Advances in both fusion techniques and instrumentation have markedly facilitated the treatment of spinal disorders. Yet a significant number of patients exist who continue to have pseudarthroses. Despite the surgical advances, the essentials of a successful spinal fusion still appear to be the effective application of sound bone-grafting principles. These principles, along with the techniques, problems, and complications associated with bone grafting, are reviewed in this chapter.

The loss of bone in the spine often presents serious difficulties not seen in other areas. The most favorable replacement would still be a bone graft that fills the defect and becomes incorporated into the spine. However, the availability of appropriate bone to replace the loss is a significant problem. Also surveyed in this chapter are alternatives to autogenous bone grafts, or autografts, including allografts, xenografts, and synthetic bone substitutes as well as other implants used to stabilize the spine.

Bone grafts have often played the roles of scaffolds, bridges, spacers, defect fillers, and bone-loss replacements. Immobilization of multiple motion segments is frequently necessary in the spine; great demands are made on bone grafts. In the lumbosacral spine, body weight and muscular forces impart loads equal to three or four times body weight.¹⁴⁹ It is not surprising that the highest rate of bone-graft failure is seen in the lumbosacral spine. Hence, the following is a discussion of technical problems and the biomechanical and physiologic characteristics of bone grafts, bone substitutes, and implants. Since our last edition on this subject in 1987, ⁶² significant progress has been made, and this updated chapter reflects the additional knowledge we have gained. At this time, the exact mechanisms controlling the physiologic processes of bone-graft incorporation and remodeling are still not completely understood. Yet, we may be able to improve our clinical results in spine surgery by appreciating the principles and information already known regarding this subject.

The Autograft
Autograft, or bone graft transplanted from one site to another in the same individual, is considered to be the most biologically suitable type of graft. Its advantages include

1. Superior osteogenic capacity
   • Contributes cells capable of immediate bone formation
   • Allows for bone induction by recipient bed where nonosseous tissue is influenced to change its cellular function and become osteogenic
2. Lack of histocompatibility differences or immunologic problems
3. Ease of incorporation
4. Lack of disease transmission

Autogenous cancellous bone has osteogenic, osteoinductive, and osteoconductive properties owing to the surviving bone cells, collagen, mineral, and matrix proteins, as well as a large trabecular surface area that is joined together as new bone forms.⁷² The disadvantages of autografts include

1. Additional incision or wider exposure, prolonged operative time, and increased blood loss and trauma
2. Increased postoperative morbidity from pain and potential infection or deformity
3. Sacrifice of normal structure and weakening of donor bone
4. Risks of significant complications
5. Limitations in size, shape, quantity, and quality (the supply is limited, especially in children)

For optimum results harvest autogenous cancellous bone in the following manner:

1. Harvest graft in thin strips (not exceeding 5 mm in thickness ⁵⁹, ¹²⁵)
   • To provide maximum exposure of superficial cells
   • To allow rapid vascularization
2. Wrap graft in a gauze soaked in patient's blood
   • Avoid exposure to high-intensity lights
   • Maintain temperature less than 42°C ⁵⁹
   • Do not store in saline or antibiotic solution ³, ¹³⁰
   • Do not use chemical sterilization ¹³⁰
3. Transfer graft to the recipient bed as soon as possible
   • To avoid exposure to air for more than 30 minutes ⁵⁹
   • To protect the viability of the surface cells
4. Place graft
   • In well-vascularized bone bed
   • In well-decorticated bone surface (cancellous site is superior)
   • With healthy soft-tissue coverage
5. Minimize surgical trauma (e.g., high-speed burring and inadequate irrigation retard healing ¹, ¹⁴³)
6. Position cancellous surface
   • On opposing cancellous surface
   • On surrounding soft tissue with good blood supply
   • So that total mass of graft is not too thick to prevent nutrient diffusion from recipient bed
7. Avoid
   • Dead space
   • Hematoma
   • Interposition of necrotic tissue
8. Minimize risk; be aware of
   • Anatomy
   • Potential complications

The iliac crest is the most versatile bone graft reserve. It is subcutaneous and easy to harvest in prone, supine, lateral, or other positions. It is expendable, and has a large reserve of cortical and cancellous bone. In addition, it allows creation of different shapes and sizes.

Anterior Iliac Crest Grafts
Anterior iliac crest bone grafts are used for anterior interbody fusion of the cervical, thoracic, or lumbosacral spine. The subcutaneous anterosuperior iliac spine and iliac crest are easily palpable. The iliac tubercle is the widest portion, and is where a large quantity of corticocancellous bone is found (see Fig. 8).

A skin incision is made parallel to, or in line with, the iliac crest. It is advantageous to center the incision over the iliac tubercle. The incision is carried down to the bone of the crest, and the muscles are elevated subperiosteally to expose the wing of the ilium.

The tensor fascia latae, gluteus medius, and gluteus minimus originate from the lateral aspect of the ilium They are innervated by the superior gluteal nerve. The abdominal muscles are also attached to the iliac crest and are segmentally innervated. The incision over the crest is, therefore, "internervous" and safe.

An appropriate osteotome or chisel may be used to outline a cortical window in the lateral iliac surface from which to procure the bone graft. Longitudinal parallel cuts may be made (Fig. 1). Strips of cancellous bone may be removed with a curved gauge. Care must be taken not to violate the inner table of the iliac wing, where hernia is a significant potential complication.

Bone graft may be obtained from the inner table of the iliac wing. However, there are risks of peritoneal perforation and significant bleeding with formation of hematoma in the retroperitoneal space.

It is important not to carry the incision to, or anterior to, the anterosuperior iliac spine. Injury to the lateral femoral cutaneous nerve or the inguinal ligament must be avoided. Detachment of the inguinal ligament may result in inguinal hernia. If bicortical bone is taken too close to the anterosuperior iliac spine, fracture may occur (see Fig. 6). Avulsion of the anterosuperior iliac spine may occur by the action of the attached muscles, such as the tensor fascia lata or sartorius.

Bone may be removed in the form of block, dowel, strips, or by way of a cortical window or "trap door" (Figs. 1, 2, 3, 4). The iliac-crest contour can be preserved by removing the bone deep to the crest, or by temporarily detaching and repositioning it later (Fig. 5). The anterosuperior iliac spine should be left intact to maintain normal appearance. The region of the iliac spine should not be weakened by removing bone adjacent to it. Fracture and displacement of the inguinal ligament may result (Fig. 6).

The wound should be closed properly. The muscles and fascia must be sutured to their original anatomic positions and the defects closed; an effective drain should be used.

Posterior Iliac Crest Grafts
The posterior iliac crest provides a large quantity of cortical cancellous bone graft. The posterosuperior iliac crest is palpable under the skin dimple in the superior medial aspect of the gluteal region. The iliac crest curves cephalad and laterally from the posterosuperior iliac spine.

An oblique, curved, or vertical incision may be made over the posterior iliac crest or in line with it. The cluneal nerves cross the iliac crest 7 to 12 cm anterolateral to the posterosuperior iliac spine (see discussion under Complications) and must be protected (see Fig. 13).

A midline spine incision may be extended distally and the posterior iliac crest approached laterally under the skin and subcutaneous fat. This avoids the use of a second skin incision.

The incision is carried down to the bone of the crest, and the muscles are elevated subperiosteally from the posterolateral surface of the ilium. This approach does not denervate the muscles. The gluteus maximus, medius, and minimus originate from the lateral surface of the ilium. The superior gluteal nerve innervates the gluteus medius and minimus, and the inferior gluteal nerve innervates the gluteus maximus. The paraspinal musculature innervated segmentally originates from the iliac crest.

It is very important to remember the following rules:

1. Stay on bone and work subperiosteally.
2. Avoid the sciatic notch and protect the sciatic nerve.
3. Protect the superior gluteal vessels (see discussion under Complications) and protect the pelvic stability.
4. Avoid the sacroiliac joint.
5. Protect the posterior sacroiliac ligaments.

The removal of bone in the vicinity of the sciatic notch can weaken the thick bone that forms the notch. This can produce instability of the pelvis. It is important to stay cephalad to the sciatic notch and remove bone only from the false pelvis. For a landmark, an imaginary line dropped anteriorly from the posterosuperior iliac spine with the patient in the prone position can be used as the caudal limit of bone removal (see Fig. 15, A and B). Care must be taken not to enter the sacroiliac joint, which may become a source of persistent pain and instability when injured.

A sharp surgical instrument (e.g., an osteotome or tip of Taylor retractor) may injure the sciatic nerve deep to the sciatic notch. Laceration of the superior gluteal vessels is a significant danger in this region. The vessels leave the pelvis via the sciatic notch. A divided vessel can easily retract into the pelvis and present a very alarming complication (see discussion under Complications [Fig. 14).

Nutrient vessels supplying the ilium found in the mid-portion of the anterior gluteal line may present troublesome bleeding and should be controlled with Gelfoam, Surgicel, bone wax, or electrocoagulation.

A less painful bone-graft donor site for lumbar spine fusion is possible by applying the following technique. A separate incision over the iliac crest is not made through the skin. The fascia is grasped with Kochers clamps and pulled medially through the wound of the lumbar surgical site. The subcutaneous tissue is carefully elevated off the fascia laterally and caudally until the fascia immediately above the posterior iliac crest and posterosuperior iliac spine is reached. A Taylor retractor is placed in the subcutaneous tissue lateral to the crest over the ilium posteriorly. The periosteum is not dissected from the ilium except from the medial aspect of the crest. The fascia is incised along the medial edge of the crest, and an elevator is used to scrape the medial surface of the crest free of periosteum to bare bone. After a window is made in the medial cortex with an osteotome, gouges are used to remove the cancellous bone between the cortical layers (Fig. 7), leaving the cortices intact laterally and posteriorly with their soft-tissue attachments. This technique minimizes postoperative donor-site pain and prevents formation of uncomfortable scar tissue over the ilium, as when the lateral cortices are removed.

