Journal Logo

Reconstructive: Trunk: Original Articles

Free Fibula Flap for Restoration of Spinal Stability after Oncologic Vertebrectomy Is Predictive of Bony Union

Mericli, Alexander F. M.D.; Boukovalas, Stefanos M.D.; Rhines, Laurence D. M.D.; Adelman, David M. M.D., Ph.D.; Hanasono, Matthew M. M.D.; Chang, Edward I. M.D.

Author Information
Plastic and Reconstructive Surgery: January 2020 - Volume 145 - Issue 1 - p 219-229
doi: 10.1097/PRS.0000000000006382
  • Free
  • Journal Club
  • Podcast

Abstract

Primary vertebral bone tumors necessitate margin-free en bloc resection; this is considered best medical practice and is associated with lower rates of recurrence and superior overall outcomes compared with radiation therapy or other less aggressive surgical options.1 In general, an en bloc resection of a spinal malignancy requires a single or multilevel vertebrectomy, severely compromising the structural integrity of the vertebral column. The standard of care for restoring spinal stability includes posterior titanium instrumentation, anterior spinal column reconstruction using an interbody alloplastic cage, and nonvascular particulate bone graft.

Considering the large size of the resection, previous radiation therapy, scarring from prior operations, and the overall deconditioned state of the patient, these constructs face numerous impediments to graft union and normal wound healing. Prior studies have demonstrated up to a 50 percent nonunion rate for nonvascular spinal bone grafts greater than 4 cm.2–5 This is a serious complication, often resulting in spinal instability, instrumentation complications, graft resorption, deformity, and potential subsequent neurologic sequelae.4–7 Vascularized bone flaps may allow for more robust fusion and long-term stability in these challenging wounds.

In vascularized bone, the osteocytes remain viable, thus minimizing creeping substitution and maximizing primary bone healing. Bone flap consolidation is accelerated, resulting in increased union rapidity, thereby decreasing morbidity.5,8–14 Furthermore, vascularized bone flaps do not undergo extensive remodeling, and, as such, weakening is less likely to occur, in contrast to nonvascularized structural bone grafts.5,15,16 In animal studies, vascularized bone flaps demonstrate superior hypertrophy, improved vertebral union, and increased stiffness and strength when compared with nonvascular controls.17,18

Studies pertaining to the use of vascularized bone flaps for the restoration of spinal stability are limited, consisting mostly of case reports and small retrospective case series.8,9,13,19–29 It is not established whether vascularized bone flaps in spinal reconstruction are superior to the standard of care (nonvascularized bone graft and alloplastic instrumentation). We hypothesized that the addition of a free fibula flap for spinal stabilization after oncologic vertebrectomy for primary bone tumors is safe and effective, and will improve bone union and overall outcomes compared with the standard of care.

PATIENTS AND METHODS

We performed a retrospective cohort study including all patients who underwent vertebrectomy to treat a primary bone tumor at our institution from 2002 to 2017. We excluded patients with metastatic disease, those who underwent a partial vertebrectomy only, those with resections involving the sacrum or ilium, and those receiving other bone flaps. This study was approved by the University of Texas M. D. Anderson Cancer Center Institutional Review Board. Patients were placed into one of two groups: those who received vertebral reconstructions using alloplastic instrumentation and nonvascularized bone graft only (control group) and those whose spinal reconstruction consisted of alloplastic instrumentation, nonvascularized bone graft, and a free fibula flap (free fibula flap group). Plastic surgery was involved in the care of all patients in the study. Surgical complications recorded included wound infections requiring antibiotics, hematoma requiring evacuation, seroma requiring drainage, and wound dehiscence requiring reoperation or dressing changes for more than 6 weeks. Instrumentation complications included hardware exposure, breakage, or symptomatic loosening. Medical complications included pneumonia, venous thromboembolism, renal failure, and cardiovascular compromise. A neurologic complication was defined as any unanticipated motor or sensory change or cerebrovascular accident.

All operations were performed in two stages. The first stage is performed in the prone position and consists of resecting the posterior vertebral elements and placing posterior spinal instrumentation and nonvascular particulate bone graft. The second stage is performed in the lateral decubitus position 2 to 4 days later; the anterior components of the vertebral body or bodies are resected, thereby allowing for an en bloc extirpation. The anterior spine is approached through a thoracotomy, lumbotomy, or anterior cervical incision, depending on the location of the tumor. During the second stage, the fibula flap elevation coincides with the anterior spinal approach. The fibula is inset and revascularized after the spine surgeons complete the tumor resection and place the interbody cage. The microvascular anastomosis between the peroneal vessels and recipient vessels is not monitored postoperatively.

