the lower abdomen by promoting the development of collateral pathways of circulation via delayed-phenomenon mechanisms,10 , 13 similar to those utilized in SIEA delay procedures.14 However, there was no reported difference between the presence or absence of usable SIEA vessels in patients with abdominal scars described in the operative reports. There is also a hypothesis that improved perfusion in multiparous women may be attributed to pregnancy per se rather than delivery via Caesarian section and subsequent scar.15 This could be explained by the gradual expansion of the abdominal wall by the pregnant uterus and consequent vascular rearrangements (e.g., extending axiality of the perforator vessels and/or collateral pathways), which could be beneficial in fasciocutaneous flap procedures.
In our cohort, the rates of fat necrosis/partial flap necrosis in subgroups of patients with and without Pfannenstiel incision were comparable. However, 3 flap failures in the study occurred in patients with Pfannenstiel scar after total abdominal hysterectomy (n = 2) or Caesarian section (n = 1), Table 5. Although not statistically significant, this finding has an important clinical relevance and should be considered in patients with Pfannenstiel scars.
Several studies have also discussed how to address and improve reconstructive outcomes in patients with Pfannenstiel,10 vertical midline,16 , 17 vertical infraumbilical,18 subcostal,19 and other scar types.9
Our data showed that flap design and harvest were not different between the scar group versus no scar group. All patients included in this study had a preoperative plan for DIEP flaps. The fTRAM, pedicled TRAM, or SIEA decision was made intraoperatively if a DIEP was not viable (after clamping trials of exposed perforators). The presence of abdominal scars did not increase the need to convert to fTRAM (1 would expect that if the abdominal scar injured perforators, Table 3). The presence of scars also did not influence which row of perforators were used or increase the number of perforators required for adequate DIEP perfusion (Table 3). In this study on bilateral flaps, cases with scars seemed to be approached as they otherwise would, for example, a mid-line scar did not necessarily mean that medial row perforators are damaged.
Although not statistically significant, the mean length of surgery in patients with abdominal scars was on average 1.5 times longer, which could be explained by the presence of the scars as well as the complexity of the dissection in the scar tissue planes and altered vascularity.
Overall, patients undergoing bilateral breast reconstruction have certain demographic and perioperative differences, as well as some technical considerations when compared with unilateral reconstructions (see table, Supplement Digital Content 1, which displays rationale to separate unilateral and bilateral DIEP breast reconstructions in publications. An assessment of these differences is beyond the scope of this investigation, http://links.lww.com/PRSGO/A542).
One of the strengths of this study is its methodology where all variables for each patient and flap, including the postoperative follow-up course, were prospectively collected into a study-specific free flap database. A dedicated physician assistant updated the database after postoperative follow-up visits so that no patient was lost for the study follow-up.
To eliminate selection bias, all consecutive patients scheduled for abdominal flap reconstruction were included in the analyses according to the intention-to-treat principle. This approach allowed studying the actual reconstructive course even in scenarios when the reconstruction changed intraoperatively, for example, converting from a DIEP to a pedicled TRAM procedure.
In addition, we stratified the study population into unilateral and bilateral groups and focused exclusively on patients with bilateral reconstructions where the effects of abdominal scars might be more consequential. Aiming to adjust for intrinsic complexity of each bilateral procedure, we analyzed reconstruction outcomes on a per patient basis, rather than counting as 2 separate unilateral reconstructions.
Conversely, the study may be underpowered because of the very low complications rate, when the differences between the study groups did not reach statistical significance. A clinically significant difference was however observed: the rate of flap failure in scar group and no scar group was 4 and zero events, respectively. As in any other retrospective study in plastic surgery, a certain preselection bias could not be avoided. Thus, patients who are poor candidates (small donor sites, high risk for other reasons) and women with CTA showing no perforators (possibly due to abdominal scarring) were not included, as they would have never been offered a DIEP.
The findings of this study could be used by plastic surgeons to facilitate preoperative planning in patients with prior abdominal procedures. In addition, the data will be helpful for gynecologists and general surgeons for planning their interventions and incisions, keeping in mind a possible need for flap reconstructions in future. The time interval from the abdominal procedures to flap harvest and number of previous procedures might also play a role, yet we have not included this information in the current study. Furthermore, data on the intraoperative assessment with fluorescence angiography were not collected prospectively and not included, although its routine use in this patient group is considered beneficial.
In conclusion, surgical outcomes of bilateral abdominal flap reconstructions in patients with abdominal scars are generally comparable with ones in patients without prior surgery. No greater risks for recipient- or donor site were demonstrated in general; however, some scar-specific problems have been identified. These procedures might have some intraoperative considerations and often require increased operative times. Apart from the traditional preoperative CTA, intraoperative imaging (e.g., fluorescence angiography) may be advocated in patients with abdominal scars. To further understand effects and vascular changes in patients with abdominal scars, more anatomical and imaging studies are needed.