Keywords
limb salvage - regenerative - pediatric - reconstruction - extremity - free flap
Complex pediatric extremity wounds with osseous, tendon, or joint exposures are a
rare yet devastating problem. As a key component of limb salvage (LS), soft tissue
reconstruction is necessary to avoid osteomyelitis and tendon or joint rupture, which
would result in disability or amputation.[1] Presently, the dominant LS paradigm employs the reconstructive elevator wherein
free flap-based LS (fLS) is the recommended first-line treatment paradigm.[2]
[3] This paradigm developed due to the significantly higher failure rates for the lower
“rungs” of the reconstructive ladder, such as skin grafts and local flaps.[4] However, fLS is resource-intensive and costly.[5]
[6] Patients who do not live near a pediatric hospital may therefore not have access
to fLS. From a technical perspective, pediatric fLS can be more difficult than in
adults due to smaller vasculature,[7]
[8]
[9]
[10] and the challenges of limiting postoperative activity and limb positioning in children.[5]
To overcome these challenges, our group developed a novel, regenerative LS (rLS) technique
using dehydrated human amnion/chorion membrane (dHACM).[11] In our pilot experience in adults, 86% of rLS patients were successfully reconstructed.
The mechanism of action for dHACM is the gradual and sustained release of over 300
growth factors, antimicrobial peptides, and immunomodulatory and anti-inflammatory
molecules. rLS leverages dHACM to stimulate the growth of stable granulation tissue
over exposed structures that traditionally do not generate significant granulation
tissue, such as cortical bone stripped of periosteum, tendon missing paratenon, and/or
joint articular capsules. Once a stable granulation base is regenerated, definitive
closure is achieved with split-thickness skin grafts (STSG).[11]
[12]
Compared with adults, pediatric patients generally have greater regenerative potential.[13] Given the success of rLS in adults, we extended the application of rLS to pediatric
patients. The aim of this study was to quantify the efficacy, safety, and cost-effectiveness
of rLS in complex pediatric extremity wounds in comparison to fLS. Despite our hospital's
status as the largest volume, free-standing pediatric hospital in the state of Louisiana,
the number of pertinent cases remained small, thus necessitating a retrospective study
design.
Methods
Study Design
This was an institutional review board-approved, retrospective cohort study. The study
population was comprised of pediatric patients (age < 18 years) presenting to a single
nonprofit children's hospital. Nine consecutive pediatric patients with complex extremity
wounds required surgical intervention over a 2-year period (December 2014–December
2016). Based on surgeon preference, these patients underwent either free fLS or rLS.
Demographic variables collected for each case included age, sex, mechanism of injury,
wound size, and the structures exposed. Operative variables of interest focused on
the following: type of flap selection, operative time (minutes), total number of follow-up
dHACM applications, total number of operations, and unplanned surgical intervention.
Complications were defined as a total or partial flap loss, wound dehiscence, and
any donor site morbidity. Postoperative variables included the number of pediatric
intensive care unit (PICU) stays, inpatient hospital stays, and time to definitive
wound closure (defined as complete wound closure without discharge). Primary outcome
measures included length of intensive care stay, rate of peri- and postoperative complication,
and time to definitive wound closure. Due to the small sample sizes of each group,
nonparametric testing utilizing the Wilcoxon Mann–Whitney U test was performed with statistical significance set at p < 0.05 (SAS version 9.4).
rLS Technique
For rLS, the management algorithm consisted of initial excisional debridement of all
contaminated, nonviable tissue followed by assessment of the size and location of
exposed critical structures. The placement of dHACM (EpiFix; MiMedx Group Inc.) was
custom fit and applied to exposed bone or tendon ([Video 1]), moistened with sterile saline, and bolstered with a nonadherent contact layer
(Adaptic; Acelity). The bolster was secured with suture or surgical staples to avoid
loss of the dHACM during outer dressing changes. The outer dressing material was customized
to the anticipated amount of wound exudate. For less exudative wounds, moist-to-dry
sterile gauze was applied and changed every 2 to 3 days. For more exudative wounds,
a silver-impregnated absorbent dressing (Mepilex Ag; Mölnycke) was applied and changed
after 7 days. Wounds were assessed weekly for infection and the generation of an adequate
granulation base in the operating room. If needed, repeat debridement was performed,
dHACM was reapplied, or definitive closure with an STSG and a negative pressure dressing
bolster was performed.
