Total Flow Rate
Both surgeons use visual estimates of the external diameter of perforating vessels as they enter the flap with a specific, five-category measurement system: less than 1 mm and not seen on computed tomographic angiography, 1 mm, 1.5 mm, 2.0 mm, and 2.5 mm and above. Concordance between the two senior authors was confirmed for all co-surgery reconstructions. These calibers were then used to calculate flow from the Poiseuille law in millimeters per second. For all calculations, the pressure differential was between a theoretical 120 and 80 mmHg to simulate systolic and diastolic blood pressures, viscosity was 40 mP for blood, and the length was 1 mm.
Postoperative Follow-Up and Assessment of Fat Necrosis and Abdominal Morbidity
Standard postoperative protocol for the senior surgeons is a 2-week, a 6-week, and then an additional 12-week follow-up followed by annual visits. Fat necrosis and abdominal morbidity data were recorded from postoperative clinic notes and imaging with the senior authors, the oncologic breast surgeons, or both.
Fat necrosis was defined as a greater than 1-cm palpable, distinct mass on physical examination or imaging that was present beyond 6 weeks postoperatively. Careful review ensured that the first evidence of fat necrosis was noted before secondary fat grafting or revision surgery.
Abdominal wounds and donor-site complications were defined as any wound with delayed healing that required local wound care or negative-pressure wound therapy, abdominal cellulitis, incisional fat necrosis or asymmetry, and any hematoma or seroma. Abdominal bulges were defined by physical examination and computed tomographic imaging. All documentation or imaging for abdominal bulges or wounds was carefully reviewed for laterality in patients undergoing bilateral breast reconstruction.
Nerve-Sparing and Nerve-Repair Techniques
If all identifiable nerves were successfully spared during perforator harvest, it was deemed a nerve-sparing procedure. If identifiable nerves had to be sacrificed to adequately mobilize the pedicle, but then all were primarily repaired, this was deemed a nerve-repairing procedure. Figure 2 depicts a nerve-sparing dissection.
A univariable logistical regression analysis was first conducted with all 28 of the above-listed patient demographic and surgical factors to calculate an odds ratio with 95 percent confidence intervals on three primary outcomes: (1) DIEP flap fat necrosis, (2) abdominal wounds or complications, and (3) abdominal bulge or hernia. All variables with a value of p < 0.15 for the calculated odds ratio in univariable analysis were then entered in a backward selection algorithm to yield the parsimonious multivariable logistic regression model. An odds ratio with a value of p ≤ 0.05 in the multivariable analysis was considered statistically significant.
Summary statistics for patient characteristics were reported using medians and ranges for continuous variables, and using counts and percentages for categorical variables. Patient characteristics were compared to analyze potentially confounding presurgical patient demographics among the significant predicting variables. For continuous patient characteristics, analysis of variance or Kruskal-Wallis nonparametric analysis of variance was used as appropriate for comparisons of three or more groups, and two-sample t test or the nonparametric Wilcoxon rank sum test was used for two-group comparisons. For categorical patient characteristics, the chi-square test or Fisher’s exact test was used as appropriate. All analyses were conducted using SAS version 9.4 (SAS Institute, Inc., Cary, N.C.).
Four hundred nine total DIEP flaps were included in our statistical analysis after strict inclusion and exclusion criteria were applied. The average age and body mass index of the patients was 50.5 and 30.7 kg/m2, respectively; 68.5 percent were Causation, 16.9 percent were black, 9 percent were Hispanic, and 5.6 percent were Asian. The average follow-up for these patients was 18.5 months, with a median of 15.75 months; 91.9 percent of flaps were bilateral and 8.1 percent were unilateral reconstructions for 32 percent immediate, 50.9 percent delayed-immediate, and 17.1 percent delayed breast reconstructions. Of these, 14.4 percent had flap fat necrosis, 21.3 percent had an abdominal wound or complication, and 6 percent had an abdominal bulge or hernia.
