Journal of Plastic, Reconstructive & Aesthetic Surgery (2014) 67, 634e639

Superior epigastric artery perforator flap forsternal osteomyelitis defect reconstruction*

R. Wettstein a,*, M. Weisser b, D.J. Schaefer a,D.F. Kalbermatten a

a Department of Plastic, Reconstructive, Aesthetic and Hand Surgery, University Hospital of Basel,Basel, Switzerlandb Department of Infectious Diseases, University Hospital of Basel, Basel, Switzerland

Received 16 December 2013; accepted 26 January 2014

KEYWORDSPropeller flap;Infection control;Thoracic wallreconstruction;Sternum osteitis

*This work has been presented at theand at the World Society for Reconstr* Corresponding author. Department

CH-4031 Basel, Switzerland. Tel.: þ41E-mail address: drwette@yahoo.co

1748-6815/$-seefrontmatterª2014Brihttp://dx.doi.org/10.1016/j.bjps.2014.0

Summary Sternal osteomyelitis after median sternotomy is associated with considerablemorbidity and mortality. Combined with radical debridement, muscle and less frequentlyomentum flaps are used to reconstruct the resulting defects. In this study, we present ourexperience with the fasciocutaneous superior epigastric artery perforator (SEAP) flap fordefect closure. After resection of the entire sternum, including the costochondral archesand the sternoclavicular joints, the repair of the defect was performed with the perforator flapwithout any re-stabilisation of the thoracic wall. A consecutive series of nine patients with amean age of 69 � 6 years were reconstructed with the SEAP flap. The mortality rate was zero.One patient developed a mediastinal haematoma and required five re-interventions by thecardiothoracic surgeons and thereafter a revision to close a small-wound dehiscence at thetip of the flap. Another two patients developed partial necrosis of the flap that could bemanaged conservatively. One patient had a revision for a seroma on the donor site, resultingin a 100% closure rate of the defect; there were revisions in two out of nine patients. The un-derlying infection was controlled by debridement, antibiotic therapy and flap closure in allcases. The overall success of the procedure was satisfactory; however, the local complicationrate was relatively high with three out of nine patients on the flap side and one of nine on thedonor site. Major advantages of the perforator flap in this highly morbid patient cohort are thatthe operation is relatively quick, muscle tissue is spared and re-education facilitated.ª 2014 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published byElsevier Ltd. All rights reserved.

24th annual meeting of the European Association of Plastic Surgeons in Antalya, 23rde25th May 2013,uctive Microsurgery (WSRM) 2013 Congress in Chicago, IL, USA, 11the14th July 2013.of Plastic, Reconstructive, Aesthetic and Hand Surgery, University Hospital Basel, Spitalstrasse 21,61 556 5482; fax: þ41 61 265 7301.m (R. Wettstein).

tishAssociationofPlastic,ReconstructiveandAestheticSurgeons.PublishedbyElsevierLtd.All rightsreserved.1.037

Sternal osteomyelitis defect reconstruction 635

Ideally, defect reconstruction after thorough debridementof sternal osteomyelitis is a safe procedure with minimaldonor-site morbidity permitting rapid recovery in this pa-tient group presenting frequently in a poor general condi-tion. Traditionally, muscle flaps, mainly pectoralis majorturnover or advancement flaps,1 which may be sufficient incases of sternal osteotomy wound dehiscence, have beenused. In cases of recalcitrant osteomyelitis necessitatingradical debridement that creates large defects afterdebridement, latissimus dorsi and vertical rectus abdominismuscle flaps or omental flaps have been used.2 Associatedweakening of the abdominal wall by muscle harvest oropening of the abdominal cavity for harvest of the omen-tum or relatively long operation time with free flaps may bepotential drawbacks of these procedures. Perforator flapsrepresent the most advanced method of tissue transfer interms of applied anatomic and physiological principles.Basically, the use of a perforator flap maximally decreasesdonor-site morbidity and spares muscle tissue that iscurrently probably most frequently used for sternal defectclosure.

