The Laryngoscope
Lippincott Williams & Wilkins
© 2007 The American Laryngological,
Rhinological and Otological Society, Inc.
Etiology of Late Free Flap Failures
Occurring After Hospital Discharge
Mark K. Wax, MD; Eben Rosenthal, MD
Objectives: Vascular compromise of free flaps most
commonly occurs in the immediate postoperative period
in association with failure of the microvascular anasto-
mosis. Rarely do flaps fail in the late postoperative period.
It is not well understood why free flaps can fail after 7
postoperative days. We undertook a case review series to
assess possible causes of late free flap failure.
Study Design: Retrospective review at two tertiary
referral centers: Oregon Health Sciences University and
University of Alabama at Birmingham.
Methods: A review of 1,530 flaps performed in 1,592
patients between 1998 and 2006 were evaluated to iden-
tify late flap failure. Late flap failure was defined as
failure occurring after postoperative day 7 or on follow-up
visits after hospital discharge. A prospective database
with the following variables was examined: age, medical
comorbidities, postreconstructive complications (fistula or
infection), hematoma, seroma, previous surgery, radia-
tion therapy, intraoperative findings at the time of de-
bridement, nutrition, and, possibly, etiologies.
Results: A total of 13 patients with late graft
failure were identified in this study population of
1,530 (less than 1%) flaps; 6 radial forearm fasciocu-
taneous flaps, 2 rectus abdominis myocutaneous
flaps, 4 fibular flaps, and 1 latissimus dorsi myocu-
taneous flap underwent late failure. The time to ne-
crosis was a median of 21 (range, 7–90) days. Etiol-
ogy was believed to possibly be pressure on the
pedicle in the postoperative period in four patients
(no sign of local wound issues at the pedicle), infec-
tion (abscess formation) in three patients, and re-
growth of residual tumor in six patients. Loss occur-
ring within 1 month was more common in radial
forearm flaps and was presented in the context of a
normal appearing wound at the anastomotic site, as
opposed to loss occurring after 1 month, which
From the Department of Otolaryngology Head and Neck Surgery
(M.K.W.), Oregon Health Sciences University, Portland, Oregon, U.S.A.;
and the Department of Otolaryngology (E.R.), University of Alabama at
Birmingham, Birmingham, Alabama, U.S.A.
Editor’s Note: This Manuscript was accepted for publication May 29, 2007.
Presented at the Triological Society Combined Sections Meeting,
Marco Island, Florida, U.S.A., February 16, 2007.
Send correspondence to Dr. Mark K. Wax, Department of Otolaryn-
gology Head and Neck Surgery, 3181 S.W. Sam Jackson Park Road, EV01,
Portland, OR 97201. E-mail: waxm@ohsu.edu
DOI: 10.1097/MLG.0b013e31812e017a
Laryngoscope 117: November 2007
happened more commonly in fibula flaps secondary to
recurrence.
Conclusion: Although late free flap failure is rare,
local factors such as infection and possibly pressure on the
pedicle can be contributing factors. Patients presenting
with late flap failure should be evaluated for residual tu-
mor growth.
Key Words: Free flap, failure, reconstruction, cancer.
Laryngoscope, 117:1961–1963, 2007
INTRODUCTION
Surgery continues to be a major treatment modality
in patients with head and neck cancer. Whether it is used
primarily or as a salvage treatment for patients who have
failed chemoradiation protocols, a significant number of
patients with head and neck cancer will undergo some
form of surgical extirpation. These patients will often be
left with large composite tissue defects. Reconstruction of
these composite defects is best accomplished with replace-
ment of similar vascularized composite tissue. Thus, free
tissue transfer has become one of the most common mo-
dalities for reconstruction in these patients.1
Contemporary survival rates for free tissue transfer
are generally accepted to be above 95%. Patients whose
free flaps fail most commonly will do so within a relatively
short period of time.2 In the majority of cases, technical
issues having to do with anastomosis or with pedicle ge-
ometry are the usual causative factors.2–5 Occasionally,
infection or unknown etiology is seen. Rarely are free flaps
lost after patients have been discharged from the hospital.
