Keywords
abdominal wall reconstruction - anterolateral thigh flap - fasciocutaneous flap -
free flap - innervated - innervation - neurotized - neurotization - plastic surgery
- systematic review
Reconstruction of large full-thickness abdominal wall defects provides unique challenges
to the plastic surgeon, especially when wide oncologic resection of soft tissue obviates
the use of locoregional options. Defect location at the epigastrium further limits
options. When abdominal wall defects are small to medium-sized, there are several
options available before free flap reconstruction is needed. Component separation
provides innervated muscle flaps for closing defects up to 20 cm (at the level of
the umbilicus).[1] Locoregional flaps are also available, including pedicled tensor fascia lata, anterolateral
thigh, rectus femoris, and latissimus dorsi flaps. Pedicled flaps are limited, as
in other regions of the body, by pedicle reach, the arc of rotation, size, unpredictable
nature of the distal end of the flap, and unpredictable function of the muscle once
transferred to the recipient site. It is accepted now that large, full-thickness abdominal
wall defects require free flap reconstruction when they are in the midline and not
amenable to component separation, or when the rectus abdominous and its fascia are
not available.
Koshima et al[2] described and Williams et al performed free tensor fascia lata flaps for abdominal
wall reconstruction to overcome the previously described limitations of pedicled flaps.[3] Since the 1980s, the tensor fascia lata has been the flap of choice for free flap
reconstruction of abdominal wall defects.[4]
[5] Tensor fascia lata and the iliotibial band may be harvested together to reconstitute
the abdominal fascia. The lateral circumflex femoral system, on which the tensor fascia
lata flap is based, was found to be unique in that it allows for composite flaps,
potential innervation, and enough soft tissue to reconstruction the entire abdominal
wall. The anterolateral thigh, anteromedial thigh, and tensor fascia lata can be harvested
together as one large flap. While we have identified excellent donor sites for free
flap abdominal wall reconstruction, challenges that remain include finding ideal recipient
vessels, and determining what adjunct techniques will maximize the functional outcome.
Significant advances have been made in the last 20 years with regard to the above
factors, especially with the advent of innervated, or neurotized, flaps. Below, the
authors review the literature on neurotized free flap abdominal wall reconstruction
in an effort to clarify the available techniques, determine functional outcomes, and
potentially to establish what the “gold standard” should be in free flap abdominal
wall reconstruction.
Methods
MEDLINE (PubMed), EMBASE, and the Cochrane Collaboration Library were thoroughly searched
by the authors from January 1975 through November 2016. Also, bibliographies of each
relevant citation were reviewed for additional sources. The following search terms
were used as both subjects and keywords: “abdominal wall reconstruction” AND (“neurotized”
OR “neurotization” OR “functional” OR “innervated” OR “innervation”).
Two independent reviewers evaluated the titles and abstracts of all studies without
language restrictions and subsequently chose studies based on the inclusion and exclusion
criteria. The authors included studies that were published in scientific journals
and involved patients who underwent a neurotized free flap for abdominal wall reconstruction.
The authors excluded studies that were focused on procedures unrelated to neurotized
free flaps for abdominal wall reconstruction, or review articles that only discussed
such reconstructions without reporting on any specific cases. Discrepancies between
the reviewers were discussed, and a third senior author (J.D.K.) decided as to whether
the study should be included or excluded. The references of each study were reviewed
for additional potential studies. The full text of studies that met criteria was reviewed
as a second stage, and additional exclusions were made. Techniques involving neurotized
free flap reconstruction of abdominal wall defects were evaluated and compared, as
were the motor, sensory, and functional outcomes related to flap transfer.
Results
The initial PubMed search yielded 264 studies. The Cochrane database search yielded
four studies. After reviewing the abstracts based on our criteria, 46 studies were
selected pertaining to abdominal wall reconstruction with free flaps. Additional review
of references of these articles yielded four more articles. After reviewing the full
text of studies and making final exclusions, the final pool of studies pertaining
to neurotized free flap reconstruction of the abdominal wall was comprised of nine
case series with a total of 16 patients ([Fig. 1]).
Fig. 1 Article search process and results totaling nine articles.
The mean age of patients in these series was 40.4 years old (range: 17–71). Eight
of the patients were males (50.0%), 6 were females (37.5%), and 2 were unknown (12.5%).