Theoretically, perforators such as superior epigastricartery perforator (SEAP) present an ideal source for fas-ciocutaneous flaps for sternal reconstruction. Previously,the use of this flap has been described for autologous breastvolume augmentation after massive weight loss,3 tumourresection and radionecrosis,4e8 or after trauma.9 Of thetotal of 19 flaps described in the literature, only three wereused for sternal osteomyelitis.10 The aim of the presentstudy is to present the results obtained with a consecutiveseries of SEAP flaps for defect closure after radical ster-nectomy for recalcitrant osteomyelitis with an emphasis oninfection control, local complications of the flap and thedonor site and duration of post-flap hospital stay.

Patients and methods

General patient data

Between May 2010 and July 2012, nine patients presentingwith osteomyelitis after sternotomy were operated on.Informed consent was taken from all patients includingapproval for publication of fotodocumentation. All patientsunderwent at least one debridement (range 1e4) andwound conditioning before defect closure, but did not un-dergo any flap procedure to close the defect before theSEAP flap procedure. American Society of Anaesthesiolo-gists (ASA) score, presence of diabetes mellitus, arterialhypertension and active smoking were assessed.

Surgical technique

All procedures were performed under general anaesthesia.After debridement, multiple biopsies were harvested formicrobiological and histological workup. Perforator vesselswere identified with a handheld Doppler probe with thepossibility of previous internal mammary artery harvest inmind. No other imaging techniques were used for operativeplanning. After perforator mapping, the skin paddle wasdesigned with the most suitable perforator serving as thepivot point. None of the flaps crossed the midline or, in

other words, the perforators were on the ipsilateral side inrelation to the flap. The inframammary fold, that is, thelower border of the pectoralis muscle, served as the upperborder of the flaps that were horizontally orientated in allcases resulting in a final flap rotation of 90�.

Flap dissection started in the distal, lateral part of theflap in the plane above the muscle fascia, which was leftintact, towards the previously identified pedicle. If therewas more than one potential pedicle identified with theDoppler probe, flap preparation continued until the first(pivot point) perforator was identified. Fine perforatorpreparation was limited to the amount necessary to trans-pose the flap into the defect (Figure 1). The flaps could beeasily rotated 90� into the defect and all donor sites couldbe primarily closed by a reverse abdominoplasty-type pro-cedure. All patients received antibiotic treatment accord-ing to previous biopsy results or on an empirical basis withlater adaptation as recommended by the infectious diseaseteam. Drains were removed if output was <20 ml/24 h. Thetypical presentation after 1-year follow-up is presented inFigure 2.

Post-operative evaluation

All patients were prospectively monitored for post-operative complications, especially for signs of infection,wound dehiscence and flap necrosis-related and donor-site-related complications. Complications were classified asmajor or minor depending on whether a reoperation wasnecessary or not. The duration of hospital stay and the 30-day mortality were also obtained.

Results

General patient data

Mean age was 69 � 7 years; there were five male and fourfemale patients; ASA score was for all patients III or IV.Body mass index (BMI) ranged between 30 and 39 kg m�2

(mean 34 � 3 kg m�2) (Table 1). Four of the patients hadeither three or four coronary artery bypasses and one pa-tient had five bypasses as the initial intervention.

Surgical data and complications

After debridement, the defect size in surface ranged be-tween 120 and 280 cm2 (mean 190 � 50 cm2). The meanoperation time was 128 � 25 min including thedebridement.

There was no recurrent or persistent infection after theflap procedures combined with antibiotic therapy. Onepatient developed a mediastinal haematoma not related tothe flap procedure and required five re-interventions by thecardiothoracic surgeons and a last revision for a tip necrosisof the flap. Two more partial flap necroses, that is, wounddehiscences, could be managed conservatively, resulting inan overall ischaemic complication rate of 3/9 (Table 2). Atthe donor site, one patient was taken back to the operationroom for a seroma, whereas haematoma and seroma in onepatient were managed conservatively. One patient

Figure 1 Presentation of a typical wound before radical debridement in a female (a, b and c) and male (d, e and f) patient. Thesituation after radical debridement and flap preparation (a, d). Directly after flap inset, there is no tenting in spite of the fact thatthe breasts are pulling laterally in the female patient (c). The donor site has been closed by advancement of the abdominal skin,which can lead to a distortion of the nipple areola complex position (f).

636 R. Wettstein et al.

underwent scar revision and flap thinning, 6 months afterthe flap procedure. One patient developed a sacral pres-sure sore during the hospitalisation and underwent adebridement and flap closure.

