Microvascular Free Flap Failure Caused by
Steven P. Davison, D.D.S.,
Craig M. Kessler, M.D.
Ali Al-Attar, M.D., Ph.D.
Background: Microvascular free flap techniques have improved over the past
large series of high-volume centers. However, despite technical and patient-
selection advances, some components of free flap failure might be attributable
to undiagnosed patient factors.
Methods: In this clinicopathologic conference article, the authors present four
patients with major free flap microvascular complications who were later diag-
nosed with multiple risk factors for hypercoagulability and biochemical abnor-
malities, and present the characteristic associated findings.
laboratory investigation revealed the biochemical abnormalities.
Conclusions: In certain patients undergoing microvascular free flap recon-
struction, undiagnosed hypercoagulability risk factors can be a source of free
flap failure that is independent of patient selection or technical factors. Diag-
nosis, management, patient counseling, and initiation of deep venous throm-
boembolism prophylaxis measures are critical clinical components of care of
these patients.(Plast. Reconstr. Surg. 124: 490, 2009.)
rator flaps, the success rate has increased from 63
percent to 98.8 percent at centers of excellence.1–5
Despite greater experience, improved technology,
and development of anastomotic devices, the aver-
age failure rate has not dropped below 1 to 3 per-
one flap (deep inferior epigastric perforator flap),
improved patient selection, and consistency of sur-
geon, yet it is not zero.
Although the vast majority of studies of free
flap failure allude to technical causes of throm-
bosis, there are some cases in which the patient is
the contributing factor to flap failure.3,6–9One of
the hidden contributing factors is undiagnosed
coagulopathy.9,10We present four cases in which
multiple free flap failures were a direct result of
underlying hidden patient predisposition to de-
veloping hypercoagulability complications. We
icrosurgical reconstruction has become
progressively more successful. From the
early microvascular flaps to current perfo-
illustrate this process. The most problematic as-
pect is that the first presentation of coagulopathy
might be microvascular complications of the free
flap. We also discuss laboratory evaluation of the
patient and long-term implications for deep ve-
nous thrombosis prophylaxis.
The coagulopathies that are represented in
our case series range from relatively common to
rare. Many of these coagulopathies are described
gous versus heterozygous). Of critical importance is
is often multifactorial such that underlying coagu-
lopathy might serve to either cause or predispose to
microvascular anastomotic thrombosis.
A 43-year-old white woman with a previous wide resection of
the cheek and orbital exenteration for lacrimal gland tumor
presented with an ethmoid sinus fistula. Sinusectomy and wider
excision of the orbit required soft-tissue filling. A superficial
inferior epigastric artery free flap to the superficial temporal
artery and vein anastomoses was attempted, which resulted in
From the Department of Plastic Surgery and the Division of
Coagulation, Georgetown University Hospital.
Received for publication October 8, 2008; accepted February
Disclosure: None of the authors has any financial
conflicts of interest.
immediate intraoperative difficulties with an arterial thrombus.
This failure was attributed initially to recipient vessel spasm and
donor vessel diameter. At the same operation, a salvage deep
inferior epigastric artery perforator flap with pedicle length
adequate to reach the retromandibular external carotid artery
was difficult to infuse, requiring multiple arterial anastomoses
and tissue plasminogen activator flush and heparin drip. The
arterial anastomosis failed 12 hours later. Delayed reconstruc-
tion was performed using free tissue grafts and artificial dermal
substitute. A postoperative hematologic workup revealed ele-
vated plasminogen activator inhibitor-1 activity (79 percent;
normal, 4 to 43 percent), protein S activity deficiency (57 per-
cent; normal, 65 to 140 percent), elevated factor VIII activity
(175 percent; normal, 50 to 150 percent), and elevated an-
tiphosphatidyl immunoglobulin G and immunoglobulin M an-
tibodies (G-phospholipid, 31 U/ml; M-phospholipid 26 U/ml;
normal, ?11 and ?25 U/ml, respectively). Furthermore, gene
testing by means of polymerase chain reaction techniques re-
vealed that she was compound-heterozygous for the C677T and
A1298C polymorphisms for the MTHFR gene mutation and was
heterozygous for the R506Q polymorphism for the factor V
A 38-year-old white woman with right breast disseminated
carcinoma in situ underwent bilateral mastectomy with imme-
diate reconstruction by means of bilateral free transverse rectus
abdominis musculocutaneus flap surgery. The patient had a
17-pack-year smoking history and a BRCA1 mutation. At-
tempted bilateral muscle-sparing transverse rectus abdominis
musculocutaneus flaps to the thoracodorsal arteries were at-
tempted. Each flap anastomosis remained patent for approxi-
mately 30 minutes and then clotted. Attempted revision to the
internal mammary artery vessels with papaverine administra-
tion and adventiectomy were attempted, but the flap failed
despite these efforts. A total of six anastomoses were attempted.