When limited quantity of cancellous bone is required, the following methods may be advantageous:

1. Currettage allows harvest of cancellous graft with least morbidity through a small round cortical window using a sharp curette as shown in Fig. 8. Cancellous bone is most abundant in the posterior aspect of the iliac crest, followed by the iliac tubercle and anterosuperior iliac spine areas.
2. A "trap door" cut in the anterior or posterior outer table of the ilium and hinged on muscles can be opened to allow access to cancellous bone. The trap door is closed at the end. Postoperative pain appears to be less with this technique. Cosmetic deformity is minimal (see Fig. 4).

Wolfe and Kawamoto ¹⁵³ reported a technique of obtaining full-thickness bone graft from the anterior ilium. An incision is made through the iliac crest. The outer ridges of the iliac crest are split obliquely with the muscular and periosteal attachments remaining. All the iliac bone beneath this split can then be removed. The edges of the crest may be reapproximated, thus minimizing cosmetic deformity, hernia, hematoma, and postoperative morbidity (Figs. 9, 10, 11, 12).

Complications
Complications involving the iliac bone-graft donor site are not uncommon. Although some of these complications may not be serious, they add to the patient's discomfort and prolong the convalescence. Complications secondary to graft removal from the ilium include

1. Major blood loss
2. Hematoma
3. Nerve injury (neuroma formation)
4. Severe pain (chronic pain)
5. Hernia
6. Cosmetic deformity
7. Fracture
8. Necessity for sacroiliac joint surgery
9. Pelvic instability
10. Hip subluxation
11. Gait disturbance
12. Peritoneal injury
13. Ureteral injury
14. Heterotopic bone formation
15. Infection

Cockin ²⁸ reviewed 118 cases of iliac crest bone graft procedures and found major complications in 3.4% of the cases and minor complaints in 6%. There were two cases of meralgia paresthetica, one of hernia, and one of hip subluxation after extensive removal of the iliac crest. The minor complaints included wound pain, hypersensitivity, and buttock anesthesia.

Younger and Chapman reviewed the medical records of 239 patients with 243 bone grafts. ¹⁵⁵ The overall major complication rate was 8.6% and included infection (2.5%), prolonged wound drainage (0.8%), large hematomas (3.3%), reoperation (3.8%), pain greater than 6 months (2.5%), sensory loss (1.2%), and unsightly scars. The minor complication rate was 20.6% and included superficial infection, minor wound problems, temporary sensory loss, and mild or resolving pain. A significantly higher major complication rate of 17.9% occurred if the surgical incision was also used to harvest the bone graft.

Nerve Injuries
Possible nerve injuries include the following:

1. Lateral femoral cutaneous nerve ²⁸, ⁸⁵, ¹⁴⁴
2. Iliohypogastric (lateral cutaneous branch) nerve
3. Superior cluneal nerve (cutaneous branches of dorsal rami L1, L2, L3) ²⁹, ³⁹
4. Middle cluneal nerve (cutaneous branches of dorsal rami S1, S2, S3)
5. Sciatic nerve
6. Ilioinguinal nerve ¹⁴
7. Femoral nerve
8. Superior gluteal nerve

Superior cluneal nerves are lateral branches of the posterior primary division of the upper three lumbar nerves that run posteriorly through the lumbosacral fascia at the lateral origin of the sacrospinatus muscle. They cross over the dorsal aspect of the posterior iliac crest and provide sensation to the skin of the buttocks. They are found 7 to 12 cm anterolateral to the posterosuperior iliac spine in the adult. When an incision is made across or parallel to the posterior iliac crest, the cluneal nerves may be injured (Fig. 13).

Painful neuritis of the buttocks has been reported. ²⁹, ³⁹ That these nerves are a cause of disability can be demonstrated by the relief of symptoms after the nerves have been infiltrated with local anesthetics. Permanent relief can be obtained by resection of the nerves with the transected ends being allowed to retract into the soft tissue.

A possible complication of anterior iliac bone graft procurement is injury to the lateral femoral cutaneous nerve. ⁸⁵, ¹⁴⁴ This nerve arises from the dorsal branches of the second and third lumbar nerve roots, crosses the ilium obliquely, and passes medial to the anterosuperior iliac spine under the inguinal ligament. Then it courses toward the sartorius muscle and divides into anterior and posterior branches. The anterior branch passes through the fascia lata and provides sensation to the anterior and lateral aspects of the thigh. The posterior branch penetrates the fascia lata and provides sensation from the region of the greater trochanter to the mid-thigh. Sometimes the nerve may have an anomalous route lateral to the anterosuperior iliac spine. ⁵³ Therefore, when dissection is carried out in the region of the anterosuperior iliac spine the nerve may be injured. It can also be entrapped in postsurgical scar tissue. Excessive retraction of the iliacus muscle when the medial wall of the ilium is exposed may also injure this nerve. Compromise to the lateral forward cutaneous nerve presents as meralgia paresthetica, which may involve numbness, paresthesias, burning, and/or pain over the anterior and lateral thigh. Most of the time meralgia paresthetica causes minor symptoms or resolves spontaneously after the local causes are corrected. For some severe and resistant symptoms, nerve injections or surgical intervention may be necessary. Surgery may involve neuroma excision or freeing the nerve from the scar tissue entrapment.

The sciatic nerve may be injured when the dissection is extended down to the sciatic notch. A surgical instrument such as an osteotome may be passed deep to the sciatic notch to cause this injury. The bony rim of the notch should be palpated before the dissection is carried to this area. An imaginary plumb line dropped from the posterosuperior iliac spine with the patient in the prone position will pass through the bony rim of the sciatic notch. This serious complication can be avoided by staying cephalad to this line (Figs. 14 and 15).

The ilioinguinal nerve may be injured when the abdominal wall is retracted medially from the anterior iliac crest. The nerve may be compressed beneath the retractor on the inner part of the wall of the ilium. It occurs when the inner cortex of the anterior ilium is exposed for removal of bone grafts. Ilioinguinal neurologic injury is characterized by pain radiating from the iliac toward the inguinal and genital areas. This complication is well-discussed by Smith and associates. ¹²¹

The iliohypogastric nerve (lateral cutaneous branch, L1 ventral rami) is found over the midlateral aspect of the iliac crest. It should be protected when working in this region (see Fig. 13).

Severe Donor-Site Pain (Chronic)
Chronic donor-site pain from the ilium was reported in 25% of 290 patients who had undergone anterior spine fusion in Summers' and Eisenstein's study. ¹²⁶ Fernyhough and co-workers explored the relationship between surgical approach and chronic posterior iliac crest donor-site pain in 151 harvests. ⁴⁷ No difference was observed in the incidence of chronic donor-site pain between harvests performed through a primary midline incision versus a separate lateral oblique incision (28% vs. 31%). Twice as many donor sites harvested for reconstructive spinal procedures were reported to have chronic pain as compared with those harvested for spinal trauma, regardless of the approach used (39% vs. 18%). Summers and Eisenstein reported the highest prevalence of donor-site pain to be in patients who had a tricortical full-thickness graft taken through a separate incision overlying the iliac crest. ¹²⁶ Patients with a clinically unsatisfactory result from spine fusion also had a significantly higher prevalence of donor-site pain. ¹²⁶

Vascular Injuries
Vascular injuries may include the superior gluteal artery (and vein), ⁴⁵, ⁶⁶ the deep circumflex iliac artery, the iliolumbar artery, and the fourth lumbar artery.

The superior gluteal artery is a branch of the internal iliac artery that curves around the rim of the sciatic notch as it leaves the pelvis. It may be injured when dissection is carried close to the sciatic notch. An osteotome or the sharp point of a Taylor retractor may enter the notch and pose similar danger to the artery. This complication can become alarming, since the divided vessel easily retracts into the pelvis (see Fig. 14).

If the superior gluteal vessel is lacerated, it can be compressed locally and exposed for ligation or clipping. A finger may be used to apply direct pressure to the vessel, against the bone. Kahn ⁶⁶ discussed the use of a Raney-modified Kerrison rongeur to remove the upper margin of the sciatic notch to expose the bleeding vessel. If the bleeding vessel is still not accessible, the patient may be positioned for a retroperitoneal or transperitoneal exposure of the vessel. Arterial occlusion by embolization or by use of a Fogerty catheter is another option.

Injury to the superior gluteal vessels can be prevented if the surgeon is aware of the anatomy in this region. The bony origin of the gluteus maximus or the roughened area anterior to the posterosuperior iliac spine is a good landmark and can be used as the caudal limit of bone removal (see Fig. 15, A and B). An imaginary plumb line dropped from the posterosuperior iliac spine with the patient in the prone position will pass through the bony rim of the sciatic notch. It is important to stay cephalad to this line.

Escalas and DeWald ⁴⁵ reported a case of combined traumatic superior gluteal arteriovenous fistula and ureteral injury complicating removal of a bone graft from the posterior ilium. The tip of a Taylor retractor accidentally dislodged and penetrated into the sciatic notch to cause this unusual injury.