Analysis showed a significant increase in the odds of fat necrosis with increasing flap weight (OR, 1.002 per 1-g increase; or OR, 1.2 per 100-g increase; p = 0.0002), and earlier year of surgery (OR, 2.51 for 2010 to 2013 versus 2014 to 2016; p = 0.01).
A significant decrease in the odds of fat necrosis was seen with lateral row (OR, 0.29; p = 0.001) and both medial and lateral row perforator flaps (OR, 0.21; p = 0.001) versus medial row–based flaps, in patients who underwent neoadjuvant chemotherapy (OR, 0.36; p = 0.01), and if indocyanine green was used to assess perfusion (OR, 0.46; p = 0.04). Overall, indocyanine green was used intraoperatively in 130 flaps (31.8 percent) and directly guided excision of areas with poor perfusion in 38.5 percent of those cases.
The average total flow rate for all flaps was 0.78 mm/second (median, 0.52 mm/second). Increasing total flow rate of the flap also decreased the odds of fat necrosis (OR, 0.62 per 1-mm/second increase; p = 0.05).
The number of perforators did not significantly affect fat necrosis rates after being controlled for by these other significant variables in the multivariable analysis. In addition, excising corners/zones of the flap prophylactically without indocyanine green angiography did not significantly affect fat necrosis rates. The results of the multivariable analysis for fat necrosis are summarized in Table 2.
Abdominal Wounds and Complications
There were significant increased odds of having an abdominal wound or donor-site complications with smoking (OR, 1.86; p = 0.02), hypertension (OR, 1.72; p = 0.04), and increasing flap weight (OR, 1.001 per 1-g increase; p = 0.001). Body mass index was not a significant factor for abdominal wounds when controlled by flap weight in the multivariable analysis. Overall, only 25.3 percent of abdominal wounds or complications required operative treatment, including dog-ear or scar revisions. The results of the multivariable analysis for abdominal wounds and complications are summarized in Table 3.
Abdominal Bulge or Hernia
Increased odds of abdominal bulge or hernia were seen with a return to the operating room on the initial hospital stay (OR, 5.05; p = 0.01), with flaps based on the lateral or both medial and lateral row perforators (OR, 3.21; p = 0.05) versus medial row perforators, and with patients who had an abdominal or umbilical wound postoperatively (OR, 2.59; p = 0.05). In total, there were 127 nerve-sparing (31 percent) and 22 nerve-repairing (5.4 percent) flap dissections. However, using mesh for rectus fascia repair in 7.8 percent of flaps and using nerve-sparing or nerve-repairing techniques did not significantly affect the odds of abdominal bulges. There were no complications from mesh use. The results of the multivariable analysis for abdominal bulges or hernias are summarized in Table 4.
Table 5 summarizes the comparisons of patient demographics between perforator row types. Similar tables were made for each significant variable after multivariable analysis (data not shown). The only differences between presurgical demographics that had a significant effect on one of the primary outcomes measures was fewer smokers (21 percent; p = 0.0008) in the lateral row cohort versus medial (36.2 percent) and both medial and lateral row (45.4 percent) cohorts, less hypertension (32.1 percent versus 42.2 percent; p = 0.041) and more neoadjuvant chemotherapy (35.9 percent versus 21.7 percent; p = 0.003) in the 2014 to 2016 versus 2010 to 2013 cohorts, and less smoking in the group that received neoadjuvant chemotherapy (27.8 percent versus 37.8 percent; p = 0.049).
Previous publications have described variables that affect free flap fat necrosis after abdominally based autologous breast reconstruction. Baumann et al. showed in their series of 228 abdominal free flaps that fat necrosis was higher in flaps with only one or two perforators, in patients who smoked, and the inclusion of zone 3 in flaps.1 Although this study used a multivariable analysis and estimates of perforator caliber, it included muscle-sparing TRAM and superficial inferior epigastric artery flaps which, as pointed out by Rozen et al.,2 have anatomical and physiologic differences. Grover et al. added to the specificity of this concept by showing in a review of 395 DIEP flaps that with a single-perforator flap there was a higher rate of fat necrosis than with a multiperforator flap.3 However, this was not a multivariable analysis and perforator location was not a measured variable. A similar result was seen with 100 flaps in addition to a correlation between higher body mass index and increased fat necrosis by using a multivariable model.4
In contrast, Mulvey et al. looked at 179 muscle-sparing TRAM and DIEP flaps and found higher odds of fat necrosis with increasing flap weight, but not with perforator number or perforator row.5 Their multivariable analysis suggested that including multiple perforators may be protective for fat necrosis, but this was not statistically significant because of the small sample size.