Throughout the case, the anastomosed vessels appeared cloudy
and flaccid. The flaps were discarded and the breast flaps were
A hematologic workup revealed that the patient carried a
PAI-1 gene mutation positive for one copy of the 4G allele, had
protein S activity deficiency (54 percent; normal, 65 to 140
A 29-year-old white woman with left buccal sulcus carcinoma
underwent wide local excision with radial forearm free flap
reconstruction. Surgery was complicated by a return to the
operating room for intermittent flow interruption and release
of the tunnel for the vessels. The patient also developed a
subsequent deep venous thrombosis from a peripherally in-
serted central catheter line.
on the cheek was reconstructed with a second radial forearm
free flap to the superficial temporal vessels. Multiple arterial
thromboses were encountered and the flap was finally well-
of Neo-Synephrine (Bayer HealthCare, Leverkusen, Germany).
Attempted salvage with a deep inferior epigastric artery perfo-
on the cheek, as inset occluded flow (Fig. 1).
A postoperative hematologic workup revealed that the pa-
tient was homozygous positive for the C677T polymorphism for
the MTHFR gene and had elevated levels of factor VIII activity
(214 percent; normal, 50 to 150 percent). She was also found
to be hypothyroid, with a thyroid-stimulating factor level of 23
mIU/liter (normal, 0.23 to 5.6 mIU/liter) and a free thyroxine
level of 0.51 ng/dl (normal, 0.58 to 1.64 ng/dl).
A 42-year-old white woman with recurrent squamous cell
cancer of the anterior floor of the mouth underwent composite
resection including partial mandibulectomy. The patient had
previously received irradiation to the site and had an extensive
history of alcohol and tobacco use.
The oral defect was reconstructed using an immediate radial
forearm free flap that was complicated by arterial clot that was
cleared by thrombectomy in the same operative setting. Of
note, the patient had had an arterial line placed previously in
the donor extremity, and despite a negative Allen test preop-
eratively, the extremity donor site became ischemic and re-
quired a radial artery reconstruction using a vein graft in the
same surgical setting.
The postoperative course was complicated by recurrent
thrombosis of the free flap artery at the site of the previous
(Fig. 2). The defect was then reconstructed in delayed fashion
using an anterolateral thigh free flap, which developed post-
operative necrosis of the distal half, although sufficient tissue
the patient developed gangrene of one toe during her hospital
stay, distal to the dorsalis pedis artery, which had been cannu-
lated with an arterial catheter (Fig. 3).
Consultation with the hematology service and a hypercoag-
ulable workup revealed that the patient was compound het-
erozygous for the C677T/A1298C polymorphisms for the
MTHFR gene mutation. After the first failed free flap, the pa-
to 80 seconds) until final flap survival was ascertained.
The intraoperative experience with these cases was similar.
The flaps initially flashed up on the first anastomosis after
rent left cheek carcinoma was left with a through-and-through
also experienced significant inflow problems and was left lying
on the cheek as inset occluded flow.
Volume 124, Number 2 • Microvascular Free Flap Failure
manipulation, with excellent hyperemic flow to the tissue.
Within 5 to 20 minutes of flap perfusion, arterial flow ceased,
with thrombus visualized at the anastomotic suture line or site
of injury. Change in recipient vessel, donor flap, or surgical
device to maintain patency by eliminating suture damage at the
artery resulted in eventual thrombosis after 2 to 12 hours. The
nature of the artery during surgical manipulation was flaccidity
despite minimal trauma or cutting back the vessel to fresh
anastomoses were as follows:
1. The arterial anastomosis does not flash when the clamp is
removed; it requires manipulation.
2. The arterial anastomosis goes down quickly (within 1 hour),
before the rest of the procedure is completed.
3. The vessels, particularly the recipient flap vessels, are flaccid.
4. Topical agents such as papaverine and lidocaine do not
appear to resolve the perfusion problem.
5. Heparin may make it worse.
6. Administration of intravascular thrombolytic agents (Acti-
vase; Genentech, Inc., South San Francisco, Calif.) provides
instant bright red bleeding from every cut surface, but
thrombus quickly reaccumulates at the anastomotic site.