Locally aggressive soft tissue masses were the etiology for 37.5% of defects (n = 6), incisional hernia for 31.3% (n = 5), motor vehicle trauma for 18.8% (n = 3), and recurrent or metastatic oncologic disease for 12.5% (n = 2). The mean defect size requiring free flap abdominal wall reconstruction was
392.2 cm2 (range: 180–700) ([Table 1]).
Table 1
Preoperative data including study type, age, and etiology
Reference
|
Type of study
|
No. of patients
|
Average age (range) (y)
|
Etiology of defect
|
Ninković et al (1998)
|
Retrospective
|
4
|
23 (17–29)
|
MVA, dermatofibrosarcoma protuberans, sarcoma (2)
|
Sasaki et al (1998)[20]
|
Retrospective
|
2
|
47.5 (27–68)
|
Squamous cell carcinoma in fistula, ovarian cancer metastases to abdomen
|
Koshima et al (1999)[21]
|
Retrospective
|
1
|
71
|
Incisional hernia
|
Koshima et al (2003)
|
Retrospective
|
1
|
48
|
Sigmoid colon rupture
|
Malheiro et al (2007)[22]
|
Retrospective
|
1
|
38
|
Complicated perforated gastric ulcer
|
Wong et al (2009)
|
Retrospective
|
2
|
47.5 (41–54)
|
Incisional hernia status-postnephrectomy, blunt trauma
|
Chalfoun et al (2012)
|
Retrospective
|
2
|
29 (23–35)
|
Motorcycle crash (2)
|
Iida et al (2013)
|
Retrospective
|
2
|
44 (43–45)
|
Dermatofibrosarcoma protuberans (2)
|
Hahn et al (2016)
|
Retrospective
|
1
|
26
|
Recurrent desmoid tumor
|
Abbreviation: MVA, motor vehicle accident.
Average time to surgery after the initial indication was 24 months (range: 0–144).
The latissimus dorsi was used in 31.3% of reconstructions (n = 5), tensor fascia lata in 25% (n = 4), rectus femoris in 12.5% (n = 2), combined tensor fascia lata-anterolateral thigh in 12.5% (n = 2), combined vastus lateralis-tensor fascia lata-anterolateral thigh flaps in 12.5%
(n = 2), and vastus lateralis-anterolateral thigh in 6.3% (n = 1). The recipient vessel was the inferior epigastric in 50% of reconstructions
(n = 8), gastroepiploic in 12.5% (n = 2), femoral in 12.5% (n = 2), lateral femoral circumflex in 6.3% (n = 1), superficial epigastric in 6.3% (n = 1), and unknown in 12.5% (n = 2). Mesh was utilized in 25% of reconstructions (n = 4)—three synthetic and one biological mesh. Osseous fixation of the flap was performed
in 25% of reconstructions (n = 4) ([Table 2]).
Table 2
Intraoperative data including defect size, flap type, neurotization type, neurroraphy,
recipient vessels, surgical delay, mesh use, and osseous fixation
Reference
|
Defect (cm2)
|
Flap
|
Neurotization
|
Nerve recipient
|
Recipient vessel
|
Average surgical delay (range) (mo)
|
Mesh
|
Osseous fixation
|
Ninković et al (1998)
|
700, 364, unknown (2)
|
Latissimus dorsi (4)
|
Motor (4)
|
Intercostal (4)
|
Superior epigastric (1), inferior epigastric (1), unknown (2)
|
78 (12–144)
|
Synthetic (2), none (2)
|
None
|
Sasaki et al (1998)
|
396, 204
|
Tensor fascia lata (1), ALT (2)
|
Motor (2)
|
Intercostal (2)
|
Inferior epigastric (2)
|
6 (0–12)
|
None (2)
|
None
|
Koshima et al (1999)
|
180
|
Rectus femoris
|
Motor
|
T10 intercostal
|
Lateral femoral circumflex
|
24
|
None
|
Yes
|
Koshima et al (2003)
|
Unknown
|
Rectus femoris
|
Motor
|
Femoral nerve branch left intact
|
Gastroepiploic
|
36
|
None
|
Yes
|
Malheiro et al (2007)
|
500
|
Latissimus dorsi
|
Motor
|
Intercostal
|
Inferior epigastric
|
12
|
Synthetic
|
None
|
Wong et al (2009)
|
375, 525
|
Tensor fascia lata (2)
|
Motor (2)
|
T10 intercostal (1), unknown (1)
|
Femoral artery (2)
|
Unknown
|
None (2)
|
None (2)
|
Chalfoun et al (2012)
|
360, 700
|
Tensor fascia lata (2)
|
Motor (2)
|
Intercostal (2)
|
Inferior epigastric (2)
|
Unknown
|
None (2)
|
Yes (2)
|
Iida et al (2013)
|
224, 195
|
Vastus lateralis/tensor fascia lata/ALT (2)
|
Motor (1), motor and sensory (1)
|
T7 intercostal (1), T10 intercostal (1)
|
Inferior epigastric (2)
|
0
|
None (2)
|
None (2)
|
Hahn et al (2016)
|
375
|
Vastus lateralis/ALT
|
Sensory
|
T5 intercostal
|
Gastroepiploic
|
0
|
Biologic
|
None
|
Abbreviation: ALT, anterolateral thigh.