Bony fusion was assessed radiologically at 3-month intervals according to Ray’s criteria,30 which specifically define union as the presence of visible bone formation in or about the fibula flap or nonvascularized graft and (1) less than 3 degrees of intersegmental position change on lateral flexion and extension views; (2) no lucent areas around hardware; (3) minimal loss of disk height; (4) no fracture of the hardware, graft, or vertebra; and/or (5) no sclerotic changes in the graft or adjacent vertebra. The primary outcome was radiographic evidence of bony union; secondary outcomes included instrumentation complications and surgical complications.

Statistical Analysis

Statistical analyses were performed using IBM SPSS (IBM Corp., Armonk, N.Y.). Either Fisher’s exact test or the chi-square test was used to compare categorical variables. The two-sided paired t test or the Mann-Whitney U test was used to compare continuous variables. Variables with statistically significant associations (p < 0.05) on univariate analysis were controlled for in a multivariate logistic regression to identify independent factors predictive of nonunion and instrumentation complications. Kaplan-Meier curves were created using the product limit method to estimate the probability of bony union and instrumentation longevity.

RESULTS

Between 2002 and 2017, 40 patients underwent resection of a vertebral primary bone tumor. In 24 patients, the spinal reconstruction consisted of titanium instrumentation posteriorly and an interbody spacer anteriorly and nonvascularized bone graft only (control group); in 16 patients, this spinal reconstruction was supplemented with a free fibula flap (free fibula flap group). The inter–vertebral body spacer consisted of either a titanium cage (22 patients), polyetheretherketone cage (14 patients), or cortical bone allograft (four patients). The median age was 36.1 years (interquartile range, 34.3 to 54.3 years) in the free fibula flap group and 46.6 years (interquartile range, 32.2 to 53.4 years) in the control group (p = 0.95). The incidences of tobacco use, preoperative and postoperative radiation therapy, neoadjuvant chemotherapy, and medical comorbidities were similar between the two groups (Table 1). The survival rates were also similar: 87.5 percent of the free fibula flap group and 91.5 percent of the control group were alive at last follow-up (Table 1). The median follow-up was 33.7 months (interquartile range, 11.9 to 43.3 months) in the free fibula flap group and 85.7 months (interquartile range, 55.7 to 115.9 months) in the control group (p = 0.001).

Table 1. - Patient Characteristics and Demographics
Free Fibula Flap (%) Control (%) p
No. 16 24
Age, yr 0.95
 Median 36.1 46.6
 IQR 34.3–54.3 32.2–53.4
Male 7 (43.8) 16 (66.7) 0.19
BMI, kg/m2 0.92
 Median 26.2 27.0
 IQR 24–29.3 23–31.8
Obese (BMI >30 kg/m2) 4 (25) 8 (33.3) 0.73
Tobacco use 2 (12.5) 8 (33.3) 0.26
>1 comorbidity 3 (18.8) 11 (45.8) 0.08
Preoperative RT* 4 (25) 2 (8.3) 0.19
Postoperative RT* 2 (12.5) 6 (25) 0.33
Chemotherapy 4 (25) 5 (20.8) 1
Survival* 14 (87.5) 22 (91.5) 0.82
Follow-up, mo
 Median 33.7 85.7
 IQR 11.9–43.3 55.7–115.9 0.001
I
QR, interquartile range; BMI, body mass index; RT, radiation therapy.
*
Number and percentage of patients alive at last follow-up.

The most common tumor abnormality was chordoma, accounting for 50 percent of the resections in the free fibula flap group and 54.2 percent of those in the control group (Fig. 1). The mean osseous defect size was 6.5 cm (range, 4.2 to 9 cm). There were more cervical vertebrectomies in the control group (6.3 percent versus 37.5 percent; p = 0.03), whereas there were more thoracic vertebrectomies in the free fibula flap group (37.5 percent versus 8.3 percent; p = 0.04) (Table 2). There was a trend toward more multilevel vertebrectomies in the fibula group, with a median of two vertebrae removed in the free fibula flap group versus one vertebra in the control group (p = 0.08) (Table 2).

Table 2. - Recipient Vessels for Free Fibula Flap
Artery/Vein No. of Patients
Lumbar segmental 5
Thoracic segmental 2
Splenic 1
Gastroepiploic 1
Aorta/hemiazygos 1
Common iliac 1
Deep inferior epigastric 1
Internal mammary 1
Subscapular 1
Subclavian 1
Facial 1
Total 16

Fig. 1.
Fig. 1.:
Pie chart illustrating the distribution of tumor abnormality for patients in the fibula flap group (left) and control group (right).