The deep circumflex iliac artery, the iliolumbar artery, or the fourth lumbar artery may cause troublesome bleeding when working on the inner table of the ilium. Occasionally, peritoneal perforation accompanies the arterial injury. The anatomic position of the arteries are illustrated in Figs. 16 and 17. It is very important to stay subperiosteal and carefully elevate the abdominal wall muscles off the crest and the iliacus muscles off the inner table of the ilium (Fig. 18).

Hernia

A hernia through the iliac bone-graft donor site may occur after the removal of a full-thickness bone graft from that site. It may appear as an iliac swelling, sometimes associated with pain or symptoms of bowel obstruction. ⁵¹, ⁷⁸, ⁹⁶, ¹⁰⁶, ¹⁰⁹ Strangulated hernia and valvulae are very rare occurrences. ²⁵ Symptoms have been reported to have occurred from 24 days ²⁵ to 15 years ¹⁴¹ after the formation of the iliac defect. ³¹

Treatment requires the reduction of the hernia and repairing the defect by

1. Using soft-tissue ⁹, ²⁵, ⁷⁸, ⁹⁶, ¹⁰⁹ advancement, imbrication, flaps, or fascial flaps
2. Using a prosthesis 106 (tantalum or Marlex mesh)
3. Using methylmethacrylate cement to reconstruct the iliac wall ⁷⁹ (Figs. 19, 20, 21)
4. Using the Bosworth technique ⁹, ³¹: removing the remaining wings of the ilium on either side of the defect, followed by layered soft-tissue closure

Pelvic Instability

Removal of a large quantity of bone graft from the posterior ilium may disrupt the mechanical keystone effect of the sacroiliac joint and the posterior sacroiliac ligament, causing instability. Lichtblau ⁷⁶ first reported such complications after a bone-grafting procedure in which the posterior sacroiliac ligaments were postulated to be interrupted. The ensuing instability transferred the stress forces to the pelvic ring, causing fractures of the superior and inferior pubic rami. Coventry and Tapper ³⁰ reported six cases of pelvic instability following removal of bone graft from the ilium. The patients with such instability often developed symptoms indistinguishable from other spinal disorders. History of clicking or thudding, as well as pain in the thigh and gluteal region, is characteristic.

Sacroiliac stability is maintained by formation of the sacrum as a keystone with interlocking eminences and depressions, along with ligamentous support mostly in the posterior and superior aspect. ¹⁴ Multiparous women with lax ligaments and anatomic variations in the sacroiliac joints are more prone to develop such pelvic instability. Radiologic examination of the entire pelvic ring is important. Changes in the sacroiliac joint, the pubic rami, and the symphysis pubis should be looked for.

Graft Sites

The Tibia
The tibia provides strong full-thickness cortical graft material and is occasionally used in spine fusion. The subcutaneous anteromedial aspect of the tibia is a convenient donor site. The periosteum should be left intact and sutured over the defect. The condyles also supply cancellous bone. However, significant risks exist when using the tibia as a donor site. Biomechanically, the tibia is changed from a closed section to an open one when a bone graft is obtained. It is markedly weakened and much less able to resist torsional and bending loads.

Frankel and Burstein ⁴⁸ discussed the effect of cortical graft removal from the tibia. They described the torque and angular deformation owing to failure of the tibia to be reduced to 30% of normal, and energy absorption capacity to 10% of normal. Even when the corners of the cutout are rounded, open section overshadows any reduction in stress concentration gained.

Fatigue fractures are fairly common, and the tibia should be immobilized in a cast for 6 to 12 months after the bone graft is obtained. ⁴¹ Thus, the disadvantages of autogenous tibial graft far outweigh the benefits.

The Fibula
Although the upper two thirds of the fibula may be removed as a bone graft, the middle one third provides the best cylindric cortical bone graft. The fibula graft is strongest in resisting compressive loading and can be depended on for longer periods of structural support in interbody fusion. For large defects in the vertebral bodies, fibula struts may be used to achieve stability. Because of the small amount of cancellous bone in the fibula, iliac cancellous graft should be supplemented to enhance osteogenesis.

Peroneal nerve injuries may occur when obtaining the graft from the proximal one third of the fibula. Valgus deformity of the ankle is a serious risk when the lower one third is violated. Significant donor-site pain and compartment syndrome have also been reported.

Ribs have been used for thoracic spine fusion. However, their modest cortex and porous cancellous bone are rarely appropriate for lumbar spine fusion.

Free-Vascularized Bone Grafts
Free-vascularized bone grafts may be used to circumvent the disadvantage of large cortical grafts, most of which become necrotic. ⁸⁹, ¹⁰⁷ Progress in microsurgical techniques is making this type of graft possible. ³⁵, ¹²⁷, ¹⁴⁵ Of the different patterns of blood supply to cortical bone, nutrient artery enters the diaphysis and divides into ascending and descending branches. The diaphyseal cortex is supplied by the radially oriented branches. Specific arteries also supply the epiphyseal and metaphyseal areas. A periosteal blood supply from the surrounding muscles enters the outer third of the cortex. ⁵⁴ Continuing circulation and increased viability of the bone grafts facilitate the problem of fracture healing. Vascularized grafts are less dependent on the recipient bed for survival, and their use is advantageous in a poorly vascularized bed with deficient soft tissue after previous surgery, trauma, infection, or irradiation. The fibula, rib, and anterior or posterior ilium may be used. However, their application is usually limited by the small size and need for time-consuming, highly specialized microvascular techniques. Superior results have been reported in clinical cases in which a vascularized fibula graft was transplanted into a segmental bony defect, larger than 6 cm. ³⁵, ⁵⁴ Rapid healing and hypertrophy of the graft were noted.

The use of free-vascularized bone graft may be advantageous in spine fusion in special circumstances. Bradford ¹¹ reported favorable results with vascularized rib pedicle grafts used in patients with post-traumatic kyphotic deformity. The superiority of vascularized rib grafts in bridging vertebral bodies was also demonstrated in canine experiments. 119 Free-vascularized fibular grafts have been applied clinically in spine fusions. ³⁸, ⁶⁷

Dupuis and co-workers ³⁸ successfully used a free-vascularized fibular graft in a case of progressive congenital kyphosis, following the work of O'Brien and Ostrup. ¹⁰⁰ In a similar situation, an avascular strut graft becomes weaker to the point of mechanical failure as it is replaced by creeping substitution, which may take 2 or more years to complete.

Muscle-pedicle bone grafting procedures were reported by Hartman and associates for failed lumbosacral spinal fusion. ⁵⁷ An iliac crest autograft with an intact quadratus lumborum muscle pedicle was used in this case.

Allografts
Allografts are the most frequently used alternatives to autografts in spine surgery. They are bones transplanted from one individual to another and are used to circumvent the problems encountered with autografts. ¹³, ⁴⁴, ⁹⁵, ¹⁴⁶ Fresh, frozen, lyophilized, and demineralized bone matrix from allografts are used clinically.

Allografts are readily available and come in a wide variety of shapes and sizes. They can provide immediate support and minimize the use of stabilization hardware or braces. Bone allografts can replace missing structures and become incorporated into the spine. They provide biologic scaffolding that is gradually replaced with the patient's own bone.

Major problems lead to the decreased effectiveness of allografts. ¹⁹, ²⁰, ²¹, ⁵⁸ Immunologic rejection of implanted graft, ⁷, ¹²⁰ delayed union, nonunion, and fracture of the graft are not uncommon. Incorporation of allografts by the host is slower. Vascular penetration is slower and less dense. There is less perivascular new bone formation when compared with autografts. Transmission of disease from allografts is also a serious concern.

The major weakness of the allograft is that it is decd and cannot contribute directly to osteogenesis, as do fresh autografts. Burwell ¹⁸ found a way around this problem by combining the osteogenic potential of autogenous marrow with allografts. The use of autogenous marrow to provide superior osteogenic capability in allografts and xenografts, as well as autografts, is finding greater clinical application (see further discussion on xenograft and synthetic implants). The use of bank bone would be very advantageous if storage problems, immunologic reactions, and infection could be eliminated. The allograft must be aseptically obtained soon after death or properly sterilized and processed early to

1. Minimize its antigenicity
2. Prevent degradation by proteolytic enzymes
3. Maintain the mechanical structure
4. Preserve the osteogenic induction property

Freezing and Freeze Drying
Freezing and freeze drying are the most widely used preservation methods that allow storage of bone in a biologically useful (but nonviable) state. ¹⁵, ⁴⁹, ⁷¹, ^{74, 82}, ⁸³

Allograft freezing is carried out as soon as possible after procurement. The length of safe storage for bone is not currently known. However, based on the knowledge of autolysis retardation by cold, lower temperatures are expected to extend the "shelf life" of allografts. ⁴⁹ At -15° C to -30° C, using a hometype mechanical freezer, long-term storage of bone is difficult. This form of freezing is not advisable because ice crystals grow rapidly in this temperature range and mechanically destroy the tissue viability. ⁸³ Freezing at -76° C is achieved in dry ice. At -60° C to -90° C, using a laboratory-type mechanical deep freezer, and at -150° C or colder, using a refrigerator with cryogenic gases, more effective preservation of bone is possible. At temperatures near -70° C, ice crystal formation is slower. ⁸³ Bones frozen to -70° C have been stored for several years and successfully applied clinically. ⁴⁹ Freezing in a cryoprotective agent such as glycerol at controlled cooling velocity may be a more effective option.