Hospitalisation and 30-day mortality

The mean overall hospitalisation time was 98 � 38 days, ofwhich 42 � 25 days were after the flap procedure, that is,most patients were discharged within 40 days after the flapprocedure; one patient with a severe, not flap-related

complication remained hospitalised for 98 days after theoperation. On the other hand, one patient could be dis-charged to rehabilitation eight days after the flap proce-dure. None of the patients died during the hospitalisation orwithin 30 days.

Discussion

The SEAP flap was successfully used in nine cases of sternalosteomyelitis. All patients recovered from surgery and the30-day mortality rate was zero. However, two flap-related

Figure 2 Preoperative view and result after 1-year follow-up in a male (a, b) and female (c, d) patient. There is a slight bulging ofthe flap and the inframammary fold is displaced downward in the donor site (b). The final result in an overweight female patientwho refused revision of a recurrently ulcerating lesion on the medial edge of the left inframammary fold that was probably relatedto the pulling of the heavy breast and limited local care and hygiene (d).

Sternal osteomyelitis defect reconstruction 637

reoperations were necessary, one for partial flap necrosisand one for a seroma at the donor site. The secondarycontour correction 6 months after uneventful initial post-operative course was not considered a complication.Debridement, flap reconstruction and intravenous antibi-otics cured all infections.

In this consecutive series of nine patients, no intra-operative change of strategy from a perforator flap to amuscle flap was necessary, as perfusion was intra-operatively deemed adequate. Only perforators on the sideof an intact internal mammary artery were used. However,

Table 1 General patient data.

Patient Age Gender BMI DM AHT Smoking

1 65 M 31 1 1 12 73 F 33 1 1 03 70 M 33 0 1 14 64 F 35 1 0 05 84 F 30 1 1 06 66 M 36 1 1 17 61 M 39 0 1 08 68 F 34 1 1 09 69 M 34 0 1 1

69 � 7 5/4 34 � 3 6/9 8/9 4/9

The age and body mass index (BMI) are indicated in years,respectively (kg m�2). Mean and standard deviation are calcu-lated for age and BMI. DM: diabetes mellitus. AHT: arterialhypertension.

adequate perfusion of SEAP flap after internal mammaryartery harvest on the same side has been described.10 Noneof the flaps used extended over the posterior axillary line,and all were horizontally oriented, resulting in a homoge-neous flap series.

In all our cases, the flap was horizontally oriented andnot vertically. Early dye injection studies showed that atransverse band just below the xiphisternum over the ribsto the axillary midline was consistently perfused by the

Table 2 Flap-related complications at donor and recip-ient sites.

Patient Wound dehiscence/necrosis

Seroma/haematoma

Reoperation

1 0 0 02 0 1 03 0 0 04 1 0 15 0 0 06 1 1 07 0 0 08 0 1 19 1 0 0Total 3/9 3/9 2/9

Note that one of the reoperations was secondary to a seroma atthe donor site of the flap.

638 R. Wettstein et al.

SEAPs, whereas in the vertical direction, the umbilicus wasfrequently the lower limit of staining.11 The watershed,that is, choke vessels, is typically located at a level corre-sponding to the midpoint between the xiphisternum andthe umbilicus.12 In that study, extensive anastomoses ealso across the midline e were found between perforatorsof the SEAP system in the subcutaneous plane. Somewhatcontradictory to the findings of a watershed in the epigas-tric area was the presence of a vertically oriented, largeperforator with an extensive anatomical territory in eightof the ten cases,12 which would speak in favour of a verticalskin paddle and corroborates the clinical experience byHamdi et al., who used vertically oriented flaps for chestwall reconstruction.4