Fourteen fibulae were placed in the interbody position; two cervical-level fibulae were positioned posteriorly, as the cross-sectional area of the fibulae was greater than that of the flanking vertebral bodies. All fibulae were designed as single struts, to provide room for the interbody cage, and were fixated to the adjacent superior and inferior vertebral bodies with miniplate titanium hardware. A variety of arterial and venous recipient vessels were used, most commonly lumbar or thoracic segmental vessels (Table 2). Twenty-five percent of arterial anastomoses were coupled, in contrast to 68.7 percent of venous anastomoses. Vein grafts were used in 18.7 percent and 25 percent of arterial and venous anastomoses, respectively. There were no intraoperative thromboses, and the flaps were not monitored postoperatively.

There were significantly more nonunions (41.7 percent versus 6.3 percent; p = 0.02), instrumentation complications (33.3 percent versus 6.3 percent; p = 0.04), and neurologic complications (25 percent versus 0 percent; p = 0.03) in the control group (Fig. 2). There was no difference in surgical complications, medical complications, reoperation for complication, or overall complications between the two groups (Fig. 2).

Fig. 2.
Fig. 2.:
Complications after vertebrectomy reconstruction using the free fibula flap versus control. There was a statistically significant increased rate of nonunions, instrumentation complications, and neurologic complication in the cage-only group (*p < 0.05).

Univariate analysis revealed that nonunion was nearly 11 times greater in the control group (OR, 10.7; 95 percent CI, 1.2 to 94.9; p = 0.03) (Table 3). Other variables significantly associated with nonunion included the existence of an instrumentation complication (OR, 23.6; 95 percent CI, 3.5 to 156.4; p = 0.001), tobacco use (OR, 15.2; 95 percent CI, 2.7 to 84.2; p = 0.002), any complication (OR, 16.3; 95 percent CI, 1.8 to 145.9; p = 0.004), a surgical complication (OR, 7; 95 percent CI, 1.5 to 33.2; p = 0.01), and possessing more than one comorbidity (OR, 5.5; 95 percent CI, 1.2 to 24.5; p = 0.03). In our multivariate logistic regression analysis, we controlled for all variables significantly associated with nonunion on univariate analysis, and for length of follow-up, and found that reconstruction without a free fibula flap (OR, 57; 95 percent CI, 1.17 to 2773; p = 0.04) was the only variable independently predictive of nonunion.

Table 3. - Univariate Analysis for Nonunion
Nonunion (%) OR (95% CI) p
Control group (no FFF) 10 (41.7) 10.7 (1.2–94.9) 0.03*
Obese (BMI >30 kg/m2) 5 (41.7) 2.6 (0.61–11.3) 0.25
>1 comorbidity 7 (63.6) 5.5 (1.2–24.5) 0.03*
Tobacco use 7 (70) 15.2 (2.7–84.2) 0.002*
Preoperative radiation therapy 2 (33.3) 1.4 (0.22–8.9) 1
Postoperative radiation therapy 2 (25) 0.85 (0.14–5.0) 0.86
Chemotherapy 2 (22.2) 0.69 (0.12–4.03) 1
Multilevel vertebrectomy 6 (37.5) 2.3 (0.55–9.4) 0.29
Cervical vertebrectomy 4 (40) 2.2 (0.48–10) 0.42
Thoracic vertebrectomy 3 (37.5) 1.8 (0.35–9.3) 0.66
Lumbar vertebrectomy 4 (18.2) 0.35 (0.08–1.5) 0.17
Instrumentation complication 7 (63.6) 23.6 (3.5–156.4) 0.001*
Medical complication 2 (25) 0.85 (0.14–5.0) 1
Neurologic complication 3 (50) 3.3 (0.54–19.4) 0.32
Surgical/wound complication 8 (50) 7 (1.5–33.2) 0.01*
Any complication 10 (47.6) 16.3 (1.8–145.9) 0.004*
F
FF, free fibula flap; BMI, body mass index.
*
Statistically significant.

Evidence of bony union was seen at a mean of 4.8 months after surgery in the free fibula flap group and at 22.4 months in the control group (p = 0.001). On Kaplan-Meier analysis, the probability of union at 30 months after surgery was nearly 94 percent in the free fibula flap group, in contrast to 50 percent in the control group (p < 0.001) (Fig. 3).

Fig. 3.
Fig. 3.:
Kaplan-Meier curve demonstrating the probability of achieving bony union in the fibula flap group versus the control group.