Lindsey, in a critique of conclusions to simply include more perforators to decrease fat necrosis, cited the importance of also considering perforator caliber and perforator location.6 , 7 Kamali et al. confirmed the importance of perforator location, with a 24.5 percent versus 8.2 percent fat necrosis occurrence from DIEP flaps based on the medial row versus the lateral row.8 This study showed a general decrease in fat necrosis within the same row with increasing perforator number, but no statistical test was conducted to assess the significance of different numbers of perforators on fat necrosis rates.
On univariable analysis of our flap data, multiple-perforator flaps had decreased odds of fat necrosis compared with single-perforator flaps (OR, 0.54; p = 0.03). This significance was lost in the multivariable model (p = 0.55). This model instead showed that flaps based on lateral and both medial and lateral row perforators (OR, 0.29 and 0.21, respectively) decreased the odds of fat necrosis independently (Table 2). In addition, increasing total flow to the flap (OR, 0.63 per 1-mm/second increase) (Table 2) was associated with a lower rate of fat necrosis. Overall, our fat necrosis rate of 14.4 percent was consistent with the rate from other recent studies.1 , 8 According to the Poiseuille law, increasing the caliber of a perforator has an exponentially larger effect on flow rates than adding additional small perforators. The results follow this concept and subsequently suggest that perforator location and perforator caliber may be more important intraoperative factors to decrease fat necrosis than simply recruiting more perforators alone.
Because of the presumptions made to calculate the total flow rate, the absolute values are less meaningful than the overall trend. In addition, we acknowledge the potential inaccuracies of estimating flow based on external diameters of blood vessels,2 but calculating perforator size solely based on preoperative computed tomographic angiography has its own limitations. For the senior authors, after being guided by computed tomographic angiography, intraoperative visual estimates of external diameter coupled with the pulse and Doppler signal best simulates what is used in the operative room to assess the quality of a perforator.
Using laser-assisted indocyanine green fluorescence angiography for assessing perfusion to abdominal free flaps has been described, with several studies correlating areas of hypoperfusion on indocyanine green angiography with postoperative flap or fat necrosis.22–27 Indocyanine green angiography was used as an adjunct to our computed tomographic angiography–guided perforator selection algorithm in 31.8 percent of flaps. Primarily, it was used after the flap had been anastomosed on the chest to subjectively assess for areas of hypoperfusion if there was clinical concern. Less commonly, it was used to subjectively assess perforator quality on the abdomen if there was a discrepancy with clinical findings versus preoperative computed tomographic angiography. This use of indocyanine green angiography decreased the odds of fat necrosis (OR, 0.46) (Table 2).
There have been studies purporting the safety of performing breast reconstruction after neoadjuvant chemotherapy with similar complication rates to patients without neoadjuvant chemotherapy,15 , 27–32 albeit with some conflicting data.33 When comparing the demographic data between flaps performed in 2010 to 2013 versus 2014 to 2016, there was less neoadjuvant chemotherapy in the earlier cohort (21.7 percent versus 35.9 percent). Interestingly, neoadjuvant chemotherapy was associated with reduced odds of fat necrosis (OR, 0.36) in the multivariable analysis even after being controlled for by year of surgery (Table 2). However, in the flap cohort that underwent neoadjuvant chemotherapy, we found there was less smoking (27 percent versus 37 percent), less immediate breast reconstructions (18.2 percent versus 38.2 percent), less adjuvant chemotherapy (16.7 percent versus 27.9 percent), and a younger age (46.9 years versus 52.1 years). These variables have previously been shown to affect fat necrosis and could have had a cumulative unmeasured effect in our study.1 , 9 , 31 Thus, we believe that the decreased odds of fat necrosis in the later period were attributable to improved surgeon experience rather than more patients with neoadjuvant chemotherapy having undergone reconstruction.