Freeze drying is a process in which the bone is first frozen to -70° C and then sublimated in a high vacuum. The bone is freeze dried until the water content is reduced to 5% or less. The freeze-dried bone graft can then be shipped and stored conveniently and indefinitely at room temperature in a vacuum container. Since freeze-dried bone is very brittle, it must be reconstituted by immersion in normal saline before use. ⁸⁴ The reconstitution time depends on the size and shape of the graft. Chips of bone may not require any rehydration, whereas larger cortical-bone may require up to 24 hours for reconstitution. ⁴⁹

In the clinical application of freeze-dried allografts in spine fusion, Malinin and Brown ⁸³ reported that the union rate of such allografts under compression load (interbody or strut grafts) was not delayed by low-grade immunologic response. This is in contrast to a high incidence of resorption with cortical graft placed under tension posteriorly in the spine.

Biomechanical Properties
Biomechanical properties of bone grafts may be changed by the techniques used for preservation, although Sedlin's study showed that freezing and thawing do not significantly alter the mechanical properties of bone. ¹¹⁷

Bright and Burstein, ¹² as well as Triantafyllou and co-workers, ¹²⁹ studied the biomechanical properties of freeze-dried and irradiated bones. Komender ⁷⁰ found that freezing to -78° C does not alter the mechanical properties of bone. Pelker and associates ¹⁰³ also concluded that freezing allograft bone to temperatures as low as that of liquid nitrogen (-196° C) does not significantly alter its biomechanical properties. They showed that freeze drying does diminish the torsional and bending strength but not the compressive strength. The data of Pelker and co-workers indicate that frozen bones are better suited than freeze-dried bones when they are subjected to torsional loads. ¹⁰², ¹⁰³

Both frozen and freeze-dried bones are acceptable when compressive forces are the primary concern. It must be remembered that the initial biomechanical properties of the bone graft will change with resorption, incorporation, and remodeling by the host. Surgical technique, internal fixation, and postoperative management must therefore be planned accordingly.

Radiation, Heat, and Chemical Treatment of Allografts and Xenografts
Although most of the aseptically procured cadaver bones do not require sterilization, allografts or xenografts have been sterilized by physical means such as high-energy radiation and heat (boiling, autoclaving); and by chemicals such as thimerosal (Merthiolate), ethylene dioxide, propiolactone, or antibiotic solutions. Some of these sterilization methods were used in the past and are discussed for historic interest only.

Radiation of at least 2 megarads is required to kill bacteria; 4 megarads inactivates some viruses. ⁸³ The same dose of radiation (2 to 4 megarads) needed to sterilize or destroy antigens also significantly impairs the inductive repair capacity of the bone graft. ¹⁶, ¹³², ¹³³ Increased solubility of collagen and glycosaminoglycan, destruction of the bone matrix fibrillar network, ¹⁶, ²² and discoloration ⁸³ of irradiated bone have been reported. In cobalt-60 irradiated bone, Ostrowski ⁹⁹ reported free radicals of unusual stability to be present, although their effect on the host tissue is unclear. ⁸³

The effect of radiation on the biomechanical properties of bone is not well defined at this time. The effect seems to be minimal with low-level radiation. Radiation doses exceeding 3 megarads are known to destroy bone matrix fibrillar network. ¹⁶, ²² There appears to be a significant drop of breaking strength of bone with more than 3 megarads. This effect is magnified when radiation is combined with freeze drying. Komender noted that 6 megarads of radiation reduced the strength in bending, compression, and torsion. ⁷⁰ Irradiation of freeze-dried bone with only 3 megarads markedly diminished the bending strength, but not the strength in compression or torsion.

Boiled bones have been used for graft material since the early part of this century. ⁵² Although some good clinical results have been reported, ⁷⁷, ¹⁵² boiled allografts and xenografts have generally produced undesirable results. Boiling destroys all inductive capacity. ¹⁵² Heat may be intended to destroy transplantation antigen, but it only denatures the antigenic proteins into another unacceptable material.

It may be tempting to use autoclaved contaminated bone in the operating room, but this produces haversian canal coagulation and denaturation of bone protein, ¹⁶, ⁸³ which severely retard host incorporation. Autoclaving destroys bone morphogenic protein activity.

Chemical processing of bone graft may present significant problems such as potential carcinogenesis and difficulty of penetration into bone. Propiolactone (1% solution) has been found to be more bacteriocidal than ethylene dioxide, which is more difficult to use. ⁸³

Merthiolate-treated grafts have, in general, produced poor results; 30% of the grafts failed in the study by Reynolds and co-workers, ¹¹⁰ with apparently three times as many failures as with autografts. Reduced callus formation and osteogenesis have been noted. When the graft fractured, there was minimal healing. When washed before use, no significant host sensitivity to Merthiolate was noted in Merthiolate-treated bone graft. ⁸³

Benzalkonium chloride completely destroys the osteogenic inductive capacity of bone, according to Urist and associates. ¹³⁵ Antibiotic solutions do not penetrate completely into bone. Their germicidal effect in bone graft is variable. In general, antibiotics appear to inhibit the osteogenic inductive capacity of bone. ³, ⁸³

A prospective study evaluating the radiographic appearance of demineralized bone matrix (DBM) chemosterilized with ethylene oxide in posterolateral lumbar spine fusion in human beings was presented by Zucherman and associates. ¹⁵⁶ Sixty-eight levels of bilateral posterolateral lumbar spinal fusions were performed using combinations of autologous bone graft (ABG) and DBM. Patients were randomly assigned to three groups: Group 1 ²⁰ (the control) consisted of bilateral placement of the ABG/DBM mixture, Group 2 ³⁰ consisted of ABG on one side with the ABG/DBM mixture on the other, and Group 3 ³⁰ consisted of ABG on one side and DBM on the other. Mean follow-up was 15 months. An orthopedist and a radiologist blinded to the content of the fusion masses rated the fusions on a scale from zero (resorption) to three (confluence) on the basis of appearance of consolidation on an anteroposterior radiograph of the lumbar spine. Results of this prospective series suggest that in posterolateral spine fusion in human beings, it is less effective than autologous bone graft. ¹⁵⁶

Bone Morphogenetic Protein
Urist states that allograft bone must be removed from the donor within 4 to 8 hours after death or within the minimal biodegradable time. ¹³² Radiation sterilization with more than 2 megarads; heating over 60° C; exposure to chemicals such as hydrogen peroxide, betapropriolactone, or benzalkonium chloride; cryolysis; immediate freeze drying; and prolonged storage at 0° C to 30° C must be avoided to preserve the inductive properties of the bone. Urist and co-workers have extensively studied osteogenic induction and discovered the biologically active factor that they called bone morphogenetic protein (BMP), which modulates recruitment of mesenchymal stem cells that differentiate into osteoblasts from surrounding tissues. ¹³¹, ¹³², ¹³⁵, ¹³⁶, ¹³⁹ Bone morphogenetic protein has been extracted from bone-forming tumors, dentin, and diaphysis of cortical bone. Studies demonstrated a number of highly active proteins with molecular weights of approximately 30,000. ⁸⁰, ¹¹⁵, ¹⁵⁴ To date, seven BMP proteins have been isolated from bone-inducing preparations using high-resolution protein purification techniques and recombinant DNA technology. ⁶ Four proteins (BMP-2, BMP-3, BMP-4 and BMP-7) have been shown to induce bone formation in animals. ⁶, ¹⁴² Their clinical application has been evaluated in the form of an injectable substance linked to different delivery systems. ¹³⁷

A chemosterilized bone that is an autodigested, antigen-extracted allograft (AAA) has been developed and clinically tested by Urist and co-workers. It is an allogeneic bone of high osteogenetic property and low immunogenicity prepared in five basic steps. Urist believes that BMP is also preserved by these measures. ¹³², ¹³⁸, ¹³⁹

Urist and Dawson ¹³⁴ reported 40 intertransverse-process fusions in 36 cases of degenerative joint and disc disease including spinal stenosis and spondylolisthesis, as well as four cases of thoracolumbar fracture dislocation. A composite of AAA cortical bone strips and local autologous bone was used in all cases. There were over 80% excellent and good results, with a nonunion rate of 12%.

Xenografts (Heterologous Bones, Heterografts)
Xenografts, or bones transplanted from other species, have been used in spine surgery. ⁶⁴, ⁸⁶ The advantages of xenografts are the almost unlimited supply from animal sources; wide range of sizes, shapes, and strength; and cortical-cancellous ratio. The application of ivory, ⁶³, ⁸¹ animal horns, ⁶³ corals, ⁶⁰, ⁶¹ and other exotic materials has been explored. Animal horns and ivory are very resistant to incorporation into the host bone. ¹¹² Fresh xenograft bones have been shown to be unacceptable. Due to the antigenicity of the foreign tissue they are unable to generate satisfactory bony repair. Invariably, they produce inflammation, fever, sequestration, resorption, or other manifestations of rejection. ¹⁶ Fibrous envelopment occurs as over a metal plate. Even when fusion takes place, sequestration of the xenograft is observed. Urist believes that xenografts should not be used in patients. ¹³²

Partially deproteinated and defatted xenografts ⁴³ were reported to have markedly decreased antigenicity resulting in a minimized immune response. However, the osteoinductive capacity is reduced by the denaturing process that destroys BMP and other osteoinductive proteins.

Bovine bones have been popular because they incorporate and remodel with less difficulty. ¹¹² Different types of preserved bovine bone have been tried since the nineteenth century, including

1. Frozen calf bone
2. Freeze-dried calf bone (Boplant) ¹⁰⁴
3. Decalcified ox bone ⁵⁵ (as well as decalcified calf and sheep bone) was evaluated experimentally and clinically, but found to be unsatisfactory.