Univariate analysis revealed nonunion (OR, 23.6; 95 percent CI, 3.5 to 156.4; p = 0.001), tobacco use (OR, 6.5; 95 percent CI, 1.3 to 33.1; p = 0.03), and surgical complication (OR, 2.3; 95 percent CI, 1.3 to 3.98; p = 0.001) to be significantly associated with an instrumentation complication (Table 4). The control group trended toward an increased risk of instrumentation failure, but did not reach statistical significance (OR, 7.5; 95 percent CI, 0.84 to 67.3; p = 0.06) (Table 4). Multivariate logistic regression analysis did not identify any variable independently predictive of an instrumentation complication. Among the seven patients with instrumentation complications in the control group, the mean time to complication was 22.5 months (range, 1 to 60 months); only one patient in the free fibula flap group experienced an instrumentation complication, consisting of hardware exposure necessitating reoperation, at 6 months after surgery. On Kaplan-Meier analysis, the probability of instrumentation complication at 60 months after surgery was 5 percent in the free fibula flap group, in contrast to 41.6 percent in the control group; however, this difference was not statistically significant (p = 0.27) (Fig. 4).

Table 4. - Univariate Analysis for Instrumentation Complication
Instrumentation Complication (%) OR (95% CI) p
Control group (no FFF) 8 (33.3) 7.5 (0.84–67.3) 0.06
Obese (BMI >30 kg/m2) 3 (25) 1.2 (0.25–5.9) 1
>1 comorbidity 4 (28.6) 1.7 (0.37–7.6) 0.69
Tobacco use 5 (50) 6.5 (1.3–33.1) 0.03*
Preoperative radiation therapy 2 (33.3) 1.9 (0.29–12.8) 0.6
Postoperative radiation therapy 0 (0) 0.74 (0.6–0.91) 0.08
Chemotherapy 2 (22.2) 0.98 (0.16–5.8) 1
Multilevel vertebrectomy 4 (25) 1.3 (0.28–5.6) 1
Cervical vertebrectomy 4 (40) 3.3 (0.68–16.3) 0.19
Thoracic vertebrectomy 2 (25) 1.2 (0.19–7.5) 1
Lumbar vertebrectomy 3 (13.6) 0.32 (0.06–1.5) 0.25
Nonunion 7 (63.6) 23.6 (3.5–156.4) 0.001*
Medical complication 1 (12.5) 0.43 (0.04–4.0) 0.65
Neurologic complication 3 (50) 4.67 (0.75–29.1) 0.11
Surgical/wound complication 9 (56.3) 2.29 (1.31–3.98) 0.001*
Any complication 9 (42.9) 1.75 (1.21–2.53) 0.001*
F
FF, free fibula flap; BMI, body mass index.
*
Statistically significant.

Fig. 4.
Fig. 4.:
Kaplan-Meier curve demonstrating the probability of not experiencing an instrumentation complication in the fibula flap group versus the control group.

DISCUSSION

This study represents the largest series of free fibula flaps for oncologic spinal reconstruction, and is the only such study providing level III data. Through our analysis, we have demonstrated that the addition of a free fibula flap to the spinal construct is independently predictive of bony union and increased rapidity of union.

Patients who require a vascularized bone flap to supplement their spinal reconstruction are rare, with such cases representing only a small percentage of all spine operations. In addition to certain oncologic resections, other indications for a bone flap may include vertebral osteomyelitis, trauma, and progressive symptomatic spinal deformities.19–22,31 At our institution, we consider patients with primary vertebral bone tumors requiring en bloc resection for curative intent to be optimal candidates for a bone flap. These patients are often young, as reflected in the median age of the patients in our study (Table 1), and have nearly normal survival, which places them at risk for a late instrumentation complication, should a strong, durable, biological union not be achieved.1,32–34 In contrast, although patients with distant spinal metastases may receive a similar resection for palliation, our practice is to perform these vertebral reconstructions without vascularized bone flap supplementation, considering the comorbidities and limited survival of patients with metastases.

Rose and colleagues reported the first bone flap used in the spine, describing a pedicled rib flap in three spina bifida cystica patients.35 Since then, various flaps have been described, including the pedicled iliac crest, free iliac crest, free rib, and free fibula flaps.23–25,36–38 The free fibula flap has emerged as the superior and preferred spinal bone flap for a variety of reasons. The utility of pedicled rib flaps is limited to the thoracic spine; furthermore, the intercostal vessels are often ligated during the vertebrectomy as they emerge from the aorta and inferior vena cava, eliminating the possibility of a vascularized bone flap.9 In addition, the rib’s thin, brittle structure and curved shape are not as mechanically favorable as the structure and shape of the fibula.23,26 In comparison with the iliac crest flap, the fibula is significantly stronger, the dissection is easier and better-suited to a two-team approach, and there is less donor-site morbidity.21,27,39,40