Video 1 Application of dHACM.
Results
Patient demographics and wound characteristics were similar between groups. The mean
ages were 11.7 years in rLS versus 9.7 years in fLS (p = 0.79). The mean wound size was 141.6 cm2 in rLS versus 154.0 cm2 in fLS (p = 0.56). All injuries were associated with exposed critical structures of the lower
extremity, including exposed bone (rLS n = 1, fLS n = 3), exposed bone and tendon (rLS n = 1, fLS n = 1), exposed bone, tendon, and cartilage (rLS n = 1, fLS n = 0), exposed tendon only (rLS n = 1, fLS n = 0), or exposed tendon and cartilage (rLS n = 1, fLS n = 0). The mean length of follow-up was 313.8 days (range: 54–498 days) in rLS and
154.0 days (range: 21–306 days) in fLS ([Table 1]).
Table 1
Patient characteristics
|
rLS
|
fLS
|
p
[a]
|
Number
|
5
|
4
|
|
Gender (M:F)
|
3:2
|
2:2
|
|
Mean age ± SD, y
|
11.7 ± 5.4
|
9.7 ± 6.7
|
0.794
|
Mean duration of follow-up, d
|
313.8 ± 309.7
|
154.0 ± 134.1
|
0.556
|
Mean wound size ± SD, cm2
|
141.6 ± 140.1
|
154.0 ± 131.4
|
0.556
|
Mechanism of injury
|
|
|
|
Motor vehicle collision
|
1
|
0
|
|
Off-road vehicle injury
|
2
|
0
|
|
Lawn mower injury
|
1
|
2
|
|
Gunshot wound
|
1
|
0
|
|
Tumor resection
|
0
|
2
|
|
Critical structures exposed
|
|
|
|
Bone
|
1
|
3
|
|
Bone + tendon
|
1
|
1
|
|
Bone + tendon + cartilage
|
1
|
0
|
|
Tendon
|
1
|
0
|
|
Tendon + cartilage
|
1
|
0
|
|
Abbreviations: fLS, flap-based limb salvage; rLS, regenerative limb salvage; SD, standard
deviation.
Note: Nine patients underwent limb salvage. There were no significant age, gender,
or wound size differences between the two groups.
a Statistical significance for Wilcoxon Mann–Whitney U test at 0.05 α level.
Overall, rLS patients required 64.5% less operative time than fLS patients (132.6
vs. 373.0 minutes, p = 0.016; [Fig. 1A]). The time to definitive wound closure was 85.8% shorter in rLS (13.8 vs. 97.5 days,
p = 0.016; [Fig. 1B]). The mean number of dHACM applications in the rLS group was 1.40 (range: 1–2).
Fig. 1 (A) Total operative time. Mean operative times were 132.6 and 373.0 minutes in regenerative
limb salvage (rLS) and flap-based limb salvage (fLS), respectively (p = 0.016). In rLS, the mean total operative time was reduced by 64.5% compared with
fLS. *Statistical significance for Wilcoxon Mann–Whitney U test set at 0.05 α level. (B) Time to definitive reconstruction. Mean times to definitive wound closure were 13.8
and 97.5 days in rLS and fLS, respectively (p = 0.016). In rLS, definitive reconstruction was achieved 85.8% faster than fLS. *Statistical
significance for Wilcoxon Mann–Whitney U test set at 0.05 α level.
rLS patients experienced no postoperative complications. In contrast, 75% of fLS patients
experienced complications including flap loss, wound dehiscence, and venous thromboses.