Deproteinized xenografts, including "os purim," "anorganic bone," "Oswestry bone," and "Kiel bone," ⁸⁶, ¹¹², ¹¹³, ¹¹⁴ have also been tried.

Kiel bone, partly deproteinized bone from freshly killed calf, sterilized either by ethylene dioxide or by gamma radiation, is commercially available. Experimental studies showed that it is weakly antigenic and does not possess active bone-inducing capacity. ¹¹⁴ Since its introduction in 1957, Kiel bone has been used in almost every possible bone-graft site, and varying success rates have been reported clinically. ⁶⁴, ⁸⁶, ¹¹²

In spinal fission, Jackson ⁶⁴ noted that Kiel-bone implant became surrounded by autogenous bone with time. For larger defects, he recommended the use of an autogenous and Kiel bone composite. McMurray ⁹² presented clinical, radiologic, and histologic data on the fate of Kiel-bone implants in four anterior spine fusions that failed. Biopsies of the Melbone implants showed invasion by fibrous tissue. There was no ossification and no incorporation into the surrounding bone. Such deproteinized bone could be invaded by host new bone when placed in an excellent vascular bed with potentially osteogenic cells. When impregnated with autogenous bone marrow cells, it may prove to be an excellent scaffolding with good bone-conduction property. ¹¹², ¹¹⁴ Plenk and associates, ¹⁰⁵ Salama. ¹¹² and Salama and co-workers ¹¹³, ¹¹⁴ reported good results using autogenous bone marrow and Kiel bone as composite grafts in patients. The red marrow can be easily aspirated from the patient's own iliac crest. The bone marrow is protected from the action of the adjacent tissues when the deproteinized bone is present to serve as an osteoconductive supportive structure.

Synthetic Implants
Synthetic implants can be prepared to fit any size or shape, but they have traditionally been considered to be subject to wear and not incorporate biologically into the host bone. ¹⁶ Metals such as titanium, ceramics, and polymers (e.g., polyethylene with porous surfaces) could allow ingrowth of bone when the interface between the implant and living bone is stable. When significant motion is present between the implant and the bone, fibrous ingrowth occurs. A number of implants fashioned from metals have been tried as replacement for bone in the spinal column (see, for example, Steffee's titanium vertebral replacement: "Total Vertebral Body and Pedicle Replacement" ¹²⁴ [Steffee A: personal communication, 1991]).

Metal scaffolds in the shape of the bone being replaced may be covered by ground autologous bone grafts of small particle size. In animal experiments, ingrowth of bone occurred over the total surface area of fiber metal implants and bone penetrated deep into the composite. ²

Titanium mesh implants have been clinically applied by Leong ⁷⁵ (Leong JCY: personal communication, 1984) and co-workers for anterior spinal fusion after discectomy in the lumbar spine. This porous implant allows ingrowth of bone and appears to obviate the use of bone graft. It acts as a spacer and can provide immediate stability, while allowing time for the slow ingrowth of bone and long-term stability.

Experimentally, porous titanium mesh blocks with a 50% void allow rapid ingrowth of bone in canine long bone. A 12-year follow-up was possible in two patients who are asymptomatic, and the implants have remained unchanged and undisplaced; 10 patients had a follow-up of more than 5-years. Of these, seven patients were asymptomatic, two had more than 70% symptomatic relief, and one retained a very stiff back. Radiologic analysis showed that disc height was maintained at 5 years with no movement between the adjacent vertebral bodies, often with bony overgrowth anterior to the implant.

Steffee and co-workers ¹²⁴ reported a total vertebral body replacement and artificial pedicle replacement with the aid of a special segmental spine plate fixation. ¹²³, ¹²⁴ (Steffee A: personal communication, 1991). Their system allows immediate and rigid fixation after extensive decompressions and radical tumor excisions. Theoretically, the total vertebral replacement may be performed using metallic or nonmetallic implants with a porous structure that allows bony ingrowth.

Nonmetallic Synthetic Implants
A growing number of other synthetic implants are being used as bone substitutes. These biosynthetic bone-graft substitutes address the disadvantages of both autogenous bone harvest and allograft problems. ⁷² According to Osborn and Nemesley 98 the chemical nature of the implant determines the biodynamics and reaction of the recipient bed in the interaction with living bone. They considered the following materials ²³, ⁹⁸:

1. Bone cement and stainless steel are biotolerant, resulting in distance osteogenesis with a fibrous layer separating the implant from bone.
2. Alumina and carbon materials are bioinert, resulting in contact osteogenesis.
3. Glass ceramic, calcium phosphate ceramics, and hydroxyapatite ceramics are bioactive, resulting in boding osteogenesis.

Bioinert porous ceramics of alumina were noted by Benum and associates ⁴ to be bound to bone by the ingrowth of bone 3 to 4 mm thick in regions exposed to compressive forces.

Evidence suggests that porous calcium phosphate ceramics are the most biocompatible synthetic-bone substitute, with the ability to become chemically bonded by living bone and with a chemical composition devoid of toxicologic liabilities. ⁶⁵ They are shown to be superior to biodegradable polymers, such as polylactic acid and polyglycolic acid, which have been considered as bone substitutes. ²⁴, ⁶⁵ The implants may be dense or porous. The minimum pore size for ingrowth of bone is shown to be 100 µm. ⁶⁸ Corals provide such porous structures ⁶⁰, ⁶¹

Holmes and co-workers ⁶¹ performed histologic and biomechanical studies in dogs using hydroxyapatite converted from sea coral calcite as bone substitute. The material was incorporated in bone and became almost as strong as the native bone. They also reported encouraging clinical application with fractures in 18 patients. Hydroxyapatite is the natural bone mineral. There is great potential of developing or discovering minerals so close to human bone that normal bone turnover would occur in a physiologic environment. Such porous mineral structure would undergo appropriate gradual and timely biodegradation and eventually be replaced by living bone. Tricalcium phosphate, for example, has the potential to undergo slow biodegradation and be replaced by bone in the living system. ⁸⁸

Another material, replam hydroxyapatite-porites (RHAP), is a ceramic with three-dimensional interconnected porous material of calcium hydroxyapatite from the exoskeleton of porites (coral). It may be carved by the surgeon before implanting. Replam hydroxyapatite-porites was approved for evaluation in spinal fission in several centers under Mooney and associates. ⁸⁸

Bioactive and biodegradable porous ceramics of hydroxyapatite or tricalcium phosphate have been studied. Jarcho ⁶⁵ stated that they are usually well tolerated and become chemically bonded to bone by natural bone-cementing mechanisms.

Porous hydroxyapatite ceramics have been used in canine experiments for the spine ³³ and other skeletal defects. Porous ceramics and autologous marrow composites were studied by Nade and associates. 91 Porous alumina, calcium aluminate calcium hydroxyapatite, and tricalcium phosphate were placed with bone marrow into intermuscular sites. Bone was found to adhere to the ceramics and to penetrate the interior if the pore size was greater than 100 µm. The marrow cells were shown to play a significant part in new bone formation into the framework. Nade and co-workers believe that the appropriate histocompatible, biodegradable ceramic material would act as a scaffold by virtue of its porosity for retention of bone-marrow cells, and provide mechanical strength while bone ingrowth progresses. This type of bone substitute would also allow a wide selection of sizes and shapes in sterile form. More recently, Ohgushi and co-workers ⁹⁴ demonstrated improved biomechanical properties with increased bone growth when porous ceramics were combined with bone marrow.

Porous biodegradable ceramic and BMP composites were evaluated by Urist and co-workers. ¹³⁷ They reported that an aggregate of B-tricalcium phosphate and bone morphogenetic protein (TCP/BMP) induced the differentiation of cartilage in 8 days and in lamella bone in 21 days. The yield of new bone was more than 12 times greater from the TCP/BMP than from the BMP alone. It is possible that a porous ceramic acts as a slow-release delivery system to distribute BMP more favorably and to potentiate its activity.

Calcium phosphate-coated metallic implants showed superior bone-bonding characteristics according to Ducheyne and associates. ³⁶ Such implants may solve the problem of weak mechanical strength of ceramics, particularly the porous ones. Ceramic implants by themselves are probably unsuitable for restoration that would have to withstand significant impact, or torsional or bending stresses, ⁶⁵ as in the spinal column.

Calcium phosphate-containing bone cements are also being developed. Calcium hydroxyapatite in powder form was used as an expander of a patient's own cancellous bone graft. It has been used in spine fusion, especially in children, when sufficient autologous bone graft is unavailable (Luque ER: personal communication, 1985). In general, the various synthetic hydroxyapatite and tricalcium phosphate ceramic or crystalline preparations have been shown to be nontoxic and biocompatible, and to have the potential to form intimate bonding with the host bone. 50 They are osteoconductive and usually not osteoinductive by themselves. These synthetic preparations have the greatest potential as a vehicle for bone marrow, BMP, and other bone-inducing agents. ⁵⁰

Collagen and Other Matrix Proteins
Major constituents of demineralized bone matrix are collagen and other matrix proteins. The mineralization process relies on the extracellular organic matrix. It is believed that fibrillar collagen, the main component of this matrix, has the capacity to serve as the structural osteoconductive scaffolding for the mineralization. Osteonectin, a noncollagen matrix protein, is complex with Collagen. ¹²⁸ Proteolipids and calcium acidic-phospholipid phosphate complexes, ¹²⁸, ¹⁴⁰ as well as bone and dentin phosphoprotein, were demonstrated to cause hydroxyapatite formation in vitro. ⁸, ¹⁰, ⁹³, ¹²⁸, ¹⁴⁰ Fibronectin, a matrix and cell membrane protein, is believed to act as a binder for mesenchymal-cell attachment to collagen matrix, which in turn allows the osteoinductor to contact cell-surface receptors. ¹⁴⁷ Wernts and coworkers demonstrated that soluble fibrillar collagen and marrow composite is more effective than even cancellous bone for bridging large bone defects. ¹⁴⁸ However, without the marrow, the collagen by itself is inferior to cancellous bone. ⁷² A combination of soluble fibrillar collagen, marrow, and ceramic of 40% tricalcium phosphate and 60% hydroxyapatite with the ceramic constituting 25% to 50% of the volume is effective in healing rat femoral segmental defects. ⁷², ⁷³ With further studies, similar synthetic composites may prove to be clinically useful.