Previous studies indicate that oncologic spinal osseous defects greater than 4 cm reconstructed with nonvascular bone graft and instrumentation only have fusion failure rates of 16 to 51 percent and a similarly high complication rate.1–5,32–34 Indeed, in our study, the mean osseous defect was 6.5 cm, and the control group nonunion rate was 41.7 percent. In contrast, only one free fibula flap patient, who was previously irradiated and was a tobacco user, developed a pseudoarthrosis, resulting in a 6.3 percent nonunion rate in the free fibula flap group, illustrating the superior biological properties conferred by a vascularized bone flap. This finding compares favorably with those of other publications on the spinal free fibula flap, which report a nonunion rate ranging from 10 to 25 percent.8,9,27,34 It is important to note that, in general, oncologic spinal resections are more morbid and are associated with a greater incidence of nonunion, malunion, and hardware complications compared to the surgical treatment of degenerative spinal disease. Indeed, union rates for lumbar, thoracic, and cervical level degenerative disease are comparatively far superior, ranging from 89.5 to 99.2 percent, depending on technique and study.41,42

One potential adverse outcome of spinal nonunion is instrumentation complication. In patients whose reconstruction unites, biomechanical stress and strain initially imparted to the instrumentation are eventually assumed by the bony construct, whereas in those with nonunion or a weaker union, the biomechanical forces fatigue the instrumentation over time, ultimately resulting in hardware failure. In general, this is a time-dependent phenomenon, which is why we consider young patients with a normal life expectancy, such as those with primary bone tumors, to be good candidates for free fibula flap reconstruction. On the basis of the improvements we documented in achieving bony union and the faster time to union, we believe our current approach represents a paradigm shift in the reconstruction of oncologic spine defects, and we advocate the use of a free fibula flap to supplement the construct. We identified a significantly greater instrumentation complication rate in the control group than in the free fibula flap group (33.3 percent versus 6.3 percent; p = 0.02). Furthermore, univariate analysis revealed a nonsignificant trend toward a reconstruction without a free fibula flap being associated with eventual instrumentation complication (OR, 7.5; 95 percent CI, 0.84 to 67.3; p = 0.06) (Table 4). This finding is similar to those of other studies, which report an instrumentation complication rate of 25 to 39.1 percent after oncologic vertebrectomy and reconstruction without vascularized bone flaps.28,43,44

Identifying adequate recipient blood vessels is a challenging aspect of using a free vascularized bone flap for spinal reconstruction. Whereas the cervical spine offers a variety of recipient blood vessel options, the thoracic and lumbar posterior trunk is notorious for its vessel paucity. Multiple segmental vessels emerge from the thoracic and abdominal aorta and inferior vena cava and wrap around the vertebra to eventually become the lumbar and intercostal vessels. These segmental arteries and veins should be considered the first-line option for anastomosis, given their proximity to the spine and favorable size match to the peroneal vessels. Unfortunately, these vessels are also routinely ligated or cauterized during the vertebrectomy, so communication with the spine surgeon and/or thoracic surgeon is of utmost importance to preserve the vessels and ensure adequate length. Secondary options, depending on location, include end-to-side to the aorta, the azygous/hemiazygous vein, internal mammary vessels, or common iliac vessels, and/or the use of vein grafts.13 The gastroepiploic blood vessels offer another option; the omentum can be delivered into the intrathoracic space through a diaphragmatic fenestration, or into the retroperitoneum along the lumbar spine, placing the gastroepiploic vessels near the spine. Other recipient vessel options include the renals, superior gluteals, and inferior mesenterics13 and, for venous anastomosis, the thoracic duct.29 Regardless of the recipient vessel choice, the microvascular anastomosis is technically challenging, as it is often performed within a narrow and deep thoracotomy or lumbotomy incision. An operating microscope with a long functional focal length is important, as are specialized, elongated microsurgical instruments.

In our series, all fibula flaps were inset in collaboration with our spine surgery colleagues, spanning the distance between the intact caudal and cranial vertebral bodies. A variety of inset techniques are documented in the literature including double-strut flaps, triple-strut flaps, and placing the fibula within the interbody cage; however, we prefer a single fibula strut combined with an adjacent alloplastic interbody cage (Figs. 5 through 7).8,9,19,27,45 Proponents of multistrut flap design cite the increased surface area for fusion as an advantage. However, like other authors, we have found cutting and positioning the struts to ensure uniform compression to be very challenging.34 Instead we advocate placing an expandable alloplastic interbody cage first, to precisely set the intervertebral distance, spinal curvature, and compression. Through the thoracotomy or lumbotomy, the fibula flap is then carefully press-fit alongside the cage and secured to the adjacent vertebral bodies with titanium miniplates and screws. We seek to minimize fibula flap hardware to avoid decreased treatment accuracy, should postoperative proton beam or electron beam radiation therapy be necessary.46