We undertook to evaluate a case series of patients whose
free flaps had died in the late postoperative phase. We
sought to evaluate whether there were any causative fac-
tors that could describe this patient population.
MATERIALS AND METHODS
All patients undergoing free tissue transfer at the Univer-
sity of Alabama Birmingham and the Oregon Health and Sci-
ences University by the two senior authors are recorded in a
microvascular reconstructive database. This database was
searched for all patients who underwent total loss of their free
flap. Patients were then divided into two groups. The first group
of patients were those who underwent flap loss within 7 days.
These were judged to be acute flap failures and were not analyzed
for the purposes of this report. Patients whose flap was noted to
Wax and Rosenthal: Free Flap Failures After Hospital Discharge
1961
 TABLE I.
Type and Number of Flaps Lost.
Type of Flap Number Lost
Radial forearm 6 area to another depends on obtaining a certain expertise
Rectus abdominus 2 and experience with microvascular reconstructive tech-
Fibular osteocutaneous 4 niques. In experienced hands, the ability to transfer this
Latissimus dorsi 1 tissue has a survival rate of well over 95%. Although
Total 13
be dead on follow-up visits were further examined. Charts were
evaluated to ensure that the free tissue transfer was alive on
discharge and then necrotic on follow-up. Patients who had de-
veloped infections while in the hospital and whose flaps were lost
were not included in the study population. Once the appropriate
patients were identified, a retrospective chart review was per-
formed. Inpatient and outpatient charts were evaluated for de-
mographic data, type of flap used, and indications for its use. The
postoperative course of the patient was reviewed, and the time
when the flap was recognized to be dead was noted. Associated
complications and the status of the patient’s surrounding dead
tissues at the time of flap failure were recorded. Any sign of
infection or tumor recurrence were specifically noted.
RESULTS
From 1998 to 2006, 1,530 flaps in 1,592 patients were
transferred. Thirteen flaps were lost more than 7 days
after transfer or after discharge. The median time to flap
loss was 21 days, with a range of 7 to 90 days. The time
course of flap loss is demonstrated in Table I. All patients
received antibiotics in the perioperative period for up to 7
days, as per the surgeon’s protocol.
The type of free tissue transfer that was lost is given
in Table II. The majority of flaps were radial forearm and
fibular osteocutaneous. The ratio of the individual type of
flap in the overall flap population compared with the ratio
of flap loss in this population was examined and was not
different.
Flaps were evaluated as to the proposed etiology of
why they were lost. Three categories were defined: pa-
tients who lost their flaps within 30 days with no physical
signs or findings that would indicate local wound issues at
the anastomotic or pedicle site (infection included); those
who lost them secondary to late infection and abscess
formation; and those who were found to have residual
tumor on debridement of the necrotic flap. There were four
patients who lost their flaps in the 1 month postoperative
period, three who lost their flaps secondary to infection,
and 6 who demonstrated recurrence when they presented
with flap loss.
TABLE II.
Time From Flap Transfer to When Flaps Were Noted to
be Unsalvageable.
Time to Flap Loss (days) Number Lost
7–14 4 the flap succumbing.
14–30 4 These infections occurred in patients who had had
⬎30 5
Laryngoscope 117: November 2007
1962
DISCUSSION
Free tissue transfer has become the standard for
reconstruction of ablative defects of the head and neck.