Microvascular free flap failure can be catego-
the latter group, previously diagnosed contribu-
tions have been stratified using advanced diagnos-
tic paradigms, and patient selection criteria have
been refined to optimize flap success rates.11–15In
contrast, our current series highlights patient fac-
tors that are not routinely or efficiently detected
that appear to have led directly to failure in these
four patients. All patients had arterial anastomosis
failures that were essentially lost on the operating
table, or were taken off the table as desperate
salvage cases with little chance of success. These
patients then required an alternative reconstruc-
tive option for defect closure.
In aggregate, these four patients had five total
free flap losses, two partial flap losses, and one
salvaged thrombus. These four patients represent
30 percent of the flap failures in 325 free flaps
performed by the senior author (S.P.D.) in over a
decade. In addition, they also had one deep vein
thrombosis and two limb ischemic events.
Hypercoagulable events resulting from genetic
and acquired causes are not rare in the general
population. The most common genetic hypercoag-
ulable states include mutations of coagulation fac-
tors V (factor V Leiden gene R506Q mutation) and
II (prothrombin G20210A mutation); mutations of
the methylene tetrahydrofolate reductase gene
ogen activator inhibitor-type 1 gene (4G/5G al-
leles), which modulates endogenous fibrinolytic
activity. Other patients without a genetic predis-
position can develop diseases, such as malignancy
or autoimmune disorders (including antiphospho-
lipid antibody syndrome), that produce a hyperco-
agulable state. Furthermore, certain medications
can promote a procoagulopathic state. The hyper-
coagulable patient can be asymptomatic and undi-
agnosed, and in fact we probably operate on many
patients whose hypercoagulable state is not
known, suspected, or significant. A thorough fam-
Fig. 2. A 42-year-old woman (the patient in case 4) underwent
resection of a recurrent anterior floor-of-mouth carcinoma. The
radial forearm free flap from the ipsilateral arm was used. The
the site of the previous arterial line.
lateral toe became necrotic.
Plastic and Reconstructive Surgery • August 2009
ily medical history is probably the most valuable
tool with which to reveal a genetic suspicion for
However, under certain circumstances, an un-
diagnosed hypercoagulopathy can become clini-
cally relevant. In particular, the patients in our
series were all undergoing microvascular free flap
situ. The combination of an underlying genetic
predisposition to hypercoagulability, plus signifi-
cant operative time (contributing an element of
stasis), plus the presence of a squamous cell car-
cinoma or ductal carcinoma in situ (and their
potential prothrombotic contribution), and the
effects of inflammation on activation of coagula-
tion, all coalesce into a clinical scenario where
our examples, these hypercoagulable events man-
ifested themselves following many hours in the
operating room, during which multiple microa-
nastomoses were unsuccessfully attempted. Given
the significant morbidity of extended operative
time and microvascular reattempts, we suggest
that certain intraoperative findings should raise
the suspicion for a hypercoagulable state, and
abandonment of microvascular repair be consid-
ered pending a thorough hematologic workup
(Table 1). This scenario again emphasizes the
value of a thorough preoperative family medical
Multiple reports have been presented in the
literature of hypercoagulopathic events contrib-
uting to reconstructive complications. In particu-
lar, head and neck reconstruction cases have been
noted to have free flap failure in patients with
underlying coagulopathies.16,17Similarly, undiag-
nosed hypercoagulopathy has been implicated in
free flap failures in reconstruction of the breast,
extremities, and torso.6,9,10,18,19Other hematologic
disorders, and even medications that mediate hy-
percoagulability, have also been insinuated as the
causes of microvascular anastomotic failure.20,21
Indeed, coagulopathy and hematologic disorders
in general have been cited as factors in free flap
outcomes but have been regarded as complex,
difficult to assess preoperatively, and of variable
Unfortunately, no routine, inexpensive, spe-
hypercoagulopathy-related perioperative compli-
cations. Screening for deep venous thrombosis
risk and measurement of prothrombin time and
partial thromboplastin time will identify some pa-
tients with possible coagulopathies; however, the
not detected preoperatively.23Once a hypercoag-
ulable state is suspected—as in multiple microves-
sel anastomotic failures—a series of hematologic
studies should be performed, often in consulta-
tion with the hematology service. The standard
deficiency, protein C and S deficiency, factor V
Leiden, and homocysteine levels (Table 2). In ad-
dition, several endocrine conditions could aggra-
vate the hypercoagulable state, and therefore thy-
roid status (thyroid-stimulating hormone, free
thyroxine) and adrenal function (free AM corti-
sone level, cortisone-stimulating test) should also
be considered. Research tools that measure en-
dogenous thrombin generation or global clot for-
mation (such as those that use the thromboelas-
togram or its variants) might be useful in future
development of effective screening tests.