All but one reconstruction had motor neurotization performed (n = 15), while only 12.5% (n = 2) had sensory neurotization performed. One reconstruction had both motor and sensory
innervation. Neurroraphy was performed to an intercostal nerve in 87.5% of cases (n = 14), and unknown in 6.3% (n = 1). The femoral nerve was left intact in one case for motor innervation of a rectus
femoris flap. At least 66.6% of patients (n = 10) who had motor neurotization regained motor function as evidenced by voluntary
contraction, hip flexion, or the ability to sit up, while 93.3% (n = 14) had “satisfactory” motor function on author's subjective description (one patient
died from metastases before follow-up). Electromyography was used to confirm innervation
in 37.5% of patients (n = 6). Both flaps that had sensory innervation were successful with Semmes–Weinstein
testing of 3.61. Average patient follow-up was 25.4 months (range: 1.5–60) ([Tables 2] and [3]).
Table 3
Postoperative data including follow-up, functional recovery, sensation, complications,
reoperation, and flap survival
Reference
|
Average follow-up (range) (mo)
|
Clinical functional recovery
|
Sensation (Semmes–Weinstein test)
|
Major complications
|
Minor complications
|
Reoperation
|
Flap survival
|
Ninković et al (1998)
|
88 (2–37)
|
75% (Unknown (1))
|
N/A
|
50% (Flap venous thrombosis [1], died from metastases [1])
|
25% (seroma [1])
|
25% (saphenous vein graft to inferior epigastric [1])
|
100%
|
Sasaki et al (1998)
|
30.7 (1.5–60)
|
50% (Unknown [1])
|
N/A
|
0%
|
0%
|
0%
|
100%
|
Koshima et al (1999)
|
54
|
Unknown
|
N/A
|
0%
|
0%
|
0%
|
100%
|
Koshima et al (2003)
|
54
|
100%
|
N/A
|
0%
|
0%
|
0%
|
100%
|
Malheiro et al (2007)
|
24
|
Unknown
|
N/A
|
0%
|
0%
|
0%
|
100%
|
Wong et al (2009)
|
12, unknown (1)
|
50% (Unknown (1))
|
N/A
|
0%
|
0%
|
0%
|
100%
|
Chalfoun et al (2012)
|
10, unknown (1)
|
100%
|
N/A
|
50% (flap venous thrombosis [1])
|
0%
|
50% (lateral femoral circumflex vein graft [1])
|
100%
|
Iida et al (2013)
|
17 (10–24)
|
100%
|
3.61 (1), N/A (1)
|
0%
|
0%
|
0%
|
100%
|
Hahn et al (2016)
|
18
|
N/A
|
3.61
|
0%
|
100% (Seroma (1))
|
0%
|
100%
|
Major complications occurred in 18.8% of patients during the acute recovery period
(venous thrombosis of flap [n = 2], died from metastasis [n = 1]). Flap survival was 100% (n = 16), including two flaps requiring a return to the operating room for evacuation
of venous thrombosis and vein grafting. Minor complications occurred in 12.5% of patients
(seroma, n = 2). No patients had developed hernias during the follow-up period ([Table 3]).
Discussion
Despite the relative complexity of the cases in this review, abdominal wall reconstruction
with neurotized free flaps appears to be a safe procedure with a high rate of success.