Methylmethacrylate Cement
Knight ⁶⁹ was the first to report the use of acrylic cement to fix the cervical spine with chronic fracture dislocation, atlantoaxial subluxation, and cervical spondylosis. He also stabilized the lumbar spine using the cement in one patient with disc disease. Scoville and co-workers ¹¹⁶ reported the use of acrylic plastic for vertebral replacement or fixation in metastatic tumor destruction of the spine.

Harrington ⁵⁶ documented the use of methylmethacrylate for vertebral body replacement and anterior stabilization of the spine with metastatic tumor. His series included 14 patients treated by anterior decompression and stabilization using metal and bone cement. The strength of methylmethacrylate is about one half that of bone. ¹⁵¹ Attempts have been made to strengthen the cement by adding fibers, ⁸⁷ but clinical data are still unavailable. After polymerization, methylmethacrylate becomes a rigid and brittle solid that can withstand significant compression. However, it fails under tension or shear forces. It is reasonable to use in the replacement of a vertebral body where compression is the predominant force present. It is important to remember that when used alone, the outer part of the cement mass is still subject to tension when bending, and will fail with time in a clinical setting. The primary indication for application of methylmethacrylate in spinal stabilization is in patients with malignant disease and limited life expectancy. ⁴², ⁹⁰ It should not be expected to provide long-term support of the spine. ³⁷

In the spinal column, methylmethacrylate cement should be used with secure metal fixation. It may be used as reinforcement for screws and hooks in cancellous bone. The cement does enhance fixation of implants by increasing the contact area, especially in osteoporotic bone.

Fig. 1 Corticocancellous bone graft is obtained from the lateral iliac surface using longitudinal parallel cuts with an osteotome or chisel. back	Fig. 2 Horseshoe-shaped corticocancellous bone graft is obtained from the iliac crest using osteotomes positioned to each other. back
Fig. 3 Dowel-cutting instrument is used to obtain an iliac graft with two tooled cancellous surfaces and cortical faces on three sides for anterior spinal interbody fusion. back	Fig. 4 Cortical "trap door" is used to gain access to iliac cancellous bone. back
Fig. 5 Large iliac graft is obtained with preservation of the iliac contour. back
Fig. 6 Illustration of attachment of inguinal ligament to anterior superior iliac spine. Detachment of inguinal ligament may lead to inguinal hernia. Injury to lateral femoral cutaneous nerve should be avoided. back	Fig. 7 A gouge is used to remove the posterior "roof" of the ilium and the cancellous bone between the cortical layers. back
Fig. 8 Cancellous bone is removed from the iliac tubercle in the anterior or posterior iliac spine region through a small cortical opening. back	Fig. 9 Wolfe and Kawamoto's technique of obtaining full-thickness bone graft from the anterior ilium. A sharp osteotome is used to make appropriate cuts shown above and in Fig.10. back
Fig. 10 The ridges of the iliac crest are split obliquely with the osteotome. The muscular and periosteal attachments should remain. back	Fig. 11 A large full-thickness bone graft can be removed as shown, using Wolfe and Kawamoto's method. back
Fig. 12 Wolfe and Kawamoto's technique of reapproximating the two fragments of the crest using wires or sutures. Figure-eight wire or suture may be passed through the bone with an awl and fixed to adjacent bone. back	Fig. 13 Illustration of the nerves that may be injured during the procedure to remove bone graft from the iliac crest. back
Fig. 15 A, The bony origin of the gluteus maximus or the roughened area anterior to the posterosuperior iliac crest is a good landmark and can be used as the caudal limit of bone removal. An imaginary plumb line dropped from the posterosuperior iliac spine with the patient in the prone position will pass through the bony rim of the sciatic notch. The superior gluteal artery is adjacent to the bony rim. B, A large amount of bone graft can be removed safely if the surgeon stays cephalad to the posterosuperior iliac spine, the sciatic notch, and the imaginary line joining them. back
Fig. 14 Illustration of the superior gluteal artery, curving around the rim of the sciatic notch as it leaves the pelvis. back	Fig. 16 Illustration of the anatomic positions of the arteries that may cause troublesome bleeding when working on the inner table of the ilium. back
Fig. 17 The anatomic locations of the peritoneal wall and the vulnerable arteries are shown relative to the iliacus muscle and the inner wall of the ilium. back	Fig. 18 It is very important to stay subperiosteally and carefully elevate the abdominal wall muscles off the crest and the iliacus muscle off the iliac wall before removing the bone from the inner table of the ilium. back
Fig. 19 The large defect left in the iliac wall may be repaired and the iliac contour restored using bone cement. Anchoring holes are made with a curette before the cement is applied. back	Fig. 20 The iliac wing defect is filled with bone cement. Malleable blades are used to repair the iliac wall as shown. back
Fig. 21 The iliac wall is reconstructed with bone cement. back

