|Home | About | Journals | Submit | Contact Us | Français|
Posterior pelvic ring reconstruction can be challenging and controversial. The choice regarding whether to reconstruct and how to reconstitute the pelvic ring is unclear. Many methods provide stability but often are technically difficult and require excessive dissection.
This unique reconstructive technique uses the anterior aspect of the iliac crest with its attached muscle pedicle to provide a biologic scaffold for healing. The construct is secured with pedicle screws into the posterior column and S1 vertebral body with a spinal rod locked in compression. No additional fixation is used proximally into the lumbar spine. The iliac crest remains attached to the gluteus medius, allowing potential abductor function.
We retrospectively reviewed six patients who underwent iliosacral resection with this reconstruction. The mean age of the patients was 41 years. Complications were recorded. One patient died 6 months postoperatively. Musculoskeletal Tumor Society 1993 (MSTS ’93) score and Toronto Extremity Salvage Score (TESS) were obtained at a minimum 1-year followup in five patients. Healing was assessed radiographically. The minimum followup was 6 months (median, 33 months; range, 6–53 months).
The mean MSTS ’93 score was 72% and mean TESS was 66. All posterior column graft sites healed. At last followup, four of the five surviving patients had a stable pseudarthrosis at the proximal sacral site. One patient had a local recurrence and experienced failure of instrumentation without collapse or rotation of the hemipelvis 3 years postoperatively.
This technique provides a simple way to reconstruct the pelvic ring after iliosacral resection with clinical outcomes comparable to those for other methods. The method is a potential alternative for reconstruction of the posterior pelvic ring after resecting the ilium although reliable healing of the sacral site needs to be improved.
Primary malignancy of the bony pelvis presents a unique and challenging problem regarding resection and reconstructive options. The tumors often are large and close to vital neurovascular structures at the time of diagnosis. Despite the potential for local recurrence, numerous authors suggest limb salvage provides better quality of life compared with hemipelvectomy [10, 12, 13, 16, 19]. With the goal of limb salvage, resection still needs to be adequate with wide and clear margins while preserving vital neurovascular structures. The difficulty in reconstruction arises in providing a stable construct that would reconstitute the integrity of the pelvic ring and allow for use of the leg with minimal collapse or rotation of the residual portion of the hemipelvis and help support function and maintain quality of life.
Depending on the location of the tumor in the pelvis, Enneking and Dunham  described three resection zones. A Type I resection includes wide excision of a portion, or the entirety of, the ilium, and a Type I-S resection involves resection of a portion of the sacrum . Many authors [1, 2, 4, 11, 15] have suggested different strategies to deal with the discontinuity of the pelvis. Some have forgone any reconstruction of Type I resections [2, 8], leaving the pelvic ring in discontinuity. Not reconstructing the pelvic ring, however, resigns the patient to a protracted period of bed rest for the anterior pelvis to form scar and prevent collapse and rotation of the hemipelvis. If, however, the patient is allowed to bear weight, a progressive leg-length discrepancy occurs as the anterior pelvis collapses to articulate with the sacrum. Furthermore, there is no reconstruction of the abductor muscles to allow for better function of the hip. Some authors have described constructs using strut grafts [1, 2], vascularized iliac pedicles , or vascularized fibulas [4, 15, 20] to connect remaining periacetabular bone to the sacrum. Although these techniques address the problems produced by not reconstructing the pelvis they have tended to be technically challenging, can result in high complication rates, and may not provide reliable union [2, 15].
We developed a method of reconstruction after Type I-S iliosacral resection that uses a gluteus medius muscle pedicle iliac bone graft and minimal spinal fusion instrumentation to reconstruct the pelvic ring. This method provides a simpler construct for posterior pelvic reconstructions and allows for reconstitution of the pelvic ring using the host bone with its muscular attachment. We describe our surgical technique and subsequent function, graft healing, and complications in a small group of patients.
