OMAC

Tcsd(a

CTgÖ

R1FOtCtHo

©P

utcomes of Immediate Vertical Rectus Abdominisyocutaneous Flap Reconstruction for Irradiatedbdominoperineal Resection Defects

harles E Butler, MD, FA C S , A Özlem Gündeslioglu, MD, Miguel A Rodriguez-Bigas, MD, FA C S

BACKGROUND: Perineal wound complications after chemoradiotherapy and abdominoperineal resection (APR)for anorectal cancer occur in up to 60% of patients, including perineal abscess and wounddehiscence. Ver tical rectus abdominis myocutaneous ( VRAM) flaps have been used in anattempt to reduce these complications by obliterating the noncollapsible dead space withvascularized tissue and closing the perineal skin defect with nonirradiated flap skin. Manysurgeons are reluctant to use VRAM flaps unless primar y closure is not possible.

STUDY DESIGN: All patients who under went chemoradiotherapy and APR during a 12-year period at theUniversity of Texas MD Anderson Cancer Center were retrospectively revie wed. Patient, tumor,and treatment characteristics and surgical complications and outcomes were compared betweenpatients who under went VRAM flap reconstr uction of wounds that could have been closedprimarily (flap group, n � 35) and those who had primar y closure of the perineal wound(control group, n � 76).

RESULTS: Overall, there were no significant differences in the incidence of perineal wound complicationsbetween the groups; the flap group had a significantly lower incidence of perineal abscess (9%versus 37%, p � 0.002), major perineal wound dehiscence (9% versus 30%, p � 0.014), anddrainage procedures required for perineal/pelvic fluid collections (3% versus 25%, p � 0.003)than the control group had. Despite flap har vest and the need for donor site closure in the flapgroup, there was no significant difference in abdominal wall complications between groupsduring the study’s mean patient followup of 3.8 years.

CONCLUSIONS: VRAM flap reconstr uction of irradiated APR defects reduces major perineal wound complica-tions without increasing early abdominal wall complications. Strong consideration should begiven to immediate VRAM flap reconstr uction after chemoradiation and APR. ( J Am Coll Surg

2008;206:694–703. © 2008 by the American College of Surgeons)

w2tpot

etclpedion

he standard treatment for locally advanced rectal adeno-arcinoma is neoadjuvant chemoradiotherapy followed byurgical resection. If the anal sphincter cannot be preser veduring tumor extirpation, an abdominoperineal resectionAPR) is per formed. When APR follows chemoradiother-py, perineal wound complications, including abscess,

ompeting Interests Declared: None.his study was funded in par t by the Shor t-Term Scientist Exchange Pro-ram, National Institutes of Health, National Cancer Institute (to Dr Azlem Gündeslioglu).

eceived August 21, 2007; Revised December 10, 2007; Accepted December0, 2007.rom the Departments of Plastic Surgery (Butler, Gündeslioglu) and Surgicalncology (Rodriguez-Bigas), University ofTexas MD Anderson Cancer Cen-

er, Houston TX.orrespondence address: Charles E Butler, MD, FACS, Department of Plas-

ic Surgery, Unit 443, University ofTexas MD Anderson Cancer Center, 1515olcombe Blvd, Houston, T X 77030. email: cbutler@mdanderson.x

prg

6942008 by the American College of Surgeons

ublished by Elsevier Inc.

ound dehiscence, and delayed wound healing occur in5% to 60% of patients.1-7 These complications are relatedo the presence of a large, noncollapsible dead space,8-10 theoor vascularity of the irradiated surrounding tissue,11 usef irradiated skin in the closure, and bacterial contamina-ion owing to bowel resection.12,13

Local-regional tissue flaps can provide bulky tissue to oblit-rate dead space, bring vascularized tissue to the irradiatedumor bed, and provide a skin paddle for cutaneous woundlosure.4-7,10,14-21 Flap reconstr uction is well accepted for ver yarge perineal defects, such as when the skin cannot be closedrimarily or a massive dead space is present (eg, after pelvicxenteration). The role of “routine” flap use when a skin pad-le is not absolutely essential for cutaneous closure of the per-

neal wound or when a large amount of tissue bulk is notbviously needed to obliterate a massive pelvic dead space hasot been established. Previous studies demonstrating im-

roved perineal wound healing after APR or pelvic exentera-

ISSN 1072-7515/08/$34.00doi:10.1016/j.jamcollsurg.2007.12.007

tofitopt

cr(atnasct

fAcwpbcVtt

MPA(co1Waiac

wieotpwtpsweA

raats

hfucct

STappmpatwttitiopslfpart

695Vol. 206, No. 4, April 2008 Butler et al VRAM Flap Perineal Reconstruction

ion with use of flaps have been limited by their small numberf patients,5,17,19 absence of a control population (similar de-ects closed primarily),7,22 the heterogeneity of perineal defectsncluded,5,15,21 and the heterogeneity of flap reconstructionechniques used.19,23-25 The current study directly comparedutcomes of patients with irradiated APR defects who hadrimary closure of the perineal wound versus flap reconstruc-ion with a standardized technique.