The reason for the observed tip necroses is probablyrelated to the difficulty in defining the territory of theperforator used. It has been found that no regularity andconsequently predictability were found in the angle orbranching pattern of abdominal perforators.12 Potential ad-vantages of the horizontal, transverse orientation of the flapare that rotation is limited to 90� leading to less torsionon thepedicle; consequently, less perforator dissection is needed.In general, the safe territory of a perforator flap depends onseveral anatomic features, such as perforator diameter,location and orientation within the flap tissue, and on thephysiologic capacity to capture adjacent territories con-nected via choke vessels. However, the amount of neigh-bouring territories connected via choke vessels that cansafely be included is not known. This is in contrast to muscleand fasciocutaneous flaps that present relatively constant,named pedicle vessels and well-defined territories. Thisdiscrepancy of the angiosome concept of a named sourcevessel to the concept of the perforator angiosome that showsmarked differences between perforators of a single-sourcevessel and between individuals has previously beenemphasised.13,14 This can lead to the typical ischaemic flapcomplications ranging from skin slough, delayed woundhealing to distal flap necrosis, which increase the risk of in-fectious complications, prolong the hospital stay andmay befatal in the case of sternal osteomyelitis.

Previous series described partial SEAP flap necrosis in twoof seven cases4 and two of three cases.10 Mah et al. recom-mended to keep the design of the flap medial to the lateralaxillary line, at bestmedial to the linea semilunaris in order toprevent tip necrosis.10 In the present series, there was onemajor tip necrosis requiring reoperation and two cases ofdelayedwoundhealing that could bemanaged conservatively.The perforator-plus concept postulates that perfusion is bet-ter; theprocedure is safer if theflap isnot completely islandedand a bridge of subcutaneous tissue and/or skin is main-tained.15e17 This view has recently been challenged by dy-namic laser-induced fluorescence videoangiography ofabdominal fasciocutaneous flaps that have shown betterperfusion if the skin bridgewas divided, that is,when the flapswere solely perfused and drained by the perforator vessels.18

Further clinical experience is probably needed to prove thatan island is safer than a peninsula. In our cases, the skin pad-dles of the flaps were completely divided, but perforatordissection was limited to the amount necessary to rotate theflap into the defect, and intramuscular preparation was notnecessary in any flap of this series, potentially resulting inwhat has been termed a ‘pedicle cluster perforator flap’.8

Overall, the complication rate was comparable to otherflap procedures in this relatively morbid patient group. Twoof nine patients were reoperated for a flap-relatedcomplication, which compares favourably to previouslypresented results with musculocutaneous flaps with areoperation rate of >40%.2 However, the infectious prob-lem may be solved with any flap tissue composition andseems to depend more on the wound preparation.2 Anobvious advantage of this perforator flap is the decrease indonor-site morbidity, as there is no weakening of the pec-toral and abdominal muscles and therefore in respiratoryactivity. The subcutaneous undermining in a reversedabdominoplasty fashion, however, can result in the typicalcomplications, such as haematoma and seroma, which ledto a reoperation in one case. The aesthetic outcome isusually not a major issue in this patient group; however,one patient required secondary flap contour correction.

Ideally, the process of healing and recovery after a flapprocedure is quick, and patients can be discharged early. Inthe present series, the mean duration of hospitalisationafter the flap procedure was 42 days. This is significantlylonger than previously reported and probably depends onmultiple factors, such as general condition of the patients,the multiple reoperations in one patient, the duration ofintravenous antibiotics recommended by the infectiousdisease team, as well as a delay in timely planning andlimited availability of rehabilitation programmes thatfollow for most of these patients.

One of the arguments against the use of fasciocutaneousflaps in the situation of sternal osteomyelitis is that thedefect is relatively deep and that bulk is needed. In thepresent series, no tenting of the flap over a mediastinalcavity was observed and all flaps seemed to contour thedefect nicely. Because all patients presented with a BMI>30 kg m�2, we cannot comment whether this was the casein patients with a thinner subcutaneous tissue layer. Theamount of soft-tissue laxity present and the defect size areprobably more vital than the thickness of the subcutaneouslayer. Infection is controlled by debridement, wound con-ditioning and antibiotic therapy for which the flap can beregarded as a transport vehicle. In purulent situations,neither a muscle nor a fasciocutaneous flap is indicated,but debridement and wound conditioning are indicated.Little is known about the importance of lymphatic drainageof flaps. Obviously, in completely mobilised perforatorflaps, there is a paucity of lymphatic drainage that may beproblematic.