Furthermore, 50% of fLS patients required an unplanned return visit to the operating
room (0% for rLS). Additional operative time due to complications averaged 70.7 minutes
in fLS ([Table 2]).
Table 2
Complications
|
rLS
|
fLS
|
Any complication
|
0
|
5
|
Total flap loss
|
–
|
1
|
Partial flap loss
|
–
|
1
|
Wound dehiscence
|
0
|
1
|
Venous thrombosis
|
0
|
1
|
Donor site complication
|
0
|
1
|
Unplanned return to operating room
|
0
|
2
|
Mean reoperative time ± SD, min
|
0
|
70.7 ± 8.96
|
Abbreviations: fLS, flap-based limb salvage; rLS, regenerative limb salvage; SD, standard
deviation.
Note: rLS produced fewer postoperative complications. rLS required fewer unplanned
returns to the operating room and less reoperative time. However, numbers in the rLS
group were too small for statistical comparison.
No rLS patients required PICU admission, whereas fLS patients averaged 1.3 days for
flap monitoring. In fLS (n = 4) patients, the free flaps used were vertical profunda artery perforator (n = 2), latissimus do.rsi (n = 1), and anterolateral thigh (n = 1) ([Table 3]). By completion of follow-up, all patients (100.0%) in both groups had resumed ambulation.
None developed infection at any phase of treatment. Example rLS cases are presented
below.
Table 3
Comparison of operative outcomes
|
rLS
|
fLS
|
p
[a]
|
% change
|
Mean PICU stay ± SD, d
|
0
|
1.3 ± 0.6
|
|
|
Mean inpatient hospital stay ± SD, d
|
9.4 ± 4.5
|
15.3 ± 11.5
|
0.564
|
–38.6%
|
Flap selection
|
|
|
|
|
VPAP
|
–
|
2
|
|
|
LD
|
–
|
1
|
|
|
ALT
|
–
|
1
|
|
|
Mean dHACM application ± SD
|
1.40 ± 0.55
|
–
|
|
|
Mean operative time ± SD, min
|
132.6 ± 84.9
|
373.0 ± 140.1
|
0.016[a]
|
–64.5%
|
Mean time to definitive wound closure ± SD, d
|
13.8 ± 1.9
|
97.5 ± 76.3
|
0.016[a]
|
–85.8%
|
Abbreviations: ALT, anterolateral thigh; dHACM, dehydrated human amnion/chorion membrane;
fLS, flap-based limb salvage; LD, latissimus dorsi; PICU, pediatric intensive care
unit; rLS, regenerative limb salvage; SD, standard deviation; VPAP, vertical profunda
artery perforator.
Note: rLS requires significantly fewer PICU stays and fewer inpatient hospital stays.
In rLS, operative time and time to definitive wound closure were significantly reduced
by 64.5 and 85.8%, respectively.
a Statistical significance for Wilcoxon Mann–Whitney U test at 0.05 α level.
Case 1: Gustilo IIIB Open Tibial Fracture
A 14-year-old boy presented with a Gustilo IIIB open tibial fracture of his left lower
extremity following an off-road vehicle accident. He underwent excisional debridement
and fracture stabilization with an external fixator ([Fig. 2A]). The exposed tibia was stripped of periosteum ([Fig. 2B]). The cortex was burred to pinpoint bleeding and a single layer of dHACM was applied.
After 1 week, adequate granulation tissue developed, and definitive closure with STSG
was successful ([Fig. 2C]). The wound completely healed in 3 weeks and remained stably closed at 18 months
([Fig. 2D, E]).