The ability to successfully transfer tissue from one body
multiple interventional modalities have been attempted to
improve the survival rate of free tissue transfer, ulti-
mately, the experience of the operator yields a survival
level that is very difficult to modify.3
There are many reasons why a free tissue transfer
may fail. In the immediate postoperative period (less than
72 hr), venous thrombosis appears to be the most common
etiology, with an occasional arterial thrombosis being
seen. Venous thrombosis is believed to be secondary to
poor pedicle geometry or to technical issues at the anas-
tomotic site.4
In addition to vascular compromise in the first 5
days, there is a poorly documented incidence of late flap
failure. During the course of transferring more than 1,000
free flaps, the senior authors have encountered a very
small number of flaps that failed after 7 days or after the
patient had left the hospital. The etiology of this loss is not
well defined. A search of the literature was not very fruit-
ful. A review of the literature looking at large flap series
and those reports that focus on flap loss or salvage in-
cluded all cases of flap loss. There was no report that dealt
with only late flap loss. In series that described late flap
loss, these cases were included with all flaps that were
salvaged or attempted and thus were not individually
described. In our series of 13 patients, we found that late
flap failure tended to occur in three time frames: 1) within
the first 2 weeks, 2) after 1 month, or 3) after 3 months. All
of the flaps that failed in the first 2 week time period were
noted to be alive and doing well on discharge from hospi-
tal. The majority of them were doing well on the patient’s
first postoperative visit and were noted at the second
postoperative visit to be dead. A number of patients made
an anecdotal observation that the flap was fine in the
evening, and then when they woke up, it was noted to be
black. Interestingly, all of these flaps had been used for
maxillary sinus or lateral scalp defect reconstructions.
The pedicle had been tunneled either under the scalp to
the superficial temporal artery or under the cheek and
over the mandible to the facial artery. We theorized that
the patient had compressed the flap pedicle during the
night. This then resulted in flap ischemia and ultimately
flap death.
The second category of patients who had late flap
failure did so 1 month to 3 months after transfer of the
flap. These patients developed late deep neck abscesses
and infections. All of them had had an uncomplicated
postoperative course, with the patient being discharged
home. In each case, the patient developed a late neck
infection ultimately with pus collecting in the wound and
previous osteoradionecrosis or who had had a fistula with
infection postoperatively. They had all been treated with
Wax and Rosenthal: Free Flap Failures After Hospital Discharge
intravenous antibiotics, and their infections had resolved
in the hospital. On discharge, they again were noted to do
well for 3 to 4 weeks when their infection then flared up,
and they developed a recurrent neck abscess. We surmise
that because the patients took 1 or 2 days, at minimum,
sometimes up to 4 or 5, before they presented with the
neck abscess, that the development of pus in the wound
was enough to overcome the natural healing process and
occluded the vascular supply to the pedicle. Once again,
comparison of these patients with others with infection
did not yield any predisposing factors.6
When tissue is transferred from one area of the body
to another, it is dependent on the vascular supply to which
it is anastomosed. Anastomosis between the donor and
recipient vessels insures adequate inflow of arterial blood
and outflow of venous blood. Anything that compromises
this arterial inflow or venous outflow will compromise the
survival of the flap. In the immediate postoperative phase,
vascular compromise is most often caused by pedicle ge-
ometry problems or by technical issues with the anasto-
mosis itself. It would appear that in our overall patient
population with more than 1,500 flaps, a small incidence
of patients lost their flaps after 1 month and suffered a
late flap failure phenomenon. Unfortunately, because this
is a retrospective review, it is difficult to ascertain what
separated these patients from the other patients in the
series who had the same type of flap performed in a
similar fashion but whose flaps did not fail. There was no
clustering of flap failures around a certain time period,
there did not appear to be a particular flap that was prone
to failure, and although we did examine the surgical ap-
proach in these cases, we were unable to conclude that the
surgical technique had created a tunnel that connected
the flap site to the vessel recipient site at too superficial a
plane.5 Interestingly, of the flaps that were lost in the
early time period, three of four were radial forearm flaps,
whereas one was a rectus flap.