Commonly used flaps for perineal reconstruction in-lude the greater omentum, gracilis myocutaneous, poste-ior thigh, and vertical rectus abdominis myocutaneousVRAM) flaps. VRAM flaps generally have greater bulknd a more reliable vascular supply to the skin paddle thanhigh-based flaps.23,26 The greater omentum is frequentlyot available or usable, particularly in cases of previousbdominal operations, and does not provide vascularizedkin for perineal wound closure. The VRAM flap has be-ome the most commonly used technique at our institu-ion for primary flap reconstruction after APR.

We hypothesized that patients who underwent VRAMlap reconstruction after preoperative chemoradiation andPR for rectal or anal carcinoma would have fewer woundomplications than patients with similar resection defectsho had primary perineal wound closure. We also antici-ated that there might be increased abdominal morbidityecause of flap harvest. In this study, we retrospectivelyompared abdominal and perineal wound outcomes ofRAM flap reconstruction versus primary closure in pa-

ients who underwent APR after neoadjuvant chemoradio-herapy for rectal or anal malignancies.

ETHODSatientsll patients with persistent anal squamous cell carcinoma

SCC) or rectal adenocarcinoma who had preoperativehemoradiotherapy and underwent APR at the Universityf Texas MD Anderson Cancer Center between September, 1993, and August 31, 2005, were evaluated for the study.e retrospectively reviewed prospectively gathered data in

n electronic database and patients’ medical records todentify patients who had a perineal wound where the skinnd soft tissues could have been closed primarily but was

Abbreviations and Acronyms

APR � abdominoperineal resectionIORT � intraoperative radiotherapySCC � squamous cell carcinomaVRAM � vertical rectus abdominis myocutaneous

losed with a VRAM flap. Control patients were those p

hose perineal wounds were closed primarily. Patients werencluded in both groups when their resections included anlliptical resection of a small (� 30% surface area) sectionf the posterior vaginal wall that was closed primarily. Pa-ients treated with total or modified pelvic exenteration orartial (� 30%) or complete vaginectomy and patientshose tissue defects required flap reconstruction to close

he perineal skin defect (ie, the defect could not be closedrimarily) were excluded. Patients who underwent recon-truction with rectus abdominis muscle-only flaps orhose reconstruction included a thigh-based flap were also

xcluded. This study was reviewed and approved by MDnderson’s institutional review board.Data extracted from the databases and medical

ecords included patient and tumor characteristics, neo-djuvant and adjuvant treatment, closure technique,nd surgical wound healing outcomes and complica-ions (abdominal wall and perineal), length of hospitaltay, and postoperative course.

Patients were divided into two subgroups: flap patientsad undergone immediate reconstruction of the APR de-ect with a pedicled VRAM flap, and nonflap patients hadndergone primary closure of the perineal wound. Thehoice of VRAM flap reconstruction had been based prin-ipally on surgeon preference for preoperative referral tohe plastic surgical service at MD Anderson.

urgical techniquehe reconstructive technique for flap patients was similar

mong all reconstructive surgeons and has been describedreviously (Figs. 1-3).27 Briefly, a 5- to 10-cm�wide skinaddle was designed vertically above the right rectus abdo-inis muscle unless previous scarring or an ostomy was

resent, in which case a left VRAM flap was used. Medialnd lateral rows of myocutaneous perforating vessels andhe intervening anterior rectus sheath fascia were includedith the flap. The inferior epigastric pedicle was mobilized

o its origin and separated from all surrounding tissue at-achments to prevent kinking or compression. The musclensertion on the pelvis was left intact to prevent tension onhe pedicle. The flap was rotated medially and transposednto the perineal defect to obliterate the maximum amountf dead space possible and to prevent tension on, and com-ression of, the flap and vascular pedicle. An elliptical-haped skin paddle was marked while the patient was in theithotomy position with legs adducted to prevent excessivelap skin paddle redundancy postoperatively. The flap skineripheral to the marked skin paddle was deepithelializednd sutured to the remnant pelvic floor edges with inter-upted permanent sutures. Closed-suction drainage cathe-ers were placed to drain the pelvic space. The flap skin

addle was then inset into the skin defect edges.

cwola

ihps

ppliis

CW

ity of

696 Butler et al VRAM Flap Perineal Reconstruction J Am Coll Surg

The abdominal donor site was subsequently closed afterontralateral colostomy placement. The fascia was closedith interrupted or running polypropylene sutures with-ut prosthetic mesh. The ipsilateral skin was underminedateral to the donor site, advanced medially, and closed overclosed-suction drain.Nonflap patients underwent layered closure of their per-

neal defects with linear closure of the skin edges. They alsoad closed-suction drainage catheters placed to drain theelvic dead space, but they did not undergo abdominal