The indications for this technique were: (1) a malignant neoplasm of the posterior ilium not involving the anterosuperior iliac crest, inferior posterior column, or acetabulum. Sacral involvement must be none or minimal such that a safe resection can be made lateral to the sacral foramen, (2) a resection plan is needed that would disrupt continuity of the pelvic ring, and (3) there must be a desire for limb salvage. A specific contraindication, in addition to bone extent, was involvement of the neurovascular structures of the sciatic notch precluding safe resection, or major tumor involvement of the anterior gluteus medius muscle.
The patient is positioned in a semimobile lateral position so that the pelvis can be seen from a near supine position to a full lateral position. An extensile incision is made from the posterosuperior iliac spine extending anteriorly along the area of the iliac crest to the anterosuperior iliac spine, then directed distally down the Smith-Petersen interval, curving posteriorly, ending just distal to the ischial tuberosity. The anterior abdominal wall is detached from the iliac crest throughout its entire length using electrocautery and the retroperitoneal space is entered. The dissection then is developed between the iliacus and psoas. The femoral nerve is identified and protected, being retracted medially. The dissection then is carried down to the anterior portion of the sacrum. The L4-L5 conjoined nerve roots are identified and reflected in a medial direction and the ala of the sacrum is identified anteriorly, superiorly, and inferiorly toward the direction of the sciatic notch. The iliacus muscle then is identified at the level of the intended iliac osteotomy just at the line of the anteroinferior iliac spine directly toward the sciatic notch.
The patient is turned to a full lateral position. A large gluteal flap is made. If the gluteus maximus is involved, a wide margin is taken through the muscle belly; otherwise, the tumor mass is dissected from the muscle. The inferior portion of the dissection is developed between the gluteus medius and gluteus maximus muscle. The entire maximus then is reflected posteriorly. The sciatic nerve and all posterior pelvic contents are protected. The flap then is turned around posteriorly such that the entire flap could be pulled away from the tumor.
After developing the gluteal flap, attention is directed at harvesting the anterior iliac crest pedicle graft (Fig. 1). A portion of the anterior iliac crest is sized to fit into the defect left by the resection (approximately 1.5 cm deep and 6–8 cm long). The bone pedicle started from the anterosuperior iliac spine directly posteriorly. Using an osteotome, the bony pedicle is created while leaving the origins of the gluteus medius attached.
The remainder of the gluteus medius and minimus is transected along the line of intended resection from the sciatic notch up toward the anteroinferior iliac spine. When the inner and outer tables were thoroughly exposed, a sponge was passed beneath the sciatic notch to protect the contents of the notch. A Gigli saw then was passed through the notch and the osteotomy was performed just superior to the hip at the level of the anteroinferior iliac spine. Once full exposure had been gained and the nerve roots protected, a large osteotome was passed through the sacral ala directly posteriorly and distally (Fig. 2).
Two screws were anchored into the sacrum and posterior column as anchors for the reconstruction using a spinal instrumentation system. Typically, we recommend using a 6.5-mm titanium rod system or a 5.5-mm, 6.0-mm, or 6.5-mm diameter cobalt-chrome rod system to perform the reconstruction (Medtronic, Inc, Memphis, TN, USA). We do not recommend using a 5.5-mm diameter titanium rod system, as it may not be strong enough to support the reconstruction. The parasagittal osteotomy of the sacrum was exposed for insertion of the pedicle screw. We used a start point in the S1 body, approximately 2 cm distal to the superior end plate, and then followed a medial trajectory toward the antermedial aspect of the promontory of the sacrum using a pedicle finder. Either a 7.5-mm or 8.5-mm diameter titanium multiaxial pedicle screw was placed in the sacrum. The screw lengths ranged from 40 to 60 mm depending on the size of the patient. We then identified another start point medially along the cut portion of the ischium and used a pedicle finder to probe in a lateral-to-medial direction distally in the remainder of the posterior column. An image intensifier was used to ensure the trajectory did not violate the notch or acetabulum. We then implanted a 7.5-mm or 8.5-mm diameter titanium multiaxial pedicle or iliac titanium screw after tapping the hole created by the pedicle finder. The screw lengths ranged from 50 to 75 mm depending on the size of the posterior column remaining. Final placement of both screws was confirmed with biplanar fluoroscopy.