Fig. 2 Case 1: Gustilo IIIB open tibial fracture. A 14-year-old male underwent initial sharp
debridement and stabilization of a Gustilo IIIB open tibial fracture with an external
fixator (A). Exposed tibia without periosteum was covered by undermining of soft tissue envelope
and advancing using tension sutures (B). A single dehydrated human amnion/chorion membrane (dHACM) was applied over the
exposed anterior tibia, adequate granulation tissue developed, and split-thickness
skin grafts (STSG) was successful (C). The wound completely healed in 3 weeks and remained stably closed at 18-month follow-up
(D, E).
Case 2: Exposed Metatarsal Joint
A 3-year-old boy presented with exposed metatarsal cartilage following a riding lawnmower
accident ([Fig. 3A, B]). He underwent initial excisional debridement followed by one dHACM treatment, which
yielded granulation tissue coverage of the exposed cartilage after 1 week. STSG with
a negative pressure dressing bolster was used to achieve definitive coverage ([Fig. 3C]). When the bolster was taken down on postoperative day 5, a small area (< 0.1 cm2) of exposed cartilage persisted ([Fig. 3D]). A second treatment of dHACM was applied. Two weeks later, the wound was fully
healed. The wound remained stably closed at 2-month follow-up, and he resumed ambulation
([Fig. 3E]; [Video 2]).
Video 2 Patient's return to ambulation at 2-month follow-up.
Fig. 3 Case 2: Exposed metatarsal joint. A 3-year-old male with exposed cartilage and tendon
on bilateral feet following lawn mower injury. Regenerative limb salvage (rLS) was
performed. He underwent debridement of all necrotic tissue (A, B). One dehydrated human amnion/chorion membrane (dHACM) treatment was performed followed
by split-thickness skin graft (C). On postoperative day 5, a small area (< 0.1 cm2) of exposed cartilage persisted (D). A second treatment of dHACM was applied for definitive wound closure. Two weeks
later, the wound was clean and fully healed with no palpable cartilage. The wound
remained stably closed at 2-month follow-up and he has resumed ambulation (E).
Case 3: Multiple Areas of Exposed Bone and Tendon
A 17-year-old girl presented with a complex open wound of her right lower leg and
foot following a motor vehicle collision ([Fig. 4A]). After initial excisional debridement, examination revealed two areas of exposed
bone and tendon. Exposed cortex was burred to pinpoint bleeding and one dHACM sheet
was applied, followed by STSG a month later ([Fig. 4B, C]). The wound remained stably closed at the 2-month follow-up and the patient resumed
ambulation ([Fig. 4D]).
Fig. 4 Case 3: Multiple areas of exposed bone and tendon. A 17-year-old female presented
following a motor vehicle collision with multiple areas of exposed bone and tendon
on her right leg and foot (A). The medial leg and foot received a single dehydrated human amnion/chorion membrane
(dHACM) treatment followed by split-thickness skin grafts (STSG) (B, C). The wound remained stably closed at 2-month follow-up and the patient resumed ambulation
(D).
Discussion
A recent systematic review of over 1,300 adult fLS patients reported success rates
ranging from 73 to 100%, with a pooled rate of 92.7%.[14] In this study, we found that pediatric rLS and fLS both yielded 100% reconstructive
success and 100% return-to-ambulation. However, rLS was advantageous in several key
areas, including shorter operative times, fewer operations, fewer complications, and
the avoidance of PICU admission. Additional advantages of rLS include the option to
apply dHACM in an outpatient setting and increased patient access to pediatric LS
by removing the requirement for microsurgical expertise.
When limited to free flaps, the fLS paradigm is technically demanding, costly, and
limits patient access.[5]
[6] These barriers have spurred innovation in fLS to reduce the reliance on free tissue
transfer, and recent, notable innovations include propeller flaps and keystone flaps.[15]
[16]
[17]
[18] However, these innovations do not alter the dominance of the fLS paradigm.
In terms of technical challenges, consistent success at fLS requires microsurgical
training and experience, as well as a sufficient case load to maintain those skills.