Figure 1. Illustration of a vertical rectus abdominis myocutaneousoriented skin paddle overlying the rectus abdominis muscle. (B) The fland one rectus abdominis muscle. (C) The flap is transposed througclose the perineal skin defect) is marked. The skin is deepithelializedto the perineal skin defect. The colostomy is brought through the conincision are closed primarily. (From: Visual Art, Houston, TX, Univers

kin undermining or subcutaneous drain placement. c

In both groups, postoperative positioning included aillow placed under the sacrum to limit pressure on theerineal wound during hospitalization. Patients were al-

owed to walk on the first postoperative day. Patients werenstructed to put their weight on the lateral gluteal andschial regions (not directly on the perineal wound) whileitting for 2 weeks postoperatively.

omplications and followupe extracted data on perineal and abdominal wall compli-

arvest and inset after abdominoperineal resection. (A) A verticallyludes skin, subcutaneous fat, a cuff of anterior rectus sheath fascia,pelvis and the size, shape, and location of the skin paddle (used toounding the marked skin paddle. (D) The flap skin paddle is suturedteral side, and the fascia and skin of the donor site and laparotomyTexas MD Anderson Cancer Center, 2007, with permission.)

flap hap inch the

surrtrala

ations to compare their occurrence in the two groups.

Dtgdwiwomatwaswputsfwrq

w

pdc

SWpperVscitiop

eoptpfwsgptttoclrtdw

RPAirrsrt

FaspadTp

697Vol. 206, No. 4, April 2008 Butler et al VRAM Flap Perineal Reconstruction

ata were recorded from the medical record and a prospec-ively maintained database. Wound dehiscences were cate-orized as major or minor. A major wound dehiscence wasefined as skin and subcutaneous tissue breakdown with orithout deep infection requiring additional intervention,

ncluding debridement or drainage and irrigation. A minoround dehiscence was a small cutaneous dehiscence withr without cellulitis requiring only dressing changes. He-atomas and seromas that required drainage were included

s complications. Pelvic and deep perineal abscess collec-ions were also considered complications. Other perinealound complications included partial or complete flap loss

nd perineal hernia, which was defined as a major protru-ion of abdominal viscera through the perineal pelvic floorith a visible and palpable bulge. Abdominal wall andarastomal hernias were defined as a palpable bulge with annderlying fascial defect, which could also be palpated byhe surgeon or identified on followup CT. Abdominal wallkin dehiscence was defined as wound breakdown withull-thickness skin separation for a distance � 2 cm with orithout infection. Fascial dehiscence was defined as sepa-

ated fascial edges visualized through an open wound re-uiring reoperation.

Patients were followed up every 1 to 2 weeks until all

igure 2. A sagittal cross-section illustration of the vertical rectusbdominis myocutaneous flap inset. The well-vascularized flap tis-ue occupies and obliterates much of the irradiated, noncollapsibleelvic/perineal dead space created by removal of the rectum andnus. The tissue bulk reduces fluid collection within and intestinalescent into the pelvic/perineal space. (From: Visual Art, Houston,X, University of Texas MD Anderson Cancer Center, 2007, withermission.)

ounds were healed and then every 3 to 6 months. The p

ercentage of patients alive or dead at last followup wasetermined. Time to death for patients who died was cal-ulated from the date of APR to death.

tatistical analysise assessed the comparability of the two groups in terms of

atient-, tumor-, and treatment-related factors. An un-aired, two-tailed Student’s t-test was used to assess differ-nces in continuous variables, such as age, body mass index,adiation dose, hospitalization time, and followup time.alues are expressed as mean and standard deviation. Chi-

quare or Fisher’s exact test was used to assess differences inategorical variables, such as gender, comorbid conditions,ntraoperative radiotherapy (IORT), recurrence, and par-ial vaginectomy. We then compared the incidence of per-neal and abdominal wall complications using chi-squarer Fisher’s exact test. We also assessed the impact of com-lications on hospital stay using Student’s t-test.Patient-, tumor-, and treatment-related factors were then

valuated to identify any independent predictors of the devel-pment of perineal or abdominal wall complications. Com-lication end points modeled were any perineal complica-ion, any major perineal complication (major dehiscence,erineal abscess, or need for an invasive drainage procedureor a perineal/pelvic fluid collection), and any abdominalall complication. Potential predictive factors used in the

tepwise logistic regression model included patient age,ender, body mass index, tobacco use, use of IORT,erformance of partial vaginectomy, tumor stage, tumorype (anal SCC or rectal adenocarcinoma), recurrentumor, use of a VRAM flap, and reported history ofhese comorbid conditions: diabetes mellitus, chronicbstructive pulmonary disease, congestive heart failure,oronary artery disease, hypertension, peripheral vascu-ar disease, renal insufficiency, hematologic disease,heumatologic disease, and endocrine disease. Odds ra-io estimates and 95% Wald CI were calculated for pre-ictive factors. In all statistical analyses, p values � 0.05ere considered statistically significant.