Management of these patients changes once a
First, the operative defect should be temporized
using dressing changes until more definitive re-
initiated, as the hypercoagulable state places the
Table 1. Clinical Findings Suggestive of a
Hypercoagulable State during Microsurgery
The arterial anastomosis does not flash when the clamp is
removed; it requires manipulation
The arterial anastomosis occludes promptly (within 1 hr)
before the rest of the case is completed
The vessels, particularly the recipient flap vessels, are
Topical agents such as papaverine and lidocaine do not
Unfractionated heparin appears to aggravate the
Administration of intravascular thrombolytic agents (tissue
plasminogen activator) provides instant bright red
bleeding from every cut surface, but thrombus quickly
reaccumulates at the anastomotic site
Table 2. Tests for Genetically Determined
Factor V Leiden gene mutation
Prothrombin A20210G gene mutation
MTHFR gene mutation
PAI-1 gene mutation
Protein C and S activity deficiency
Antithrombin III deficiency
Plasma homocysteine level
Volume 124, Number 2 • Microvascular Free Flap Failure
patient at high risk for postoperative venous throm-
boembolism/deep venous thrombosis. Third, plans
for long-term anticoagulation, if any, need to be
coordinated with the hematology service. Some hy-
percoagulable patients (those with lupus or malig-
nancy) might need protracted anticoagulation,
whereas other patients simply require counseling
on anticoagulation strategies during future scenar-
ios associated with high risk of venous thromboem-
bolism, such as operations, long plane flights, preg-
nancy, and other conditions that might unmask
their predisposition to hypercoagulability.
The plastic surgical literature has only several
case reports of microvascular free flaps compli-
cated by hypercoagulable states; yet if these con-
ditions are present in up to 10 percent of the
population, why are reports of these complica-
tions so scarce? The main reason is that micro-
surgical reconstruction, as in our institution, is
becoming significantly more aggressive. The pa-
tients in this series all had locally advanced squa-
mous cell carcinoma of the head and neck or
breast adenocarcinoma. Most of the patients pre-
sented as recurrences, with preoperative irradia-
tion to this site. All of the patients had significant
tobacco and/or alcohol comorbidity, with nutri-
tional deficiency. Every patient had surgical extir-
pation of the tumor in the same surgical setting
as the initial microsurgical reconstruction, ne-
cessitating operative times in excess of 10 hours
and intraoperative tumor manipulation. Finally,
these patients also had at least one genetic hy-
percoagulable mutation; some patients had up
to three separate mutations. Hypercoagulable
events are often a clinical reflection of syner-
gistic genetic abnormality with environmental
conditions, and this synergy is exemplified in
this patient population. We suspect that as the
indications for free flap reconstruction con-
tinue to expand, this scenario will become in-
More than other malignancies, squamous cell
carcinoma appears to have particular prothrom-
botic effects. It is noteworthy that three of the
recurrent) squamous cell carcinoma, with tumor
manipulation and extirpation immediately pre-
ceding the initial microvascular attempt. The pro-
carcinoma, and the potential for aggravating the
condition with surgical manipulation, could have
in this case series.23–29
Patients with multiple coagulopathic abnor-
have a hematology consultation that includes a
thorough coagulopathic workup. The hematolo-
gist, ideally with the plastic surgeon, should then
counsel the patient as to his or her condition and
the associated risks. Importantly, the patient
should be advised on the use of prophylaxis for
deep venous thrombosis and pulmonary embolus
during future surgical procedures, prolonged im-
mobilization, long car rides, and plane flights.
Steven P. Davison, D.D.S., M.D.
Da Vinci Plastic Surgery
3301 New Mexico Avenue, Suite 236
Washington, D.C. 20016
1. Godina M. Preferential use of end-to-side arterial anastomo-
ses in free flap transfers. Plast Reconstr Surg. 1979;64:673–682.
2. Fogdestam I. Experiences with free flap surgery. Ann Chir
3. Bui DT, Cordeiro PG, Hu QY, Disa JJ, Pusic A, Mehrara BJ.
Free flap reexploration: Indications, treatment, and out-
comes in 1193 free flaps. Plast Reconstr Surg. 2007;119:2092–
4. Okazaki M, Asato H, Takushima A, et al. Analysis of salvage
treatments following the failure of free flap transfer caused
by vascular thrombosis in reconstruction for head and neck
cancer. Plast Reconstr Surg. 2007;119:1223–1232.
5. Chen KT, Mardini S, Chuang DC, et al. Timing of presen-
tation of the first signs of vascular compromise dictates the
salvage outcomes of free flap transfers. Plast Reconstr Surg.