The last group of patients represents a subpopulation
in which revascularization of the flap did not occur. Once
the tissue is transferred up to the head and neck area, it
is dependent on its vascular pedicle. However, as time
goes on, a revascularization phenomenon occurs, with in-
growth of blood vessels and communication between the
donated tissue and the recipient bed. Animal studies have
demonstrated that in a rat fasciocutaneous model, if the
pedicle is left intact for 7 days, one can expect almost 100%
survival of the flap.7 The rat groin flap is unlike most flaps
used in head and neck reconstruction. It is very thin,
almost like a full-thickness skin graft. Furthermore, the
recipient bed in this model is not comparable with the
usual head and neck patient. Thus, the translation of this
finding to humans is tenuous at best. The amount of time
required until the pedicle can be divided in humans re-
mains unknown. Anecdotally, we have divided the pedicle
a number of times for a variety of reasons. Usually, it has
Laryngoscope 117: November 2007
been after 3 months, and we have not encountered flap
loss phenomena.
In our third subgroup of patients, residual tumor was
present and regrew at the site where the flap was inset.
We would surmise that the residual tumor prevented ad-
equate revascularization of the flap, and as the tumor
grew and invaded into the surrounding tissues, it gradu-
ally occluded the vascular pedicle. Once the vascular pedi-
cle was occluded, with a consequent lack of revasculariza-
tion, the flap was no longer able to survive and became
necrotic. Because of the length of time the flap had been in
place in most of these cases, infection did not develop, but
exploration revealed a necrotic flap with recurrent tumor.
The retrospective nature of this review along with the
small number of cases makes it difficult to draw any
conclusions as to whether other ancillary measures could
have been instituted to prevent late flap loss. Certainly,
the flaps that were lost early and perhaps because of
compression constitute an interesting series. We have
taken to instructing patients to avoid lying on the flap side
or putting any form of compression on the site where the
vascular pedicle is. However, we note that the external
jugular vein is frequently the only vascular outflow for a
significant number of flaps, and this has not proven to be
a problem that contributes to flap loss from compression.
CONCLUSION
Loss of free tissue that has been transferred to the
head and neck for reconstruction after oncologic ablation
is a devastating phenomenon. There are a small number
of patients who will lose their flaps 1 week or more after
discharge. Possible compression of the vascular pedicle,
late infection, and residual tumor are contributing factors
and should be considered in these cases.
BIBLIOGRAPHY
1. Rosenthal EL, Wax MK. Simplification of microvascular recon-
struction. Head Neck 2004;26:930 –936.
2. Haughey BH, Wilson E, Kluwe L, et al. Free flap reconstruc-
tion of the head and neck: analysis of 241 cases. Otolaryn-
gol Head Neck Surg 2001;125:10 –17.
3. Suh JD, Sercarz JA, Abemayor E, et al. Analysis of outcome
and complications in 400 cases of microvascular head and
neck reconstruction. Arch Otolaryngol Head Neck Surg
2004;130:962–966.
4. Singh B, Cordeiro PG, Santamaria E, Shaha AR, Pfister DG,
Shah JP. Factors associated with complications in micro-
vascular reconstruction of head and neck defects. Plast
Reconstr Surg 1999;103:403– 411.
5. Genden EM, Rinaldo A, Suarez C, Wei WI, Bradley PJ,
Ferlito A. Complications of free flap transfers for head
and neck reconstruction following cancer resection. Oral
Oncol 2004;40:979 –984.
6. Huang RY, Sercarz JA, Smith J, Blackwell KE. Effect of sali-
vary fistulas on free flap failure: a laboratory and clinical
investigation. Laryngoscope 2005;115:517–521.
7. Partsafas AW, Jorgensen SA, Bascom DA, Wax MK. Effects of
two sealents (Floseal and Tisseel) on fasciocutaneous flap
revascularization. Arch Facial Plast Surg 2003;5:399 – 402.
Wax and Rosenthal: Free Flap Failures After Hospital Discharge
1963