References

1. Albrektsson T: The healing of autologous bone grafts after varying degrees of surgical trauma, J Bone Joint Surg 62B:403, 1980. back
2. Andersson GBJ and others: Segmental replacement of the femur in baboons with fiber metal implants and autologous bone grafts of different particle size, Acta Orthop Scand 53:349, 1982. back
3. Bassett CAL: Clinical implications of cell function in bone grafting, Clin Orthop 87:45, 1972. back
4. Benum P and others: Porous ceramics as a bone substitute in the medial condyle of the tibia: an experimental study in sheep: long-term observations, Acta Orthop Scand 48:150, 1977. back
5. . Blakemore ME: Fractures at cancellous bone graft donor sites, Injury 14:519, 1983.
6. Bolander ME: Inducers of osteogenesis. In Bone and cartilage allografts: biology and clinical applications, Park Ridge, IL, 1991, American Academy of Orthopedic Surgeons, p 75. back
7. Bonfiglio M, Jetter WS: Immunological response to bone, Clin Orthop 87:19, 1972. back
8. Boskey AL, Posner AS: The role of synthetic and bone extracted Ca-phospholipid-P04 complexes in hydroxyapatite formation, Calcif Tissue Res 23:251, 1977. back
9. Bosworth DM: Repair of herniae through iliac crest defects, J Bone Joint Surg 37A:1069, 1955. back
10. Bovan-Salvars BD, Boskey AL: Relationship between proteolipids and calcium phospholipid phosphate complexes in calcification, Calcif Tissue Int 30:167, 1980. back
11. Bradford DSL: Anterior vascular pedicle bone grafting for the treatment of kyphosis, Spine 5:318, 1980. back
12. Bright R, Burstein A: Material properties of preserved cortical bone, Trans Orthop Res Soc 3:210, 1978. back
13. Brown KLB, Cruess RL: Bone and cartilage transplantation in orthopaedic surgery, J Bone joint Surg 64A:270, 1982. back
14. Brown LT: The mechanics of the lumbosacral and sacroiliac joints, J Bone joint Surg. 19:770, 1937. back
15. Brown MD and others: A roentgenographic evaluation of frozen allografts versus autografts in anterior cervical spine fusions, Clin Orthop 119:23,, 1976. back
16. Burchardt H: The biology of bone graft repair, Clin Orthop 174:28, 1983. back
17. Burchardt H and others: Freeze-dried allogenic segmental cortical-bone grafts in dogs, J Bone joint Surg 60A:1082, 1978.
18. Burwell RG: A study of homologous cancellous bone combined with autologous red marrow after transplantation to a muscular site, J Anat 95:613, 1961. back
19. Burwell RG: Studies in the transplantation of bone. V. the capacity of fresh and treated homografts of bone to evoke transplantation immunity, J Bone joint Surg 45B:386, 1963. back
20. Burwell RG: Studies in transplantation of bone. VIII. treated composite homo-autografts of cancellous bone, J Bone Joint Surg 48B:532, 1966. back
21. Burwell RG: The fate of bone grafts. In Apley AG, editor: Recent advances in orthopaedics, Baltimore, 1969, Williams & Wilkins. back
22. Burwell RG: The fate of freeze-dried bone allografts, Transplant Proc. (suppl 1) 8:95, 1976. back
23. Burwell RG: The function of bone marrow in the incorporation of a bone graft, Clin Orthop 200:125, 1985. back
24. Cameron HU: Evaluation of a biodegradable ceramic, J Biomed Mater Res 11:179, 1977. back
25. Challis JH and others: Strangulated lumbar hernia and volvulus following removal of iliac crest bone graft, Acta Orthop Scand 46:230, 1975. back
26. Chalmers J, Rush J: Observations on the induction of bone in soft tissues, J Bone Joint Surg 57B:36, 1975.
27. Cobey ML: A national bone bank survey, Clin Orthop 110:333, 1975.
28. Cockin J: Autologous bone grafting-complications at the donor site, J Bone Joint Surg 5311:153, 1971. back
29. Cooper JW: Cluneal nerve injury and chronic post surgical neuritis, J Bone Joint Surg 49A:199, 1967. back
30. Coventry MB, Tapper EM: Pelvic instability: a consequence of removing iliac bone for grafting, J Bone Joint Surg 54A:83, 1972. back
31. Cowley SP, Anderson LD: Brief note: hernias through donor sites for iliac-bone grafts, J Bone Joint Surg 65A:1023, 1983. back
32. Curtiss PhH and others: Immunological factors in homologous bone transplantation, J Bone Joint Surg 41A:1481, 1959.
33. Dawson E: The fate of bone substitution with porous hydroxyapatite implants in the dog spine. In Transactions of the 27th Annual Meeting, Orthopaedic Research Society Vol 6, 1981. back
34. deBoer HH: The history of bone grafts, Clin Orthop 226:292, 1988.
35. Doi K and others: Anterior cervical fusion using the free vascularized fibular graft, Spine 13:1239, 1988. back
36. Ducheyne P and others: Effect of hydroxyapatite impregnation on skeletal bonding of porous coated implants, J Biomed Mater Res 14:225, 1980. back
37. Dunn EJ: The role of methylmethacrylate in the stabilization and replacement of tumors of the cervical spine, Spine 2:15, 1977. back
38. Dupuis PR and others: Anterior free vascular transplant of the fibula for the treatment of kyphosis, J Bone joint Surg 64B:259, 1982. back
39. Drury BJ: Clinical evaluation of back and leg pain due to irritation of the superior cluneal nerve, J Bone Joint Surg 49A:199, 1967. back
40. Bone harvesting and transplantation, Lancet 2:730, 1981 (editorial).
41. Edmonson AS and others: Campbell's operative orthopaedics, ed 6, St. Louis, 1980, Mosby. back
42. Eftekhar NS, Thurston CW: Effect of irradiation on acrylic cement with special reference to fixation of pathological fractures, J Biomech 8:53, 1975.
43. Elves MW, Salama R: A study of the development of cytotoxic antibodies produced in recipients of xenografts (heterografts) of iliac bone, J Bone joint Surg 56B:331, 1974. back
44. Enneking WF and others: Autogenous cortical bone grafts in the reconstruction of segmental skeletal defects, J Bone Joint Surg 62A:1039, 1980. back
45. Escalas F, Dewald RL: Combined traumatic arteriovenous fistula and ureteral injury: a complication of iliac bone-grafting, a case report. J Bone joint Surg 59A:270, 1977. back
46. Evarts CM: Surgery of the musculoskeletal system, New York, 1983, Churchill Livingstone.
47. Fernyhough JC and others: Chronic donor site pain complicating bone graft harvesting from the posterior iliac crest for spinal fusion, Spine 17:1474, 1992. back
48. Frankel VH Burstein AH: Orthopedic biomechanics, Philadelphia, 1970, Lea & Febiger. back
49. Friedlander GE: Current concepts review - bone banking, J Bone Joint Surg 64A:307, 1982. back
50. Friedlander G, Huo M: Bone grafts and bone graft substitutes. In Frymoyer JW, editor: The adult spine: principles and practice. New York, 1991, Raven Press, p 565. back
51. Froimson AI, Cummings AG, Jr: Iliac hernia following hip arthrodesis, Clin Orthop 30:89, 1971. back
52. Gallie WE: The use of boiled bone in operative surgery, Am J Orthop Surg 16:373, 1918. back
53. Ghent WR: Further studies on meralgia paresthetica, Can Med Assoc J 85:871, 1961. back
54. Goldberg VM, Shaffer JW, Stevenson S: Biology of vascularized bone grafts. Bone and cartilage allografts: biology and clinical applications, Park Ridge, IL, 1991, American Academy of Orthopedic Surgeons, p 13. back
55. Gupta D and others: Bridging large bone defects with a xenograft composited with autologous bone marrow: an experimental study, Int Orthop (SICOT) 6:79, 1982. back
56. Harrington KD: The use of methylmethacrylate for vertebral body replacement and anterior stabilization of pathological fracture dislocations of the spine, due to metastatic disease, J Bone Joint Surg 63A:36, 1981. back
57. Hartman JR and others: A pedicle bone grafting procedure for failed lumbosacral spinal fusion, Clin Orthop 178:223, 1983. back
58. Heiple KG and others: A comparative study of the healing process following different types of bone transplantation, J Bone Joint Surg 45A:1593, 1963. back
59. Heppenstall RB: Fracture treatment and healing, Philadelphia, 1989, W.B. Saunders. back
60. Holmes RE: Bone regeneration within a coraline hydroxyapatite implant, Plast Reconstr Surg 63:626, 1979. back
61. Holmes R and others: A coralline hydroxyapatite bone graft substitute: preliminary report, Clin Orthop 188:252, 1984. back
62. Hsu K, Zucherman J, White A: Bone grafts and implants in spine surgery. In White AH, Rothman RH, Ray CD, editors: Lumbar spine surgery: techniques and complications, St. Louis, 1987, Mosby, p 343. back
63. Hughes CW: Rate of absorption and callus stimulating properties of cow horn, ivory, beef bone and autogenous bone, Surg Gynecol Obstet 76:665, 1943. back
64. Jackson JW: Surgical approaches to the anterior aspect of the spinal column, Ann R Coll Surg Engl 48:83, 1971. back
65. Jarcho M: Calcium phosphate ceramics as hard tissue prosthetics, Clin Orthop 157:259, 1981. back
66. Kahn BP: Superior gluteal artery laceration: a complication of iliac bone graft surgery, Clin Orthop 140:204, 1979. back
67. Kaneda K, Kurakami C, Minami A: Free vascularized fibular strut graft in the treatment of kyphosis, Spine 13:1273, 1988. back
68. Klawitter JJ, Hulbert SF: Application of porous ceramics for the attachment of load bearing orthopaedic applications, J Biomed Mater Res 2:161, 1971. back
69. Knight G: Paraspinal acrylic inlays in the treatment of cervical and lumbar spondylosis and other conditions, Lancet 2:147, 1959. back
70. Komender A: Influence of preservation on some mechanical properties of human haversian bone, Mater Med Pol 8:13, 1976. back
71. Kreuz FP and others: The preservation and clinical use of freeze-dried bone, J Bone Joint Surg 33A:297, 1974. back
72. Lane JM and others: Clinical application of biosynthetics - bone and cartilage allografts: biology and clinical applications, Park Ridge, IL, 1991, American Academy of Orthopedic Surgeons. back
73. Lane JM, Sandhu HS: Current approaches to experimental bone grafting, Orthop Clin North Am 18:213, 1987. back
74. Langer F and others: The immunogenicity of fresh and frozen allogeneic bone, J Bone joint Surg 57A:216, 1975. back
75. Leong JCY and others: The use of porous titanium mesh implant after discectomy in patients with deranged lumbar intervertebral disc-the five-year results of a prospective trial, Twelfth Annual Meeting of the International Society for the Study of the Lumbar Spine, 1985 (abstracts). back
76. Lichtblau S: Dislocation of the sacro-iliac joint: a complication of bone-grafting, J Bone Joint Surg 44A:193, 1962. back
77. Lloyd-Roberts GC: Experiences with boiled cadaveric bone, J Bone Joint Surg 34B:428, 1952. back
78. Lotem M and others: Lumbar hernia at an iliac bone graft donor site: a case report, Clin Orthop 80:130, 1971. back
79. Lubicky JP, Dewald RL: Methylmethacrylatae reconstruction of large iliac crest bone defect donor site, Clin Orthop 164:252, 1982. back