After confirmation of the screw position, we measured a 6.5-mm titanium or 5.5-, 6.0-, or 6.5-mm cobalt-chrome rod to fit the distance spanned between the two screws, then cut and contoured the rod, placed it into the screws, and secured it in place with the appropriate set screws using the torque and countertorque device. Once the rod was in position, the muscle pedicle iliac crest graft was fashioned to fit on the inside of the pelvis keyed into the ischium and sacrum. After the graft was placed, compression was applied across the construct, ensuring that the graft did not kick out. After compression was applied, we finalized tightening of the setscrews using the torque and countertorque device (Fig. 3).
The gluteus maximus muscle and the entire Smith-Petersen interval were repaired. We then repaired the abdominal wall back to the gluteus maximus fascia. The skin was closed in layers and a large bulky dressing was applied.
Patients were permitted toe-touch weightbearing for 6 weeks and then were advanced to 50% weightbearing for an additional 6 weeks. Radiographs were taken at 2, 6, and 12 weeks, 6 months, and then yearly (Fig. 4). If there was evidence of bridging trabecular bone or consolidation between the strut autograft and pelvis  at 12 weeks, the patient was allowed to advance weightbearing as tolerated. Patients with concerns for delayed union or pseudarthrosis were still advanced in weightbearing as long as they had no pain.
We retrospectively reviewed six patients with sarcoma of the posterior pelvis treated with this procedure from November 2006 to July 2010. During that time we treated a total of 19 patients with primary malignant tumors of the ilium. The mean age of the patients was 41 years (range, 17–73 years). Three patients were younger than 35 years. The cohort included two males and four females. Two patients had postradiation sarcoma, two had chondrosarcoma, one had fibrosarcoma, and one had Ewing’s sarcoma (Table 1). All but one patient had a minimum of 12 months of followup: one patient died 6 months after surgery as a result of her disease. The minimum followup for the five surviving patients was 12 months (median, 33 months; range, 13–53 months). Institutional review board approval was obtained to review patient records retrospectively.
Patients were followed with serial radiographs and physical examinations at 2, 6, and 12 weeks, 6 months, and then annually thereafter. The Toronto Extremity Salvage Score (TESS)  and the Musculoskeletal Tumor Society 1993 score (MSTS ’93)  were collected at latest followup. Four patients were seen in followup and one was contacted by phone. No MSTS ’93 score or TESS was available for the patient who died at 6 months.
Postoperative complications were recorded and graded based on the complication classification system described by Dindo et al.  and recently adapted for orthopaedic surgery by Sink et al. . Grade I complications had no clinical relevance on outcome; these included postoperative fever, nausea, constipation, and urinary tract infection. Grade II complications were events that resulted in deviation of postoperative care, pharmacologic treatment, or closer observation. These included superficial skin infection, transient neurapraxia, delayed union, or an asymptomatic nonunion. Grade III complications required surgical interventions. These included a superficial or deep wound infection that required surgical débridement, symptomatic failure of hardware that required revision surgery, or local recurrence. Grade IV complications result in permanent disability and include permanent nerve injury, major vascular injury, pulmonary embolism, or stroke. A Grade V complication is death. Radiographic evaluation of the construct was obtained at the latest office followup to assess union of the acetabular and sacral sites.
Procedures lasted for a median of 256 minutes (range, 212–339 minutes). The median reported blood loss was 1400 cc (range, 900–6000 cc). Patients received a median of 6.5 units of packed red blood cells (range, 4–16 units) during their hospitalization.
The average MSTS ’93 score was 72% (range, 33–93) and the average TESS was 66.2 (range, 53–88). On clinical examination, two patients (both younger than 35 years) had adequate abductor function with no Trendelenburg sign. The remaining patients had a positive Trendelenburg sign. One patient (Patient 5) continued to have substantial pain and limitations of function after surgery. The reason for the pain could not be elucidated despite multiple evaluations at our institution and at other quaternary referral centers. All six patients had radiographic union of the vascularized iliac crest with the posterior column, but only one achieved healing of the junction of the iliac crest autograft and sacrum (Fig. 5).