Fig. 5.
Fig. 5.:
Postoperative computed tomographic images demonstrating placement of the free fibula flap in parallel with the interbody cage. The posterior spinal instrumentation is highlighted in blue. (Left) T5 to T7 vertebrectomy, 12 months after surgery. (Center) T8 to T11 vertebrectomy, 6 months after surgery. (Right) L2 vertebrectomy 18 months after surgery.
Fig. 6.
Fig. 6.:
C4 to C5 vertebrectomy, 12 months after surgery. The posterior onlay fibula flap spans C2 to C7; the anterior interbody cage is between C2 and C6.
Fig. 7.
Fig. 7.:
(Above) Intraoperative photograph after T8 to T9 vertebrectomy, posterior instrumentation, and placement of an expandable interbody polyetheretherketone cage (visualized at the base of the defect). The patient is positioned such that the top of the photograph is anterior and the bottom is posterior. The aorta is visualized anterior to the cage and the spinal cord is posterior. (Below) Intraoperative photograph after placement of free fibula flap in parallel with the interbody cage. The fibula is fixated to the T7 vertebral body cranially and the T10 vertebral body caudally. The pedicle is anastomosed to thoracic segmental vessels from the aorta and inferior vena cava.

Criticisms of the free fibula flap for spinal reconstruction include the added surgical morbidity and additional operative time. On the contrary, we did not identify any statistically significant difference in the incidence of wound-healing complications or surgical complications between the free fibula flap and control groups, despite the additional lower extremity surgical site in the free fibula flap cohort. Regarding timing, a bone-only fibula flap can be harvested in 60 to 90 minutes, simultaneously with the anterior approach to the spine. Therefore, the only added time should be that required for placement and fixation of the flap and the microvascular anastomosis, which, in experienced hands, should take no longer than 60 minutes. Similar to other authors, we do not monitor our spinal free fibula flaps after revascularization. Given the multiple vital structures within the immediate surgical field (lung, heart, aorta, inferior vena cava, spinal cord), and the fact that the patient has endured a 2-day, two-stage physiologically demanding procedure, we believe an emergent flap take-back is contraindicated; therefore, postoperatively, we do not monitor the patency of our microvascular anastomosis. Because of this, we cannot be certain there were no flap losses; however, postoperative radiology did not reveal flap resorption or abscess formation—imaging findings potentially indicative of flap nonviability.

A strength of our study is the homogeneity of our patient population and anatomical location, considering that inclusion was limited to oncologic defects of the mobile spine and that sacropelvic reconstructions were excluded. In addition, ours is the only study to date that includes a comparison analysis, thereby providing level III data. As such, ours is the only study on this topic using a rigorous statistical analysis and, specifically, a multivariable logistic regression. Our project is limited by its retrospective design and associated biases. In addition, because this application of the fibula flap is a relatively recent surgical innovation, our follow-up times were dissimilar between the two groups. However, we controlled for the discrepancy in our regression analysis and still identified that reconstruction without a free fibula flap is independently predictive of nonunion.

CONCLUSIONS

The free fibula flap for spinal reconstruction after oncologic vertebrectomy offers significant advantages for patients with primary vertebral bone tumors. Supplementing the spinal reconstruction with a free fibula flap is associated with fewer nonunions, instrumentations complications, and neurologic complications compared with spinal reconstructions using standard techniques (alloplastic instrumentation and nonvascularized bone graft only). Most importantly, a spinal reconstruction without a free fibula flap is independently predictive of nonunion and increased time to union, compared with a reconstruction using a free fibula flap. The free fibula flap for spinal reconstruction represents yet another opportunity for plastic surgeons to collaborate with other specialists to optimize patient outcomes.

ACKNOWLEDGMENT

The authors would like to thank Jessie Liu, Ph.D., for assistance with creating the Kaplan-Meier curves.