80. Luyten FP and others: Purification and partial amino acid sequence of osteogenin, a protein initiating bone differentiation, J Biol Chem 264:13370, 1989. back
81. Magnusson PB: Holding fractures with absorbable materials-ivory plates and screws, JAMA 61:1514, 1913. back
82. Malinin TI: University of Miami tissue bank: collection of postmortem tissues for clinical use and laboratory investigation, Transplant Proc (suppl) 8:53, 1976. back
83. Malinin TI: Cadaver bone allografts-bone banks. In Turek SL, editor: Orthopaedics, principles and their application, Philadelphia, 1984, Lippincott. back
84. Malinin TI, and Brown MD: Bone allograft in spinal surgery, Clin Orthop 154:168, 1981. back
85. Massey EW: Meralgia paresthetica secondary to trauma of bone graft, j Trauma 20:342, 1980. back
86. McMurray GN: The evaluation of Kiel bone in spinal fusions, J Bone joint Surg 64B:100, 1982. back
87. Mittelmeier JH and others: PMMA cement with carbon fiber reinforcement and apatite ingredients: mechanical properties and tissue reaction in animal tests, First World Biomaterials Congress, Bade, Austria, 1980. back
88. Mooney J, Derian C: Synthetic bone graft, lumbar spine surgery: techniques and complications. In White AH, Rothman RH, Ray CD, editors: Lumbar Spine Surgery Techniques and Complications. St. Louis, 1987, Mosby, p 471. back
89. Moore JB and others: A biomechanical comparison of vascularized and conventional autogenous bone grafts, Plast Reconstr Surg 73:382, 1984. back
90. Murray JA and others: Irradiation of polymethylmethacrylate: in vitro gamma radiation effect, J Bone joint Surg 56A:311, 1974. back
91. Nade S and others: Osteogenesis and bone marrow transplantation, the ability of ceramic materials to sustain osteogenesis from transplanted bone marrow cells: preliminary studies, Clin Orthop 181:255,1983.
92. Nade S, Burwell RG: Decalcified bone as a substrate for osteogenesis: an appraisal of the interrelation of bone and marrow in combined grafts, J Bone joint Surg 59B:189, 1977. back
93. Nawrot CJF and others: Dental phosphoproteininduced formation of hydroxylapatite during in vitro synthesis of amorphous calcium phosphate, Biochemistry 15:3445, 1976. back
94. Ohgushi H, Goldberg VM, Caplan AI: Heterotopic osteogenesis in porous ceramics induced by marrow cells, j Orthop Res 7:568, 1989. back
95. Oikarinen J, Korhonen LK: The bone inductive capacity of various bone transplanting materials used for treatment of experimental bone defects, Clin Orthop 140:208, 1979. back
96. Oldfield MD: Iliac hernia after bone grafting, Lancet 248:810, 1945. back
97. Osborn JF, Newesely H: Dynamic aspects of the implant-bone-interface, Dental Implants 111:123, 1980.
98. Osborn JF, and Newesely H: The material science of calcium phosphate ceramics, Biomaterials 1:108, 1980. back
99. Ostrowski K: Current problems of tissue banking, Transplant Proc 1:126, 1969. back
100. Ostrup LT: Distant transfer of a free living bone graft by microvascular anastomoses, and experimental study, Plast Reconstr Surg 54:274, 197J. back
101. Pappas AM: Current methods of freezing and freeze-drying, Cryobiology 4:358, 1968.
102. Pelker RR and others: Biomechanical properties of bone allografts, Clin Orthop 174:54, 1983. back
103. Pelker R and others: The effects of preservation on allograft strength, Trans Orthop Res Soc 7:283, 1982. back
104. Pierson AP and others: Bone grafting with Boplant: results in thirty-three cases, J Bone joint Surg 50B:364, 1968. back
105. Plenk H Jr, Hollmann K, Wilfert KH: Experimental bridging of osseous defects in rats by the implantation of Kiel bone containing fresh autologous marrow, J Bone joint Surg 54B:735, 1972. back
106. Pyrtek LJ, Kelly CC: Management of herniation through large iliac bone defects, Ann Surg 152:998, 1960. back
107. Ray RD: Vascularization of bone grafts and implants, Clin Orthop 87:43, 1972. back
108. Ray RD, Holloway JA: Bone implants, J Bone joint Surg 39A:1119, 1957.
109. Reid RL: Hernia through an iliac bone-graft donor site: a case report, J Bone joint Surg 50A:757, 1968. back
110. Reynold CF and others: Clinical evaluation of the merthiolate bone bank and homogenous bone grafts, J Bone joint Surg 33A:873, 1951. back
111. Rhinelander FW: Tibia] blood supply in relation to fracture healing, Clin Orthop 105:34, 1974.
112. Salama R: Xenogeneic bone grafting in humans, Clin Orthop 174:113, 1983. back
113. Salama R, Weissman SL: The clinical use of combined xenografts of bone and autologous red marrow: a preliminary report, J Bone joint Surg 60B:111, 1978. back
114. Salama R and others: Recombined grafts of bone and marrow, J Bone joint Surg 55B:402, 1973. back
115. Sampath TK, Coughlin JE, Whetstone RM and others: Bovine osteogenic protein is composed of dimers of OP-1 and BMP 2A, two members of the transforming growth factor Beta superfamily, J Biol Chem 265:13198, 1990. back
116. Scoville WB and others: The use of acrylic plastic for vertebral replacement or fixation in metastatic disease of the spine, j Neurosurg 27:274, 1967. back
117. Sedlin E: A rheologic model for cortical bone, Acta Orthop Scand (suppl) 36:83, 1965. back
118. Seres JL: Fusion in the presence of severe metastatic destruction of the cervical spine (case report), j Neurosurg 28:592, 1968.
119. Shaffer JW and others: The superiority of vascularized compared to nonvascularized rib grafts in spine surgery shown by biological and physical methods, Spine 13:1150, 1988.
120. Smith RT: The mechanism of graft rejection, Clin Orthop 87:15, 1972. back
121. Smith SE and others: Ilioinguinal neuralgia following iliac bone-grafting, J Bone joint Surg 66A:1306, 1984. back
122. Spence WT: Internal plastic splint for stabilization of the spine, Clin Orthop 92:325, 1973.
123. Steffee A, Biscup R, Sitkowski D: Segmental spine plates with pedicle screw fixation: a new internal fixation device for disorders of the lumbar and thoracic spine, Clin Orthop 203:45, 1986. back
124. Steffee A, Sitkowski D, Topham L: Total vertebral body and pedicle replacement, Clin Orthop 203:203, 1986. back
125. Stringa G: Studies on the vascularization of bone grafts, J Bone joint Surg 39B:395, 1957. back
126. Summers BN, Eisenstein SM: Donor site pain from the ilium, a complication of lumbar spine fusion, J Bone joint Surg 71B:677, 1989. back
127. Taylor GI and others: The free vascularized bone graft: a clinical extension of microvascular techniques, Plast Reconstr Surg 64:745, 1979. back
128. Termine JD, Kleinman HK, Whitson SWK and others: Osteonectin: a bone-specific protein linking mineral to collagen, Cell 26:99, 1981. back
129. Triantafyllou N and others: The mechanical properties of the lyophilized and irradiated bone grafts, Acta Orthop Belg 41:35, 1975. back
130. Turek SL: Orthopaedics, principles, and their application, Philadelphia, 1984, Lippincott. back
131. Urist MR: Bone: formation by autoinduction, Science 150:893, 1965. back
132. Urist MR: Practical applications of bone research on bonegraft physiology. In Instructional course lectures, American Academy of Orthopaedic Surgeons, vol 25, St. Louis, 1976, Mosby. back
133. Urist MR: Surface-decalcified allogenic bone implants, Clin Orthop 56:37, 1968. back
134. Urist MR, Dawson E: Intertransverse process fusion with the aid of chemosterilized autolyzed antigen-extracted allogeneic (AAA) bone, Clin Orthop 154:97, 1981. back
135. Urist MR, Nogami H: Morphogenetic substratum for differentiation cartilage in tissue culture, Nature 225:1051, 1970. back
136. Urist MR, Strates BJS: Bone morphogenetic protein, J Dent Res 50:1392, 1971. back
137. Urist MR and others: B-tricalcium phosphate delivery system for bone morphogenetic protein, Clin Orthop 187:277, 1984. back
138. Urist MR and others: A chemosterilized antigenextracted autodigested allo-implant for bone banks, Arch Surg 110:416, 1975. back
139. Urist MR and others: Human bone morphorgenic protein (BMP), Proc Soc Exp Biol Med 173(2):194, 1983. back
140. Veis A: The role of acidic proteins ire biological mineralization: ions in macromolecular and biological systems. In Everett DH, Vincent B, editors: Colston paper 29, Bristol, England, 1978, Society Technica, p 259. back
141. Verheugen P and others: Hernie illique apres prelevement osseux: subobstruction, Acta Chir Belg 1051:1056, 1965. back
142. Wang EA, Rosen J, D'Alessandro JS and others: Recombinant human bone morphogenetic protein induces bone formation, Proc Natl Acad Sci USA, 87:2220, 1990. back
143. Watson-Jones R: Transplantation of bone. In Fracture and joint injuries, vol 1, ed 4, Baltimore, 1955, Williams & Wilkins. back
144. Weikel AM, Habal MB: Meralgia paresthetica: a complication of iliac bone procurement, Plast Reconstr Surg 60:572, 1977. back
145. Weiland AJ: Current concept review: vascularized free bone transplants, J Bone joint Surg 63A:166, 1981. back
146. Weiland AJ, Phillips TW, Randolph MA: Bone grafts: a radiologic, histologic, and biomechanical model comparing autografts, allografts, and free vascularized bone grafts, Plast Reconstr Surg 74:368,1984 back
147. Weiss RE, Reddi AH: Role of fibronectin in collagaenous matrix induced mesenchymal cell proliferation and differentiation in vivo, Exp Cell Res 133:243, 1981. back
148. Werntz J, Lane JM, Piez C and others: The repair of segmental bone defects with collagen and marrow, Orthop Trans 10:346, 1986. back
149. White AA 111, Panjabi MM: Clinical biomechanics of the spine, Philadelphia, 1978, Lippincott. back
150. White AA III and others: Spinal stability: evaluation and treatment, Instructional course lectures. The American Academy of Orthopaedic Surgeons, vol 30, St. Louis, 1981, Mosby.
151. Wilde AH, Greenwald AS: Shear strength of self curing acrylic cement, Clin Orthop 106:126, 1975. back
152. Williams G: Experiences with boiled cadaveric cancellous bone for fractures of long bones, J Bone joint Surg 46B:398, 1964. back
153. Wolfe SA, Kawamoto HK: Taking the iliac-bone graft: a new technique, J Bone Joint Sur 60A:411, 1978. back
154. Wozney JM, Rosen J, Celeste AJ and others: Novel regulators of bone formation: molecular clones and activities, Science 242:1528, 1988. back
155. Younger EM, Chapman MW: Morbidity at bone graft donor sites, J Orthop Trauma 3:192, 1989. back
156. Zucherman JF, Brack S, Hsu ICY and others: Radiographic comparison of autogenous and freeze-dried ethylene oxide sterilized demineralized corticocancellous allograft bone in posterolateral spine fusion-a prospective study (Submitted for publication). back