In conclusion, local fasciocutaneous flaps for recon-struction of sternal osteomyelitis defects present an idealoption with complete muscle preservation and conse-quently no weakening of the respiratory forces. Eventhough the precise territory that can safely be included inthe SEAP flap is currently not known, the limited number ofcases and anatomic studies published in the literature andintraoperative estimation permit to obtain a relatively saferesult with an acceptable rate of complications and com-plete control of infection.

Conflict of interest/funding

None of the authors has any conflict of interest or funding.

Sternal osteomyelitis defect reconstruction 639

References

1. Schmid DW, Orasch-Jorg C, Wettstein R, Kalbermatten DF,Todorov A, Pierer G. Successful therapy for a patient with aninfected ascending aortic graft and sternal osteomyelitiswithout graft removal. Eplasty 2008;8:43e.

2. Wettstein R, Erni D, Berdat P, Rothenfluh D, Banic A. Radicalsternectomy and primary musculocutaneous flap reconstruc-tion to control sternal osteitis. J Thorac Cardiovasc Surg 2002;123:1185e90.

3. Szychta P, Anderson WD. Islanded pedicled superior epigastricartery perforator flaps for bilateral breast augmentation withmastopexy after massive weight loss. J Plast Reconstr AesthetSurg 2011;64:1677e81.

4. Hamdi M, Van Landuyt K, Ulens S, Van Hedent E, Roche N,Monstrey S. Clinical applications of the superior epigastric ar-tery perforator (SEAP) flap: anatomical studies and preopera-tive perforator mapping with multidetector CT. J PlastReconstr Aesthet Surg 2009;62:1127e34.

5. Uemura T. Superior epigastric artery perforator flap: pre-liminary report. Plast Reconstr Surg 2007;120:1ee5e.

6. Woo KJ, Pyon JK, Lim SY, Mun G-H, Bang SI, Oh KS. Deep su-perior epigastric artery perforator “propeller” flap forabdominal wall reconstruction: a case report. J Plast ReconstrAesthet Surg 2010;63:1223e6.

7. Ziegler K, Schmidt M, Huemer GM. A pedicled superior epigastricartery perforator (SEAP-) propeller flap for 2-cavity reconstruc-tion after oncologic rib resection.Microsurgery 2011;31:335e6.

8. Oni G, Sharma R, Rao R, Unger J, Saint-Cyr M. Bilateral superiorepigastric pedicle perforator flaps for total chest wallcoverage. J Plast Reconstr Aesthet Surg 2011;64:1104e7.

9. Hallock GG. The superior epigastric rectus abdominis muscleperforator flap. Ann Plast Surg 2005;55:430e2.

10. Mah E, Rozen WM, Ashton MW, Flood S. Deep superior epigas-tric artery perforators: anatomical study and clinical applica-tion in sternal reconstruction. Plast Reconstr Surg 2009;123:1719e23.

11. Boyd JA, Taylor GI, Corlett R. The vascular territories of thesuperior epigastric and the deep inferior epigastric systems.Plast Reconstr Surg 1984;73:1e16.

12. Tregaskiss AP, Goodwin AN, Acland RD. The cutaneous arteriesof the anterior abdominal wall: a three-dimensional study.Plast Reconstr Surg 2007;120:442e50.

13. Rozen WM, Grinsell D, Koshima I, Ashton MW. Dominance be-tween angiosome and perforator territories: a new anatomicalmodel for the design of perforator flaps. Microsurgery 2010;26:539e45.

14. Rozen WM, Ashton MW, Le Roux CM, Pan W-R, Corlett RJ. Theperforator angiosome: a new concept in the design of deepinferior epigastric artery perforator flaps for breast recon-struction. Microsurgery 2010;30:1e7.

15. Sharma RK, Mehrotra S, Nanda V. The perforator “plus” flap: asimple nomenclature for locoregional perforator-based flaps.Plast Reconstr Surg 2005;116:1838e9.

16. Mehrotra S. Perforator-plus concept: adding versatility totraditional flaps. Plast Reconst Surg 2009;123:219ee20e.

17. Mehrotra S. Perforator-plus flaps: a new concept in traditionalflap design. Plast Reconstr Surg 2007;119:590e8.

18. Me�si�c H, Kirkebuen KA, Bains R. The importance of a skinbridge in peripheral tissue perfusion in perforator flaps. PlastReconstr Surg 2012;129:428ee34e.