In specific cases, achieving microvascular anastomosis outside the zone of injury
can prove difficult. Our third case is one such example: the zone of injury extended
from foot to groin ([Fig. 4]). Alternatively, injuries such as those to the plantar surface often require ultra-thin
flaps that may be unavailable given the growing epidemic of obesity. In the pediatric
population, smaller vessel size and increased arterial spasm make free tissue transfer
challenging even for experienced microsurgeons.[7]
[8]
[9]
[10] Multiple studies have shown that anastomoses involving vessels with a diameter < 0.7 mm
increase the risk of flap failure due to the technical difficulty.[5]
[6]
[19] Furthermore, postoperative activity restrictions and positioning requirements are
notoriously difficult to impose on pediatric patients, yet crucial for flap survival.
In contrast, rLS does not require microsurgical expertise but also preserves all fLS
options as reconstructive lifeboats.
fLS costs are high, with mean 2-year costs of $81,316 and an estimated lifetime cost
of over $200,000 in pediatric patients.[20] The Lower Extremity Assessment Project study found that fLS' high costs are partly
due to prolonged hospital stays (range: 2–85 days), multiple surgical procedures (1–7
procedures), and frequent rehospitalization (60.7% required readmission).[21] We found similar patterns in the pediatric fLS patients in this study: they required
significantly longer hospital stays, more surgical procedures, and experienced more
perioperative complications. By requiring significantly fewer inpatient resources,
rLS should yield cost savings.
Patient access to fLS is often limited because surgeons with the training, experience,
and willingness to perform free-tissue transfers for trauma tend to cluster in urban
environments and academic medical centers.[22] The result is that LS candidates in low socioeconomic or rural areas may never be
offered fLS due to the lack of hospitals offering specialty care.[22] Pediatric LS candidates are further burdened by the sparsity of pediatric hospitals
with microsurgical capabilities. Fewer than 15.0% of pediatric hospitals are considered
high-volume centers (> 1,000 pediatric surgical admissions per year).[22] In fact, 71.0% of U.S. hospitals admit fewer than 250 pediatric surgery patients
per year and, as a whole, see less than 20.0% of all pediatric plastic surgery cases
nationwide.[22] The low pediatric LS volume feeds back to the technical challenges of fLS, as studies
have shown that higher case volumes drive improved outcomes, especially in the pediatric
population.[22]
[23]
[24]
[25]
[26]
[27]
[28] Restoring the reconstructive ladder for LS would address these challenges.
Other non-fLS strategies are reported in the literature, such as the use of acellular
wound matrices (AWMs) that serve as scaffolding for the ingrowth of granulation tissue.
These matrices include collagen matrices, human acellular dermal matrices, and porcine
urinary bladder matrices.[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38] We distinguish rLS from these strategies because dHACM functions not through scaffolding,
but instead through the elution of > 300 bioactive molecules that modulate inflammation,
promote tissue regeneration, and fight infection.[39]
[40]
[41] Whereas we have experience no wound infections during the granulation tissue phase
in any of our adult and pediatric rLS patients, the use of AWM for coverage of exposed
bone has a 9 to 20.0% infection rate.[42]
[43]
Several limitations exist in this study. First, due to the low incidence of pediatric
LS, the sample sizes were small and not adequately powered to detect many potentially
significant differences between the two cohorts. Second, the retrospective nature
of the study design suggests the possibility of selection bias. Third, our findings
are limited to patients presenting to a single institution in an urban setting, and
therefore limit the external validity of our results.
Conclusion
This study demonstrates that rLS is a safe and efficacious reconstructive option for
pediatric LS patients. rLS patients achieved definitive closure more quickly than
their fLS counterparts. Moreover, rLS patients experienced less operative time and
required fewer surgeries, and therefore utilized fewer inpatient resources. Further
studies are needed to establish the long-term impact rLS may have on the pediatric
LS population and its' overall cost-utility relative to free flap reconstruction.