ESULTSatient, tumor, and treatment characteristicstotal of 111 patients were included in the study. Tumors

n the flap group consisted of 30 (24 primary and 6 recur-ent) rectal adenocarcinomas and 5 (1 persistent and 4ecurrent) anal SCCs. Tumors in the nonflap group con-isted of 75 primary rectal adenocarcinomas and 1 recur-ent anal SCC. A considerably greater proportion of pa-ients in the flap group had anal SCC, recurrent tumors,

artial vaginectomy, and IORT (Table 1).

av(e

tttp

have

698 Butler et al VRAM Flap Perineal Reconstruction J Am Coll Surg

There were no substantial differences between the flapnd nonflap groups in mean patient age (54.3 � 13.8 yearsersus 56.2 � 13.6 years, respectively) or body mass index26.5 � 6.5 versus 28.5 � 5.8). There were also no differ-

Figure 3. Photographs of a vertical rectus abdominis myocutaneous fland shape of the skin island that will be interposed between the irradiaperipheral flap skin has been deepithelialized. (C) The skin paddle hasof the healed perineal skin 3 weeks postoperatively, after the sutures

nces between the groups in tobacco use, tumor stage, or e

he incidence of any comorbid condition (diabetes melli-us, coronary artery disease, congestive heart failure, hyper-ension, peripheral vascular disease, chronic obstructiveulmonary disease, renal insufficiency, hematologic dis-

set. (A) Perineal defect after abdominoperineal resection. (B) The sizeerineal skin defect edges have been marked, and the remainder of theinset into the perineal skin defect. (D) Inferior and (E) posterior viewsbeen removed. (Photographs courtesy of Charles E Butler, MD.)

ap inted pbeen

ase, rheumatologic disease, or endocrine disease). There

w2

bGIoi

POpncfdlppvmws

ciadncnhc

mc

AOd

T

C

P

A

T

C

ARPI

T

R

P*I

699Vol. 206, No. 4, April 2008 Butler et al VRAM Flap Perineal Reconstruction

as a greater proportion of female patients in the flap (n �8; 80%) than nonflap (n � 23; 30%) group (p � 0.001).

The preoperative radiation dose did not differ substantiallyetween the flap (52.2 � 12.5 Gy) and nonflap (49.2 � 3.9y) groups. Ten patients (29%) in the flap group received

ORT (mean 12 Gy; range 10 to 15 Gy) in addition to pre-perative external-beam radiotherapy, but none of the patientsn the nonflap group received IORT (p � 0.001; Table 1).

erineal complicationsverall, there was no difference in the incidence of any

erineal wound complication between the flap (46%) andonflap (46%) groups. Fewer patients had severe compli-ations in the flap group. Incidence of perineal abscessormation (9% versus 37%, p � 0.002) and major woundehiscence (9% versus 30%, p � 0.014) was considerably

ower in the flap than the nonflap group (Table 2). Feweratients required drainage procedures to treat perineal/elvic fluid collections (seroma or abscess) in the flap (3%)ersus nonflap (25%) group (p � 0.003). The incidence ofinor perineal complications—hematoma, seroma, minoround dehiscence, and perineal hernia—did not differ

ubstantially between groups.In both groups, the majority of perineal wound dehis-

ences were treated with wet-to-moist saline-soaked dress-ng changes, with or without bedside debridement. Oper-tive debridement of the perineal wound followed byressing changes was performed in 8 patients (11%) in theonflap group and 3 patients (9%) in the flap group, in-luding 1 patient with complete flap necrosis. There wereo partial flap failures in the flap group, but one patientad necrosis leading to complete flap loss. The patient with

able 1. Patient, Tumor, and Resection Defect Characterist

haracteristicFlap group (n � 35)

n %

nal SCC 5 14ecurrent tumor 9 26artial vaginectomy 19 54ORT 10 29

Rectal Analn % n

umor stageI 2 6 0II 13 37 0III 7 20 1IV 2 6 0

ecurrent 6 17 4 1

athologic stage of primary tumors based on American Joint Committee onComparison of tumor stage distribution between flap and nonflap group.ORT, intraoperative radiotherapy; SCC, squamous cell carcinoma.

omplete flap necrosis and loss underwent flap debride-

ent and subsequent reconstruction with bilateral pedi-led gracilis myocutaneous flaps.

bdominal wall complicationsverall, there was no substantial difference in the inci-

ence of any abdominal wall complication between the flap

able 2. Perineal and Abdominal Wound Complications

omplication

Flapgroup

(n � 35)

Nonflapgroup

(n �76)p Valuen % n %

erineal complicationsSeroma 0 5 7 0.177Hematoma 0 3 4 0.550Perineal abscess 3 9 28 37 0.002Major wound dehiscence 3 9 23 30 0.014Minor would dehiscence 9 26 9 12 0.065Perineal hernia 2 6 4 5 1Complete flap loss 1 3 0 0.2Any perineal complication 16 46 35 46 0.97

bdominal wallcomplications

Seroma 0 0 —Hematoma 0 0 —Infection 0 1 1 1Skin dehiscence 4 11 11 14 0.772Fascial dehiscence 2 6 2 3 0.589Delayed healing 1 3 8 11 0.276Incision hernia 2 6 6 8 1Parastomal hernia 4 11 12 16 0.772Any abdominal wall