There were no free flap losses in any of the case series, and the two cases requiring
a return to the operating room for venous thrombosis were salvaged with vein grafts.
The other associated complications were seromas that were managed successfully with
percutaneous drainage. With regard to functionality, motor innervation of free flaps
appears to be successful in over 93% of cases and sensory innervation in 100% of cases.
Although functionality was often measured subjectively, or on physical examination,
motor innervation was additionally confirmed with electromyogram in over one-third
of cases. Although longer term data are needed, none of these patients have developed
hernias during an average follow-up period of more than 2 years. Although this is
a small set of patients, it is clear that neurotized free flaps have great potential
to restore long-term function to the abdominal wall. However, there are many considerations
when designing free flaps for abdominal wall reconstruction for which we do not have
a “gold standard”: flap type, recipient vessel, mesh use, osseous fixation, and motor
and/or sensory innervation. Below the authors discuss some of these considerations
and lessons learned from this systematic review.
Recipient Vessels
Potential abdominal wall recipient vessels include the superior and inferior epigastric,
intercostal, superficial circumflex iliac, and gastroepiploic vessels. More distant
recipients who have been used in thoracic and abdominal wall reconstruction include
the thoracoacromial, thoracodorsal, and cervical vessels.[6]
When there is large tumor expiration of the abdominal wall, several of these options
may be precluded. Also, use of extraperitoneal vessels will restrict the type of flap
that can be used. Flaps with a short pedicle, or those with a pedicle located at the
center of the flap, will not be able to be used when the recipient's vessels sit at
the periphery of the abdominal wound. Use of peripheral recipient vessels may also
result in kinking of the pedicle and subsequent venous thrombosis, which occurred
in multiple cases.[5]
[6]
[7] None of the flaps in these series that were anastomosed to the gastroepiploic vessels
had complications. The gastroepiploic vessels are likely the most versatile, and often
overlooked, recipient for free flap abdominal wall reconstruction.[8] The vessels are simple to locate when the peritoneum is open—they can be found just
inferior to the pylorus, parallel to the greater curvature of the stomach. The gastroepiploic
vessels can be dissected out to a length of around 10 cm with a diameter in the range
of 2 to 3 mm.[8]
[9] Also, if the pedicle is at the deep surface of the flap, an uninterrupted fascial
closure can be performed without having to leave an open passage for the pedicle,
as would be necessary when using extraperitoneal vessels.[6] When a mesh underlay is used to reconstitute the fascial layer, a “window” at an
intercostal space can be made for the pedicle to run through.[10] This allows for uninterrupted mesh closure of the fascial layer, as well as prevention
of pedicle kinking if the pedicle had been run around the edge of the mesh ([Fig. 2]).
Fig. 2 The gastroepiploic vessels can be tunneled through a window in an intercostal space
when utilizing an underlay mesh closure.
There are situations in which all abdominal vessels have been damaged and thus are
not available (i.e., scarring and fibrosis from prior surgeries). One possibility
is to create an arteriovenous shunt, or Corlett's loop, between the femoral artery
and the long saphenous vein. The arteriovenous loop can then be rotated superiorly
toward the abdominal defect and divided to provide arterial inflow and venous outflow.[11]
Neurotized Flaps
Reconstruction of full-thickness defects of the abdominal wall is unique in that while
the integrity of musculofascial layer may be restored, dynamic stability is not always
returned. The importance of the stability provided by the rectus abdominous muscles
is demonstrated by the reduced abdominal flexion and rotational strength, pain, hernias,
and bulges that result after transverse rectus abdominis flap transfer and the secondary
displacement of the oblique muscles.[12] Innervated, or neurotized, flaps allow for the potential passive tone of the abdominal
wall, prevention of flap atrophy, and even active muscle contraction. If flap atrophy
occurs, as with a noninnervated flap, abdominal wall laxity may result.[13] Therefore, a majority of large abdominal wall reconstruction being performed, either
with acellular dermal matrix and fasciocutaneous free flaps or with noninnervated
musculocutaneous flaps, have the same potential consequences—an abdominal wall that
lacks dynamic strength and may result in chronic pain and weakness for the patient.