In addition, the size of the flap is important to consider, as flap weight increases not only the odds of fat necrosis (OR, 1.2 per 100-g increase) (Table 2) but also abdominal wound or complication rates (OR, 1.1 per 100-g increase) (Table 3). The relationship between flap weight and fat necrosis is depicted in Figure 3, with an increase in the probability of fat necrosis from 11.25 percent to 25 percent from a 500-g flap to a 1500-g flap, respectively. The other factors associated with increased odds of abdominal wounds and complications were active smoking (OR, 1.869) (Table 3) and hypertension (OR, 1.72) (Table 3), which is consistent with other studies.5 , 15 , 16 , 20
The presence of an abdominal wound itself was independently associated with increased risk of abdominal bulge or hernia (OR, 2.59) (Table 4). This scenario is particularly true with umbilical necrosis, which can cause a loss of domain, and one should critically evaluate umbilical stalk height to try and prevent the sequelae of these donor-site wounds.20 , 34 Returning to the operating room on the initial hospital stay significantly increased the odds of abdominal bulge (OR, 5.05) (Table 4). Because 95.7 percent of these take-backs were for recipient-site hematomas or flap congestion, we believe this association with abdominal bulge is attributable to the urgent nature of these cases and flexed positioning not being well maintained.
Rozen et al. suggested that the important motor nerves supplying the rectus enter with the lateral row perforators, thus placing them at risk when harvesting a laterally based flap.11–13 , 19 This is particularly salient in the setting of our results that, along with previous data,8 suggest that the addition of a lateral row perforator decreases fat necrosis. However, Garvey et al. published data on 615 abdominal free flaps and found no difference in donor-site bulge or hernia (overall rate, 4.6 percent) between medial and lateral harvest.14 Our overall abdominal and hernia rate of 6.1 percent is comparable. However, their criteria excluded all flaps based on both rows. With the idea that in either a lateral or both medial and lateral row perforator harvest there is potential insult to the lateral motor nerves,11 we compared lateral and both medial and lateral row–based flaps together versus medial row flaps and found an increased risk of bulge rates (OR, 3.21) with the former (Table 4). Conversely, neither nerve-sparing, nerve-repairing, nor mesh repair was found to significantly reduce abdominal bulge rates (Table 4). However, given the low incidence of this complication and the low number of nerve-sparing, nerve-repairing, and mesh fascia repairs, the analysis lacks adequate power to extrapolate this negative finding and confidently conclude that neither mesh nor nerve preservation has a role in abdominal bulging. Therefore, additional, larger studies will be needed in the future to clarify their efficacy.
This study aims to elucidate the predicting factors for fat necrosis and abdominal morbidity in the patient undergoing DIEP flap reconstruction to aid in preoperative risk stratification and to help prioritize specific perforator selection parameters to optimize outcomes. Our results suggest that using larger caliber perforators and perforators from the lateral row alone, or in addition to medial row perforators, can decrease fat necrosis rather than by simply harvesting more perforators alone. However, lateral and both medial and lateral row perforator flaps come at the cost of increasing abdominal bulge rates. In addition, we found that the use of intraoperative indocyanine green angiography decreased the odds of fat necrosis, whereas increasing flap weight increased these odds. Larger flaps, smoking, and hypertension led to a higher rate of abdominal wounds. Returning to the operating room during the initial hospital stay and abdominal wounds themselves were associated with higher abdominal bulge rates.
The authors thank Hong Zhu, Ph.D., and Jingsheng Yan, Ph.D., of the Division of Biostatistics, Department of Clinical Sciences, Simmons Comprehensive Cancer Center, University of Texas Southwestern Medical Center, for expedient and accurate work.