Nonflap group (n � 76)p Valuen %

1 1 � 0.011 1 � 0.0014 5 � 0.0010 � 0.0001

Rectal Analn % n %

1.0*1 1 0

45 59 025 33 04 5 00 1 1 � 0.00003

r system (6th ed.).28

ics

%

3

1

Cance

complication 10 29 25 33 0.649

(eontie

IOs3am0ppvdp1w

FNtdf(dmcpVd(ao

FTb2nAhef

DTi

racaai

wrvsspvfcrecsabclrd

rddpwfascfr2ih

ssttaprs

700 Butler et al VRAM Flap Perineal Reconstruction J Am Coll Surg

29%) and nonflap (33%) groups, nor were there differ-nces in specific complications (Table 2). Despite harvestf skin, subcutaneous fat, rectus fascia, and rectus abdomi-is muscle from the abdominal wall for the VRAM flaps,he incidence of abdominal skin or fascial dehiscence andncisional or parastomal hernia was not substantially differ-nt in the flap group.

mpact of complications on hospitalizationverall, there was no substantial difference in mean hospital

tay between the flap (9.1 � 7.2 days) and nonflap (10.2 �.8 days) groups. The occurrence of at least 1 postoperativebdominal or perineal wound complication increased theean hospital stay by 80.4% (p � 0.039) and 91.3% (p �

.0002) for flap and nonflap patients, respectively. Mean hos-ital stay did not differ substantially between flap and nonflapatients who had a postoperative wound complication de-elop. Mean hospital stay was 9.3 � 3.4 days versus 7.4 � 3.0ays for patients in the flap versus nonflap group with noostoperative complications and 16.8 � 13.9 days versus4.2 � 10.0 days for those in the flap versus nonflap groupho had at least 1 postoperative complication.

actors predictive of complicationso patient-, tumor-, or treatment-related factor was found

o be predictive for the development of a perineal or ab-ominal wall complication. Predictive factors were identi-ied when development of a major perineal complicationmajor dehiscence, perineal abscess, or need for an invasiverainage procedure for perineal/pelvic fluid collection) wasodeled. The odds ratio of developing a major perineal

omplication was 5.5 (95% CI, 1.5 to 16.2; p � 0.001) foratients who had 2 or more comorbid conditions. Use of aRAM flap was found to be protective: the odds ratio forevelopment of a major perineal complication was 0.3295% CI, 0.11 to 0.9; p � 0.03) for patients who receivedVRAM flap. No other factors were found to be predictiver protective for major perineal complications.

ollowup and overall survivalhere was no substantial difference in mean followup timeetween the flap (3.2 � 2.3 years) and nonflap groups (4.2 �.7 years). At last followup, 71% of flap patients and 70% ofonflap patients were alive; the difference was not substantial.mong the patients who died during the followup period andad at least 1 postoperative complication, there was no differ-nce in the mean time to death after operation between thelap (4.3 � 0.9 years) and nonflap (3.6 � 2.5 years) groups.

ISCUSSIONhis study demonstrates that VRAM flap reconstruction of

rradiated APR defects in patients with anal or rectal cancer t

educes the incidence of major perineal wound dehiscencend perineal abscess without increasing abdominal woundomplications, as compared with nonflap closure. Obliter-tion of perineal and pelvic dead space with a VRAM flaplso reduces the need for drainage procedures to treat per-neal abscess or seroma.

Previous studies have demonstrated reduced perinealound complications using flap reconstruction of perineal

esection defects.5,7,14-19 Perineal defects resulting from pel-ic exenteration, sacrectomy, and recurrent anal cancer re-ections are frequently large and often require flap recon-truction to close the perineal skin defect, obliterate theelvic space, or reconstruct a new vagina after completeaginectomy. In the current study, perineal defects wererom an APR (with or without partial vaginectomy) andould all be closed primarily; patients whose defects were aesult of large skin or multivisceral resections or pelvic ex-nteration were excluded. In the patients studied, flap re-onstruction was not absolutely required for wound clo-ure, so the nonflap group served as a relevant andppropriate control group with similar defects. Relativeenefits and disadvantages of VRAM flap reconstructionould be accurately assessed. To our knowledge, this is theargest controlled study that directly compares VRAM flapeconstruction with primary closure of irradiated perinealefects after APR for anal or rectal cancer.The perineal complication rates in both groups were

elatively high, confirming the morbidity of APR after ra-iotherapy. The overall perineal wound complication rateid not differ substantially between the flap and nonflapatients. However, rate of major perineal complicationsas substantially higher in the nonflap group, even though

lap patients had greater incidences of recurrent tumors,nal cancers, IORT, and partial vaginectomy with the re-ection. Despite a potential bias toward larger or moreomplex perineal defects in flap patients, use of VRAMlaps reduced major perineal wound complications. Thiseduction occurred without an increase in early (3.2 years �.3 years followup) abdominal wall complications (includ-ng incisional and parastomal hernias) resulting from flaparvest.Our findings are in agreement with previous smaller