Several complications were noted in the postoperative evaluations of this cohort. Grade I complications included one patient with a postoperative ileus which resolved with observation. Grade II complications included two patients who had urinary retention that ultimately resolved but required prolonged catheterization in the meantime. Grade III complications developed in four patients. One patient experienced asymptomatic loosening of the sacral screw and rod fracture at 34 months postoperatively noted on his annual radiographs. The same patient had a local recurrence, which required reexcision (Fig. 6). At the time of re-excision, the instrumentation was removed because the patient had a stable pseudarthrosis intraoperatively. One patient experienced wound necrosis that was treated with repeat surgical débridement and a vacuum-assisted closure. Another patient had a radiation-induced injury to her incision 1 year after surgery and required débridement and revision of her scar. One patient was diagnosed with a deep wound infection 12 months after the index surgery. The patient had débridement with removal of all hardware and retention of the iliac graft. The posterior column site was well healed and the sacral site had a stable pseudarthrosis. There were no Grade IV or Grade V complications.
The first patient of the series was a 31-year-old woman who carried two pregnancies to term postoperatively without any medical complications. Both children were delivered by Cesarean section.
Reconstruction of the pelvic ring is a challenging task without a clear method to restore the strength or anatomy of the resected bone. Forgoing reconstruction of the pelvic ring, although reported, is not a demonstrably superior option to reconstructing the pelvic ring . Skeletally immature patients also may have hip dysplasia, leg length discrepancy and prgoressive hip pain may develop. Current reconstructive options may not be robust enough [2, 11] or tend to be complex, requiring excessive dissection to stabilize the construct to the lumbar spine . The purpose of our study was to investigate a novel reconstructive option, which uses autologous bone with its muscular pedicle to augment a simple pedicle screw construct. We describe our surgical technique and subsequent function and healing in a small initial cohort of patients.
We note limitations of our study. First, this is a small cohort of patients as a result of the rarity of pelvic tumors. This small cohort reflects the rarity of these reconstructions as one surgeon would see in his or her practice. A larger cohort and multicenter trial would help delineate the potential benefits of this reconstructive technique on a larger scale. Second, the complication rate in this series was high. The complication rate was similar to complication rates for pelvic surgery and reconstructions that have been reported in the literature, with some reporting the complication rates ranging from 40% to 75% [2, 14]. Third, we noted a high incidence of pseudarthrosis at the sacral site. All but one of the patients who had a pseudarthrosis develop received either postoperative chemotherapy or radiation. For four of the five patients, this was asymptomatic. Additionally, the pseudarthrosis may mimic the normal sacroiliac joint which has shown physiologic micromotion . We anticipate that if late hardware failure occurs, there should be minimal effect on the patient as the posterior pelvis had scarred in a stable position. The one patient with hardware failure in this series was asymptomatic and underwent reoperation only for local recurrence.
We have noted similar or improved function using this technique compared with other reported methods (Table 2). Beadel et al.  argued against reconstruction after iliosacral resection. At a mean followup of 45 months, they reviewed 16 patients who underwent Type I pelvic resections. Twelve patients had no reconstruction and four had reconstructions with contralateral iliac crest autograft or fibular allograft. The MSTS scores for the two groups were comparable (58% nonreconstruction versus 51% reconstruction), but their scores were lower than those of our patients. The mean TESS in their series was higher than the TESS for our autograft reconstructions (72 versus 66). According to Beadel et al. , patients who underwent reconstructions had a greater need for chronic pain medication and use of walking aids after surgery. They attributed the higher incidence of complications and difficulty with ambulation to the lack of abductor function. Sabourin et al.  followed 24 patients with tumors affecting the sacroiliac joint for a mean of 4.8 years. They used spinal instrumentation along with nonvascularized iliac autograft, tibial autograft, or allograft. At final followup, the average MSTS score was 61% in 12 patients but they noted a substantially lower score in patients who had a hemisacrectomy. If the patients who had a hemisacrectomy were excluded, the MSTS score increased to 77%, which was comparable to our findings. Akiyama et al.  reported their experience using nonvascularized fibula single- or double-strut autograft fixed with small fragment plates and screws in 10 patients. The mean MSTS score of their patients was 75% with no reported postsurgical complications and minimal pelvic shortening after the procedure.