REFERENCES

1. Lewandrowski KU, Hecht AC, DeLaney TF, Chapman PA, Hornicek FJ, Pedlow FX. Anterior spinal arthrodesis with structural cortical allografts and instrumentation for spine tumor surgery. Spine (Phila Pa 1976) 2004;29:1150–1158; discussion 1159.
2. Bradford DS, Daher YH. Vascularised rib grafts for stabilisation of kyphosis. J Bone Joint Surg Br. 1986;68:357–361.
3. Saraph VJ, Bach CM, Krismer M, Wimmer C. Evaluation of spinal fusion using autologous anterior strut grafts and posterior instrumentation for thoracic/thoracolumbar kyphosis. Spine (Phila Pa 1976) 2005;30:1594–1601.
4. Bradford DS, Ganjavian S, Antonious D, Winter RB, Lonstein JE, Moe JH. Anterior strut-grafting for the treatment of kyphosis: Review of experience with forty-eight patients. J Bone Joint Surg Am. 1982;64:680–690.
5. Lonstein JE, Winter RB. Long multiple struts for severe kyphosis. Clin Orthop Relat Res. 2002;394:130–138.
6. Jenkins DH, Hodgson AR, Yau AC, et al. Stablization of the spine in the surgical treatment of severe spinal tuberculosis in children. Clin Orthop Relat Res. 1975;110:69–80.
7. Rajasekaran S, Soundarapandian S. Progression of kyphosis in tuberculosis of the spine treated by anterior arthrodesis. J Bone Joint Surg Am. 1989;71:1314–1323.
8. Houdek MT, Rose PS, Bakri K, et al. Outcomes and complications of reconstruction with use of free vascularized fibular graft for spinal and pelvic defects following resection of a malignant tumor. J Bone Joint Surg Am. 2017;99:e69.
9. Moran SL, Bakri K, Mardini S, Shin AY, Bishop AT. The use of vascularized fibular grafts for the reconstruction of spinal and sacral defects. Microsurgery 2009;29:393–400.
10. Noorda RJ, Wuisman PI, Fidler MW, Lips PT, Winters HA. Severe progressive osteoporotic spine deformity with cardiopulmonary impairment in a young patient: A case report. Spine (Phila Pa 1976) 1999;24:489–492.
11. Han CS, Wood MB, Bishop AT, Cooney WP III.. Vascularized bone transfer. J Bone Joint Surg Am. 1992;74:1441–1449.
12. Weiland AJ, Moore JR, Daniel RK. Vascularized bone autografts: Experience with 41 cases. Clin Orthop Relat Res. 1983;174:87–95.
13. Winters HA, Kraak J, Oosterhuis JW, de Kleuver M. Spinal reconstruction with free vascularised bone grafts; approaches and selection of acceptor vessels. Scand J Surg. 2013;102:42–48.
14. Shin AY, Bishop AT. Vascularized bone grafts for scaphoid nonunions and Kienböck’s disease. Orthop Clin North Am. 2001;32:263–277, viii.
15. Arata MA, Wood MB, Cooney WP III.. Revascularized segmental diaphyseal bone transfers in the canine: An analysis of viability. J Reconstr Microsurg. 1984;1:11–19.
16. Cutting CB, McCarthy JG. Comparison of residual osseous mass between vascularized and nonvascularized onlay bone transfers. Plast Reconstr Surg. 1983;72:672–675.
17. Shaffer JW, Field GA. Rib transposition vascularized bone grafts: Hemodynamic assessment of donor rib graft and recipient vertebral body. Spine (Phila Pa 1976) 1984;9:448–449.
18. Shaffer JW, Field GA, Goldberg VM, Davy DT. Fate of vascularized and nonvascularized autografts. Clin Orthop Relat Res. 1985;197:32–43.
19. Aliano KA, Agulnick M, Cohen B, et al. Spinal reconstruction for osteomyelitis with free vascularized fibular grafts using intra-abdominal recipient vessels: A series of three cases. Microsurgery 2013;33:560–566.
20. Lee MJ, Ondra SL, Mindea SA, Fine NA, Dumanian GA. Indications and rationale for use of vascularized fibula bone flaps in cervical spine arthrodeses. Plast Reconstr Surg. 2005;116:1–7.
21. Wuisman PI, Jiya TU, Van Dijk M, Sugihara S, Van Royen BJ, Winters HA. Free vascularized bone graft in spinal surgery: Indications and outcome in eight cases. Eur Spine J. 1999;8:296–303.
22. Erdmann D, Meade RA, Lins RE, McCann RL, Richardson WJ, Levin LS. Use of the microvascular free fibula transfer as a salvage reconstruction for failed anterior spine surgery due to chronic osteomyelitis. Plast Reconstr Surg. 2006;117:2438–2445; discussion 2446–2447.
23. Kaneda K, Kurakami C, Minami A. Free vascularized fibular strut graft in the treatment of kyphosis. Spine (Phila Pa 1976) 1988;13:1273–1277.