tudies in which VRAM flaps were used for perineal recon-truction. Major perineal complications ranging from 25%o 46% have been reported when irradiated perineal resec-ion defects are closed without flap reconstruction,1,3-5,7 butcute perineal wound complication rates have been re-orted to range from 3.5% to 15.8% after VRAM flapeconstruction.7,18 Despite removal of abdominal skin,ubcutaneous fat, anterior rectus sheath fascia, and the rec-

us abdominis muscle with VRAM flap harvest, the early

atwp

rconlw(vltwo

soaisfqcavrictitarbrcoatdicnsna

o

vlcscttaamsqhpcfmi

matwWtcjctrtmifpcbdrsglfascs

ra

701Vol. 206, No. 4, April 2008 Butler et al VRAM Flap Perineal Reconstruction

bdominal wall complication rates did not differ betweenhe flap and nonflap groups. This finding is in agreementith a smaller previous study involving 19 VRAM flapatients.28

Unlike previous studies, the longer mean followup pe-iod of � 3 years in our study allowed the evaluation ofomplications that would be expected to occur later afterperation, such as incisional, parastomal, or perineal her-ia. Certainly, hernias can occur during additional fol-

owup. Despite flap harvest from the abdominal wall, thereas a trend toward a lower incidence of early abdominal

incisional or parastomal) hernia in the flap group (17%ersus 24%), but the difference was not significant. Theower abdominal wall complication rate further supportshe relatively low morbidity of VRAM flap harvest, evenhen the abdominal wall donor site is followed for a meanf � 3 years postoperatively.

Although the VRAM flap is well suited for perineal recon-truction, there are other flap options, including greatermental and thigh-based flaps (eg, gracilis, posterior thigh,nd anterolateral thigh flaps). An advantage of the VRAM flapncludes its large, reliable skin paddle for perineal skin recon-truction or vaginal reconstruction, if required. The VRAMlap also has excellent bulk for dead space obliteration, ade-uate reach to the perineum and sacrum, reliable vascularity,onsistent anatomy, relative ease of harvest, and no need fordditional scars in a separate donor area. Potential disad-antages of the VRAM flap are generally related to theemoval of skin, fascia, and rectus muscle from the abdom-nal wall. Although not observed in this study, theoreticonsequences of VRAM harvest include greater tension onhe abdominal skin and fascial closure, potentially increas-ng the risk of wound separation.17,29-31 Absence of one ofhe two rectus abdominis muscles in the abdominal walllso has implications for ostomy placement and potentialeplacement in the future.26 Generally, a colostomy isrought out through the left rectus muscle, and the rightectus muscle is used for VRAM flap reconstruction. Oc-asionally, a second ostomy is required during a subsequentperation, such as an ileal conduit for urinary diversionfter salvage total pelvic exenteration or colostomy reloca-ion for treatment of parastomal hernia. If one rectus ab-ominis muscle has been used for flap reconstruction, there

s no intact, virgin rectus muscle through which the ostomyan be placed. In either situation, we typically place theew ostomy through the abdominal wall on the flap donoride, lateral to the empty rectus sheath. Although this tech-ique has not been formally studied, we have not observedn increase in ostomy-related complications using it.

The advantages of flap reconstruction might be more

bvious for larger defects, such as those resulting from pel- s

ic exenteration and APR, with extended skin resection forocally advanced anal SCC or with sacrectomy for rectalancer invasion into the sacrum. Previous studies involvinguch cases demonstrated relatively few perineal woundomplications and few hospital readmissions and reopera-ions for wound healing complications by using flap tissueo reconstruct the perineal defect.7,15,17,21,32 Although thedvantages of flap reconstruction for standard APRs, suchs in our study, might seem less striking than for massive,ultivisceral resections, the overall impact on patients and

urgeons can be momentous, owing to the relative fre-uency with which these procedures are performed and theigh incidence and morbidity of associated perineal com-lications. Although most irradiated APR defects can belosed primarily without flap reconstruction, use of VRAMlaps in our study was associated with a 4-fold reduction inajor perineal wound dehiscence and a 10-fold reduction

n perineal abscess formation.Although there was a substantial overall reduction inajor perineal wound morbidity without increased early

bdominal wall morbidity in the flap group, we recognizedhat the flap patients might not have encountered perinealound complications at all if they had not received a flap.e used logistic regression to identify any individual fac-

ors predictive of any perineal or abdominal wall compli-ation. We found no such factors. When modeled for ma-or perineal complications, patients with two or moreomorbid conditions were more than five times more likelyo have a major perineal complication develop (one thatequired invasive treatment). Even more interesting, pa-ients receiving VRAM flaps were a third as likely to have aajor perineal complication develop. Unfortunately, this

nformation is not especially useful for selecting patientsor treatment. The conundrum of factors that truly renderatients to be “high risk” for perineal complications is notlearly elucidated. Patients truly at high risk will clearlyenefit from VRAM flap reconstruction. It is less clear theegree of benefit that patients not truly at high risk willeceive from a VRAM flap. A greater number, but notpecific type, of comorbid conditions can indicate poorereneral health, but such a criterion does not directly trans-ate to specific indications for using or not using a VRAMlap to reconstruct APR defects. Future prospective studiesre needed to identify factors and conditions that can helpelect the patients who are truly at high risk for perinealomplications and would benefit the most from flap recon-truction in this setting.