Using this construct, our patients obtained excellent healing of the autograft at the posterior column but reliable healing of the graft to the lateral aspect of the sacrum was not achieved. At latest followup however, there was only one asymptomatic failure of fixation. For the reconstruction with allograft and screws used by Beadel et al. , only two of the four reconstructed pelves had long-lasting union of the bone graft. Similarly, Sabourin et al.  reported complete healing in 58% of their patients, whereas the remainder had a pseudarthrosis. They attributed nonhealing to adjuvant radiotherapy, chemotherapy, and use of allograft and metastatic recurrence. The double strut graft technique used by Akiyama et al.  did not fare any better, with union being achieved in at least one of the two strut grafts in all patients. Four of six patients who had a double-strut graft did not achieve healing of the upper graft.
Complications and reoperations are prevalent for pelvic tumor surgery. Using, the Dindo-Clavien system , Grades III, IV, and V complications can be compared between techniques since they result in increased morbidity and mortality. Four of the six patients in our cohort required an additional surgical intervention. Although a high rate, it is similar to those reported in the literature for pelvic reconstructions. Beadel et al.  reported 33% Grade III complications in the unreconstructed pelves and 75% in the reconstructed ones. Sabourin et al. reported a 75% rate of Grade III complications . In their cohort of 24 patients, 10 required débridement for scar necrosis and six for hematoma evacuation. Other authors who have used vascularized grafts have reported lower rates of complications. Sakuraba et al.  performed a double-barreled vascularized fibular graft on five patients with a mean followup of 12 months. There was only one Grade III complication, a reoperation for infection that required removal of the vascularized graft. Similarly Chang et al.  reported on six patients who underwent reconstruction with a vascularized double-barrel strut graft. They reported only one Grade II complication, a failure of fixation that required prolonged observation. Nishida et al.  reported on a small series of patients in whom they used vascularized iliac bone pedicle in Type I resections from the deep iliac artery with reconstruction with pedicle screw and rod construction. Only one patient had skin necrosis resulting from pedicle screw protrusion.
Our findings support the feasibility of using a gluteus medius muscle iliac crest pedicle graft in conjunction with minimal spinal instrumentation for reconstruction of the pelvic ring after Type I and I-S resections. Complications using this technique did not differ greatly from those of other published techniques. This technique has several advantages over others. Autogenous living bone is used and may contribute to better healing at the interfaces. A portion of the gluteus medius muscle remains attached to the iliac graft possibly providing abductor function. We also identified potential areas for improved stability of the construct to reduce nonunion rates at the sacral sites. Biomechanical studies can help elucidate weaknesses of the construct and potential modifications for improving its strength and stability.
One of the authors (JMB) certifies that he has or may receive payments or benefits, during the study period, an amount in excess of $10,000–$100,000, from Stryker, Inc (Mahwah, NJ, USA), CoreLink, Inc (St Louis, MO, USA), Globus Medical, Inc (Audubon, PA, USA), DePuy, Inc (Warsaw, IN, USA), K2M, Inc (Leesburg, VA, USA), One of the authors (DJM) certifies that he has or may receive payments or benefits, during the study period, an amount in excess of less than $10,000, from Stryker, Inc (Mahwah, NJ, USA) and Smith and Nephew, Inc (Memphis, TN, USA). The institution of one of the authors (JMB) has received funding from Complex Spine Study Group/K2M, Inc (Leesburg, VA, USA).
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research editors and board members are on file with the publication and can be viewed on request.
Clinical Orthopaedics and Related Research neither advocates nor endorses the use of any treatment, drug, or device. Readers are encouraged to always seek additional information, including FDA-approval status, of any drug or device prior to clinical use.
Each author certifies that his or her institution approved the human protocol for this investigation, that all investigations were conducted in conformity with ethical principles of research, and that informed consent for participation in the study was obtained.