24. Doi K, Kawai S, Sumiura S, Sakai K. Anterior cervical fusion using the free vascularized fibular graft. Spine (Phila Pa 1976) 1988;13:1239–1244.
25. Sumiura S. Free vascularized fibula graft applied to anterior body fusion of the cervical spine: Clinical and experimental study (in Japanese). Nihon Seikeigeka Gakkai Zasshi 1989;63:308–319.
26. Hubbard LF, Herndon JH, Buonanno AR. Free vascularized fibula transfer for stabilization of the thoracolumbar spine: A case report. Spine (Phila Pa 1976) 1985;10:891–893.
27. Winters HA, van Engeland AE, Jiya TU, van Royen BJ. The use of free vascularised bone grafts in spinal reconstruction. J Plast Reconstr Aesthet Surg. 2010;63:516–523.
28. Sciubba DM, De la Garza Ramos R, Goodwin CR, et al. Total en bloc spondylectomy for locally aggressive and primary malignant tumors of the lumbar spine. Eur Spine J. 2016;25:4080–4087.
29. Winters HA. The thoracic duct as a venous receptor vessel for free flap transplantation. Plast Reconstr Surg. 1998;101:872–873.
30. Ray CD. Threaded fusion cages for lumbar interbody fusions: An economic comparison with 360 degrees fusions. Spine (Phila Pa 1976) 1997;22:681–685.
31. Shin DE, Kim HS, Ahn CS, Lee DH, Lee SC. Anterior debridement and strut graft with pedicle screw fixation for pyogenic spondylitis. Asian Spine J. 2007;1:91–97.
32. Fernyhough JC, White JI, LaRocca H. Fusion rates in multilevel cervical spondylosis comparing allograft fibula with autograft fibula in 126 patients. Spine (Phila Pa 1976) 1991;16(Suppl):S561–S564.
33. Emery SE, Hughes SS, Junglas WA, et al. The fate of anterior vertebral bone grafts in patients irradiated for neoplasm. Clin Orthop Relat Res. 1994;300:207–212.
34. Yanamadala V, Rozman PA, Kumar JI, et al. Vascularized fibular strut autografts in spinal reconstruction after resection of vertebral chordoma or chondrosarcoma: A retrospective series. Neurosurgery 2017;81:156–164.
35. Rose GK, Owen R, Sanderson JM. Transposition of rib with blood supply for the stabilization of a spinal kyphosis. J Bone Joint Surg Br. 1975;57:112117.
36. Oshima Y, Miyoshi K, Mikami Y, Kawamura N. Pelvic ring reconstruction with a vascularized pedicle iliac bone graft for a large sacral schwannoma. J Spinal Disord Tech. 2005;18:200–202.
37. Métaizeau JP, Czorny A, Miahle C, Kuhnast M, Prévot J. Use of vascularized bone grafts in surgery of the spine: Apropos of 6 cases (in French). Rev Chir Orthop Reparatrice Appar Mot. 1989;75:166–171.
38. Aydinli U, Akin S, Tirelioglu O, Ozturk C, Yalcinkaya U, Sadikoglu Y. A new autogenous graft choice in pelvic reconstructions: Free vascularized rib (a case report). Arch Orthop Trauma Surg. 2006;126:57–62.
39. Winters HA, Van Loenen DK. A comparison between fibula and iliac crest in mandibular reconstruction. Eur J Plast Surg. 2007;29:205–208.
40. Wittenberg RH, Moeller J, Shea M, White AA III, Hayes WC. Compressive strength of autologous and allogenous bone grafts for thoracolumbar and cervical spine fusion. Spine (Phila Pa 1976) 1990;15:1073–1078.
41. Schroeder GD, Kepler CK, Millhouse PW, et al. L5/S1 fusion rates in degenerative spine surgery: A systematic review comparing ALIF, TLIF, and axial interbody arthrodesis. Clin Spine Surg. 2016;29:150–155.
42. Rao RD, Gore DR, Tang SJ, Rebholz BJ, Yoganandan N, Wang M. Radiographic changes in the cervical spine following anterior arthrodesis: A long-term analysis of 166 patients. J Bone Joint Surg Am. 2016;98:1606–1613.
43. Shah AA, Paulino Pereira NR, Pedlow FX, et al. Modified en bloc spondylectomy for tumors of the thoracic and lumbar spine: Surgical technique and outcomes. J Bone Joint Surg Am. 2017;99:1476–1484.
44. Kavadi N, Tallarico RA, Lavelle WF. Analysis of instrumentation failures after three column osteotomies of the spine. Scoliosis Spinal Disord. 2017;12:19.
45. Ackerman DB, Rose PS, Moran SL, Dekutoski MB, Bishop AT, Shin AY. The results of vascularized-free fibular grafts in complex spinal reconstruction. J Spinal Disord Tech. 2011;24:170–176.
46. Vrints I, Depreitere B, Vranckx JJ. Multilevel cervical reconstruction with no remaining hardware: The potential of a vascularised fibular strut graft. J Plast Reconstr Aesthet Surg. 2012;65:e344–e347.
Copyright © 2019 by the American Society of Plastic Surgeons