The strengths of this study include strict inclusion crite-ia to maximize the perineal defect similarity, an appropri-te control group, a relatively large number of patients,

imilar extirpative and reconstructive techniques used for

atsadmhphdttStpst

cuaphaesputncaoecte

Awttfw

A

SAA

DC

ACSatCDcDt

R

1

1

1

1

702 Butler et al VRAM Flap Perineal Reconstruction J Am Coll Surg

ll patients, and a � 3-year mean followup period. Despitehe strict inclusion and exclusion criteria, this retrospectivetudy has some limitations. Patients were not randomlyssigned to receive a VRAM flap or primary closure; theecision to perform flap reconstruction, rather than pri-ary closure, was based on surgeon preference. This might

ave resulted in some selection bias: the decision to refer aatient preoperatively to the plastic surgery service mightave been because of the surgeon’s perception that theefect would be more difficult to heal or of greater magni-ude. This supposition is supported by the observationshat patients in the flap group had considerably more analCCs, recurrent tumors, partial vaginectomies, and IORThan those in the nonflap group. This potential bias towardotentially greater wound severity in the flap group furtherupports the benefits of VRAM reconstruction even whenhe perineal defect can be closed primarily.

There is a trend at our institution, and other majorancer centers nationwide, to refer patients who will bendergoing APR to the plastic surgery service preoper-tively, regardless of the nature or magnitude of thelanned resection. This is likely largely a result of theigh overall perineal wound complication rates observedfter APR and the commitment at our center to preop-ratively coordinate multidisciplinary cases with recon-tructive surgeons interested and experienced in theseatients. We strongly believe that patients scheduled tondergo APR or pelvic exenteration benefit from mul-idisciplinary evaluation and treatment from a coordi-ated team, including medical oncologists, surgical on-ologists, radiation oncologists, enterostomal therapists,nd reconstructive surgeons. Urologists and gynecologicncologists are also included in the team when pelvicxenteration is considered. Preoperative planning andommunication between members of the treatmenteam are critical to optimize patient education, operativefficiency, and patient outcomes.

In conclusion, VRAM flap reconstruction of irradiatedPR defects reduces major perineal wound complicationsithout increasing early abdominal wall morbidity. Addi-

ional prospective studies will be important to develop cri-eria to identify the patients who will benefit most fromlap reconstruction and evaluate abdominal wall morbidityith a longer followup.

uthor Contributions

tudy conception and design: Butlercquisition of data: Butler, Gündeslioglunalysis and interpretation of data: Butler, Gündeslioglu,

Rodriguez-Bigas

rafting of manuscript: Butler, Gündesliogluritical revision: Butler, Gündeslioglu, Rodriguez-Bigas

cknowledgment: We thank Drs Elisabeth Beahm, Davidhang, Pierre Chevray, Steven Kronowitz, Michael Miller,cott Oates, Gregory Reece, Geoffrey Robb, Roman Skoracki,nd Ron Yu from the Department of Plastic Surgery for con-ributing reconstructive cases and Drs John Skibber, Georgehang, Steven Curley, Barry Feig, and Kelly Hunt from theepartment of Surgical Oncology for contributing extirpative

ases to this study. We also thank Joe Ensor, PhD, from theivision of Quantitative Sciences for assistance with the sta-

istical analysis.

EFERENCES

1. Touran T, Frost DB, O’Connel TX. Sacral resection: operativetechnique and outcome. Arch Surg 1990;125:911–913.

2. Arnold PG, Lovich SF, Pairolero PC. Muscle flaps in irradiatedwounds: an account of 100 consecutive cases. Plast Reconsr Surg1994;93:324–327.

3. McAllister E, Wells K, Chaet M, et al. Perineal reconstructionafter surgical extirpation of pelvic malignancies using thetranspelvic transverse rectus abdominis musculocutaneous flap.Ann Surg Oncol 1994;1:164–168.

4. Yeh KA, Hoffman JP, Kusiak JE, et al. Reconstruction withmyocutaneous flaps following resection of locally recurrent rec-tal cancer. Am Surg 1995;61:581–589.

5. Shibata D, Hyland W, Busse P, et al. Immediate reconstructionof the perineal wound with gracilis muscle flaps following ab-dominoperineal resection and intraoperative radiation therapyfor recurrent carcinoma of the rectum. Ann Surg Oncol 1999;6:33–37.

6. Khoo AKM, Skibber JM, Nabawi AS, et al. Indications forimmediate tissue transfer for soft tissue reconstruction in visceralpelvic surgery. Surgery 2001;130:463–469.

7. Buchel EW, Finical S, Johnson C. Pelvic reconstruction usingvertical rectus abdominis musculocutaneous flaps. Ann PlastSurg 2004;52:22–26.