1. Baumann DP, Lin HY, Chevray PM. Perforator number predicts fat necrosis in a prospective analysis of breast reconstruction with free TRAM, DIEP, and SIEA flaps. Plast Reconstr Surg. 2010;125:1335–1341.
2. Rozen WM, Whitaker IS, Chubb D, Ashton MW. Perforator number predicts fat necrosis in a prospective analysis of breast reconstruction with free TRAM, DIEP, and SIEA flaps. Plast Reconstr Surg. 2010;126:2286–2288; author reply 22882289.
3. Grover R, Nelson JA, Fischer JP, Kovach SJ, Serletti JM, Wu LC. The impact of perforator number on deep inferior epigastric perforator flap breast reconstruction. Arch Plast Surg. 2014;41:63–70.
4. Bozikov K, Arnez T, Hertl K, Arnez ZM. Fat necrosis in free DIEAP flaps: Incidence, risk, and predictor factors. Ann Plast Surg. 2009;63:138–142.
5. Mulvey CL, Cooney CM, Daily FF, et al. Increased flap weight and decreased perforator number predict fat necrosis in DIEP breast reconstruction. Plast Reconstr Surg Glob Open 2013;1:1–7.
6. Lindsey JT. Perforator number does not predict fat necrosis. Plast Reconstr Surg. 2011;127:1391–1392; author reply 1392.
7. Lindsey JT. Integrating the DIEP and muscle-sparing (MS-2) free TRAM techniques optimizes surgical outcomes: Presentation of an algorithm for microsurgical breast reconstruction based on perforator anatomy. Plast Reconstr Surg. 2007;119:18–27.
8. Kamali P, Lee M, Becherer BE, et al. Medial row perforators are associated with higher rates of fat necrosis in bilateral DIEP flap breast reconstruction. Plast Reconstr Surg. 2017;140:19–24.
9. Li L, Chen Y, Chen J, et al. Adjuvant chemotherapy increases the prevalence of fat necrosis in immediate free abdominal flap breast reconstruction. J Plast Reconstr Aesthet Surg. 2014;67:461–467.
10. Douglas HE, Wilkinson MJ, Mackay IR. Effects of perforator number and location on the total pedicle flow and perfusion of zone IV skin and fat of DIEP flaps. J Plast Reconstr Aesthet Surg. 2014;67:212–218.
11. Rozen WM, Ashton MW, Murray AC, Taylor GI. Avoiding denervation of rectus abdominis in DIEP flap harvest: The importance of medial row perforators. Plast Reconstr Surg. 2008;122:710–716.
12. Rozen WM, Ashton MW, Kiil BJ, et al. Avoiding denervation of rectus abdominis in DIEP flap harvest II: An intraoperative assessment of the nerves to rectus. Plast Reconstr Surg. 2008;122:1321–1325.
13. Rozen WM, Tran TM, Barrington MJ, Ashton MW. Avoiding denervation of the rectus abdominis muscle in DIEP flap harvest III: A functional study of the nerves to the rectus using anesthetic blockade. Plast Reconstr Surg. 2009;124:519–522.
14. Garvey PB, Salavati S, Feng L, Butler CE. Abdominal donor-site outcomes for medial versus lateral deep inferior epigastric artery branch perforator harvest. Plast Reconstr Surg. 2011;127:2198–2205.
15. Selber JC, Kurichi JE, Vega SJ, Sonnad SS, Serletti JM. Risk factors and complications in free TRAM flap breast reconstruction. Ann Plast Surg. 2006;56:492–497.
16. Lee KT, Mun GH. Effects of obesity on postoperative complications after breast reconstruction using free muscle-sparing transverse rectus abdominis myocutaneous, deep inferior epigastric perforator, and superficial inferior epigastric artery flap: A systematic review and meta-analysis. Ann Plast Surg. 2016;76:576–584.
17. Wormer BA, Clavin NW, Lefaivre JF, et al. Reducing postoperative abdominal bulge following deep inferior epigastric perforator flap breast reconstruction with onlay monofilament poly-4-hydroxybutyrate biosynthetic mesh. J Reconstr Microsurg. 2017;33:8–18.