Neurotized tensor fascia lata, rectus femoris, and latissimus dorsi flaps have been
used to provide dynamic stability of the lower abdominal wall successfully.[5]
[13] Neurotized composite flaps of the anterolateral thigh have also been described as
having success using the vastus lateralis as the muscle component with its femoral
nerve branch for neurroraphy.[14] There is still, however, the debate as to the optimal method in performing neurotized
free flap abdominal wall reconstruction. The results with neurotized, or reinnervated,
free muscle flaps for abdominal wall reconstruction have been questioned by some who
feel that the best possible outcome occurs by performing a free flap, but leaving
the motor nerve intact.[13] This would obviously require a very tedious dissection. The technical details of
flap inset and fixation are also critical. Some surgeons argue that innervated flaps
for abdominal wall reconstruction are not useful when sutured to the fascial edges
of the defect alone, as this is not a stable point of fixation. Bony stabilization
of one end of the flap to the pelvic ring has been performed and may be helpful in
maintaining dynamic stability. Cases using osseous fixation are lacking, but have
proven successful in one small series of patients.[5]
When considering the optimal abdominal wall reconstruction, skin, and fascial continuity
would be restored, the musculofascial component would remain innervated with the ability
for contraction, and the skin component would be sensate. While this is the theoretical
ideal, we are rarely able to achieve this level of reconstruction. While nine case
series exist describing free flaps for innervated abdominal wall reconstruction, it
can be seen from the above systematic review that few describe successful sensory
reinnervation.[10]
[14] Iida et al described using a free combined vastus lateralis and anterolateral thigh
flap for abdominal wall reconstruction in which the both the sensory and motor components
were neurotized. The femoral nerve to the vastus lateralis and the lateral femoral
cutaneous nerve were anastomosed to the intercostal nerve. Electromyography confirmed
contraction of the vastus lateralis component while Semmes–Weinstein testing confirmed
sensation.[14] Hahn et al reported a second case of successful sensory innervation of a free combined
vastus lateralis and anterolateral thigh flap for abdominal wall reconstruction.[10]
Applications to Composite Tissue Allotransplantation of the Abdominal Wall
The concepts critical to functional free flap abdominal wall reconstruction are also
applicable to composite tissue allotransplantation of the abdominal wall. While abdominal
wall transplantation has been performed successfully in patients undergoing concurrent
solid organ transplantation, it has come across several challenges including poor
functional outcomes and controversy with regard to the optimal anatomic configuration
of the allograft.[15] These problems have limited the applicability of elective abdominal wall transplants
in patients not requiring solid organ transplantation. The lack of innervation and
lack of bony stability of these allografts have been flagged as potential causes for
poor functionality with a resultant hernia and bulge.[16]
[17]
From the results of the above review, it can be seen that neurotization of free flaps
appears to result in excellent motor and sensory outcomes for abdominal wall reconstruction,
and thus suggests that abdominal wall allografts for “end-stage hernias” may be successful.
The results also beg the question as to whether osseous fixation of abdominal wall
free flaps and allografts should be performed. In cadaveric studies performed by Singh
et al the abdominal wall allograft was designed as an osteomyocutaneous graft with
ribs harvested contiguously.[16] Chalfoun et al reported osseous fixation of their free flaps for abdominal wall
reconstruction to the pubis, which is another possible addition to allograft design.[5] Minimizing ischemic time of the abdominal wall allografts is important to the feasibility
of the operation.[18] Recipient vessels in abdominal wall transplantation have almost exclusively been
the inferior epigastric, common iliac, or circumflex iliac vessels.[19] It can be seen from the above review that the gastroepiploic vessels may be a convenient
and expedient alternative that has been used in free flap abdominal wall reconstruction
successfully.
Conclusion
While innervated free flaps have been used in facial reanimation, phalloplasty, extremity
reconstruction, and decubitus ulcer reconstruction, only nine published case series
exist describing innervated free flap abdominal wall reconstruction. A majority of
neurotized free flap reconstructions for abdominal wall defects have been performed
for motor innervation, which is almost invariably successful. The addition of sensory
innervation to free flap reconstruction of the abdominal wall would more completely
satisfy the “like for like” principle of reconstructive surgery. The methods utilized
in neurotized abdominal wall reconstruction should be applied to composite tissue
allotransplantation of the abdominal wall to improve functionality, outcomes, and
applicability.