8. Hilsabeck JR. The presacral space as a collector of fluid accumu-lations following rectal anastomosis: tolerance of rectal anasto-mosis to closed suction pelvic drainage. Dis Colon Rectum1982;25:680–684.

9. Miller LB, Steele G, Cady B, et al. Resection of tumors in irra-diated fields with subsequent immediate reconstruction. ArchSurg 1987;122:461–466.

0. Baird WL, Hester TR, Nahai F, et al. Management of perinealwounds following abdominoperineal resection with inferior glu-teal flaps. Arch Surg 1990;125:1486–1489.

1. Farid H, O’Connel TX. Methods to decrease the morbidity ofabdominoperineal resection. Am Surg 1995;61:1061.

2. Irvin TT, Goligher JC. A controlled clinical trial of three differ-ent methods of perineal wound management following excisionof the rectum. Br J Surg 1975;62:287–291.

3. Borel Rinkes IH, Wiggers T. Gracilis muscle flap in the treat-ment of persistent, infected pelvic necrosis. Eur J Surg 1999;

165:390–391.

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

703Vol. 206, No. 4, April 2008 Butler et al VRAM Flap Perineal Reconstruction

4. Kroll SS, Pollock R, Jessup JM, et al. Transpelvic rectus abdo-minis flap reconstruction of defects following abdominal-perineal resection. Am Surg 1989;55:632–637.

5. Radice E, Nelson S, Mercill R, et al. Primary myocutaneous flapclosure following resection of locally advanced pelvic malignan-cies. Br J Surg 1999;86:349–354.

6. Hay JM, Fingerhut A, Paquet JC, et al. Management of thepelvic space with or without omentoplasty after abdominoperi-neal resection for carcinoma of the rectum: a prospective multi-center study. The French Association for Surgical Research. EurJ Surg 1997;163:199–206.

7. Tei TM, Stolzenburg T, Buntzen S, et al. Use of transpelvicrectus abdominis musculocutaneous flap for anal cancer salvagesurgery. Br J Surg 2003;90:575–580.

8. Chessin DB, Hartley J, Cohen AM, et al. Rectus flap reconstruc-tion decreases perineal wound complications after pelvic chemo-radiation and surgery: a cohort study. Ann Surg Oncol 2005;12:104–110.

9. Ferenschild FT, Vermaas M, Hofer SO, et al. Salvage abdomi-noperineal resection and perineal wound healing in local recur-rent or persistent anal cancer. World J Surg 2005;29:1452–1457.

0. Vermaas M, Ferenschild TFJ, Hofer SOP, et al. Primary andsecondary reconstruction after surgery of the irradiated pelvisusing a gracilis muscle flap transposition. Eur J Surg Oncol2005;31:1000–1005.

1. Houvenaeghel G, Ghouti L, Moutardier V, et al. Rectus abdo-minis myocutaneous flap in radical oncopelvic surgery: a safe

and useful procedure. Eur J Surg Oncol 2005;31:1185–1190.

2. de Haas WG, Miller MJ, Temple WJ, et al. Perineal woundclosure with the rectus abdominis musculocutaneous flap aftertumor ablation. Ann Surg Oncol 1995;2:400–406.

3. Menon A, Clark MA, Shatari T, et al. Pedicled flaps in thetreatment of nonhealing perineal wounds. Colorectal Dis 2005;7:441–444.

4. Galandiuk S, Jorden J, Mahid S, et al. The use of tissue flaps asan adjunct to pelvic surgery. Am J Surg 2005;190:186–190.

5. Cordeiro PG, Pusic AL, Disaz JJ. A classification system andreconstructive algorithm for acquired vaginal defects. Plast Re-constr Surg 2002;110:1058–1065.

6. Pemberton JH. How to treat the persistent perineal sinus afterrectal excision. Colorectal Dis 2003;5:486–489.

7. Butler CE, Rodriguez-Bigas MA. Review, pelvic reconstructionafter abdominoperineal resection: is it worthwhile? Ann SurgOncol 2005;12:91–94.

8. Greene FL, Page DL, Fleming ID, et al. AJCC cancer stagingmanual. 6th ed. New York: Springer-Verlag; 2002:113�130.

9. Logan SE, Mathes SJ. The use of a rectus abdominis myocuta-neous flap to reconstruct a groin defect. Br J Plast Surg 1984;37:351–353.

0. Soper JT, Havrilesky LJ, Secord AA, et al. Rectus abdominismyocutaneous flaps for neovaginal reconstruction after radicalpelvic surgery. Int J Gynecol Cancer 2005;15:542–548.

1. Smith HO, Genesen MC, Runowicz CD, et al. The rectus ab-dominis myocutaneous flap: modifications, complications, andsexual function. Cancer 1998;83:510–520.

2. Jain AK, DeFranzo AJ, Marks MW, et al. Reconstruction of pelvicexenterative wounds with transpelvic rectus abdominis flaps: a case

series. Ann Plast Surg 1997;38:115–122; discussion 122�123.