18. Butler PD, Wu LC. Abdominal perforator vs. muscle sparing flaps for breast reconstruction. Gland Surg. 2015;4:212–221.
19. Uda H, Tomioka YK, Sarukawa S, Sunaga A, Sugawara Y. Comparison of abdominal wall morbidity between medial and lateral row-based deep inferior epigastric perforator flap. J Plast Reconstr Aesthet Surg. 2015;68:1550–1555.
20. Mirzabeigi MN, Wilson AJ, Fischer JP, et al. Predicting and managing donor-site wound complications in abdominally based free flap breast reconstruction: Improved outcomes with early reoperative closure. Plast Reconstr Surg. 2015;135:14–23.
21. Wan DC, Tseng CY, Anderson-Dam J, Dalio AL, Crisera CA, Festekjian JH. Inclusion of mesh in donor-site repair of free TRAM and muscle-sparing free TRAM flaps yields rates of abdominal complications comparable to those of DIEP flap reconstruction. Plast Reconstr Surg. 2010;126:367–374.
22. Griffiths M, Chae MP, Rozen WM. Indocyanine green-based fluorescent angiography in breast reconstruction. Gland Surg. 2016;5:133–149.
23. Duggal CS, Madni T, Losken A. An outcome analysis of intraoperative angiography for postmastectomy breast reconstruction. Aesthet Surg J. 2014;34:61–65.
24. Komorowska-Timek E, Gurtner GC. Intraoperative perfusion mapping with laser-assisted indocyanine green imaging can predict and prevent complications in immediate breast reconstruction. Plast Reconstr Surg. 2010;125:1065–1073.
25. Pestana IA, Coan B, Erdmann D, Marcus J, Levin LS, Zenn MR. Early experience with fluorescent angiography in free-tissue transfer reconstruction. Plast Reconstr Surg. 2009;123:1239–1244.
26. Pestana IA, Zenn MR. Correlation between abdominal perforator vessels identified with preoperative CT angiography and intraoperative fluorescent angiography in the microsurgical breast reconstruction patient. Ann Plast Surg. 2014;72:S144–S149.
27. Deutsch MF, Smith M, Wang B, Ainsle N, Schusterman MA. Immediate breast reconstruction with the TRAM flap after neoadjuvant therapy. Ann Plast Surg. 1999;42:240–244.
28. Song J, Zhang X, Liu Q, et al. Impact of neoadjuvant chemotherapy on immediate breast reconstruction: A meta-analysis. PLoS One 2014;9:e98225.
29. Monrigal E, Dauplat J, Gimbergues P, et al. Mastectomy with immediate breast reconstruction after neoadjuvant chemotherapy and radiation therapy: A new option for patients with operable invasive breast cancer. Results of a 20 years single institution study. Eur J Surg Oncol. 2011;37:864–870.
30. Schaverien MV, Munnoch DA. Effect of neoadjuvant chemotherapy on outcomes of immediate free autologous breast reconstruction. Eur J Surg Oncol. 2013;39:430–436.
31. Azzawi K, Ismail A, Earl H, Forouhi P, Malata CM. Influence of neoadjuvant chemotherapy on outcomes of immediate breast reconstruction. Plast Reconstr Surg. 2010;126:1–11.
32. Decker MR, Greenblatt DY, Havlena J, Wilke LG, Greenberg CC, Neuman HB. Impact of neoadjuvant chemotherapy on wound complications after breast surgery. Surgery 2012;152:382–388.
33. Mehrara BJ, Santoro TD, Arcilla E, Watson JP, Shaw WW, Da Lio AL. Complications after microvascular breast reconstruction: Experience with 1195 flaps. Plast Reconstr Surg. 2006;118:1100–1109; discussion 11101111.
©2018American Society of Plastic Surgeons
34. Cho MJ, Teotia SS, Haddock NT. Predictors, classification, and management of umbilical complications in DIEP flap breast reconstruction. Plast Reconstr Surg. 2017;140:11–18.