Nikolis A. Heparin-induced thrombocytopenia syndrome as
a cause of flap failure: A report of two cases. J Plast Reconstr
Aesthet Surg. 2008;61:78–83.
7. Wax MK, Rosenthal E. Etiology of late free flap failures
occurring after hospital discharge. Laryngoscope 2007;117:
8. Bozikov K, Arnez ZM. Factors predicting free flap compli-
cations in head and neck reconstruction. J Plast Reconstr
Aesthet Surg. 2006;59:737–742.
9. Baumeister S, Follmar K, Zenn M, Erdmann D, Levin LS.
Strategy for reoperative free flaps after failure of a first flap.
Plast Reconstr Surg. 2008;122:962–971.
10. Vekris MD, Ovrenovits M, Dova L, et al. Free functional
a potential risk factor: A case report. Microsurgery 2007;27:
11. Serletti JM, Higgins JP, Moran S, Orlando GS. Factors af-
fecting outcome in free-tissue transfer in the elderly. Plast
Reconstr Surg. 2000;106:66–70.
12. Nahabedian MY, Momen B, Manson PN. Factors associated
with anastomotic failure after microvascular reconstruction
of the breast. Plast Reconstr Surg. 2004;114:74–82.
13. Nahabedian MY, Singh N, Deune EG, Silverman R, Tufaro
AP. Recipient vessel analysis for microvascular reconstruc-
tion of the head and neck. Ann Plast Surg. 2004;52:148–155;
Plastic and Reconstructive Surgery • August 2009
optimization of microsurgery in atherosclerotic patients.
ing curve in microvascular head and neck reconstruction.
Arch Otolaryngol Head Neck Surg. 1997;123:1332–1335.
16. Ayala C, Blackwell KE. Protein C deficiency in microvascular
head and neck reconstruction. Laryngoscope 1999;109:259–
17. Di Spilimbergo SS, Rampazzo A, Squaquara R, Penzo M,
of head and neck reconstruction resolved with systemic ther-
apy. Plast Reconstr Surg. 2008;121:1508–1509; discussion
18. Olsson E, Hoijer P. Activated protein C resistance due to
factor V Leiden, elevated coagulation factor VIII, and post-
operative deep vein thrombosis in late breast reconstruction
with a free TRAM flap: A report of two cases. Br J Plast Surg.
19. Labow BI, Green AK, Upton J. Homocystinuria: An unrec-
ognized cause of microvascular failure. Plast Reconstr Surg.
20. Al-Sukhun J, Koivusalo A, Tornwall J, Lindgvist C. COX-2
major lower extremity trauma: Mechanism and possible role
in free flap failure. Ann Plast Surg. 1996;36:489–494.
22. Hoijer P, Olsson E. Elevated coagulation factor VIII, post-
operative thrombosis and flap failure in late breast recon-
struction with a free TRAM flap: A case report. J Plast Reconstr
Aesthet Surg. 2006;59:102–104.
23. Davison SP, Venturi ML, Attinger CE, Baker SB, Spear SL.
Prevention of venous thromboembolism in the plastic sur-
gery patient. Plast Reconstr Surg. 2004;114:43E–51E.
24. Avances C, Jacot W, Senesse P, Culine S. Prompt resolution
of acute disseminated intravascular coagulation with do-
cetaxel and cisplatin in hormone refractory prostate cancer.
J Urol. 2002;168:1496.
25. Navarro M, Ruiz I, Martin G, Cruz JJ. Patient with dissemi-
nated intravascular coagulation as the first manifestation of
adenocarcinoma of the prostate: Risks of prostatic biopsy.
Prostate Cancer Prostatic Dis. 2006;9:190–191.
26. Ng HJ, Crowther MA. Malignancy-associated venous throm-
limb gangrene and thrombotic microangiopathy. Thromb
27. Koenig MA, Maleszewski J, Winters B. Multiple hemorrhagic
strokes from DIC associated with occult large cell carcinoma.
Neurocrit Care 2006;5:210–212.
28. Zurada JM, Ward KM, Grossman ME. Henoch-Scho ¨nlein
purpura associated with malignancy in adults. J Am Acad
Thromb Haemost. 1999;82:1005–1007.
Contacting the Editorial Office
To reach the Editorial Office, please use the following contact information:
Plastic and Reconstructive Surgery
Rod J. Rohrich, M.D., Editor-in-Chief
St. Paul’s Hospital
5909 Harry Hines Boulevard
Dallas, Texas 75235-8820
Volume 124, Number 2 • Microvascular Free Flap Failure