Chronic harvest site pain is the most frequently reported complication in most studies. Reported rates of chronic pain vary greatly in the literature, depending on study design and study population. In a study by Goulet et al., 30 of 71 (42.3%) patients undergoing spinal surgery reported persistent donor site pain at 6 months compared to 3 of 16 (18.7%) undergoing nonspinal surgery (16). At 2 years, these figures had decreased to 15 of 71 (21.1%) and 1 of 16 (6.2%), respectively. In studies of selected patients undergoing maxillofacial reconstruction, rates are generally low. Laurie et al. reported a 10% rate of moderate pain with exercise at 2 years in this group (19). Rates are generally higher in studies specifically addressing populations of spine surgery patients, but even within this population there is a wide range of reported rates, depending largely on study design. Based on chart review data, Younger and Chapman reported a minimal 2.5% rate of persistent pain at 6 months (14). At the other extreme, Summers and Eisenstein reported a 49% rate of chronic donor site pain (25% “severe” and 24% “acceptable”) in a series of 290 patients undergoing posterior lumbar spinal fusion (10). Frymoyer et al. reported a rate of 37% in patients 10 years after lumbar spinal fusion surgery, and De Palma et al. reported persistent discomfort in 36% of patients 1 year following anterior graft harvest for cervical spine fusion (20,21).

It has been suggested that the higher reported rates in spine surgical patients may reflect the difficulty in distinguishing persistent back pain from harvest site pain. This is supported by the finding that the underlying diagnosis appears to be a factor in determining the likelihood of persistent harvest site pain. Fernyhough et al. studied 151 spinal fusion patients and found that the rate of chronic harvest site pain in patients undergoing surgery for back pain and degenerative spinal disease was twice as high as in patients undergoing surgical stabilization for acute spinal trauma (22).

Certain aspects of surgical technique appear to play a significant role in the incidence of harvest site pain. The rate and severity of postoperative pain appears to correlate with the extent of soft-tissue dissection. Use of percutaneous techniques with large-bore needles or trephines has been associated with less pain but provides limited quantities of bone graft that are inadequate for most spinal fusion procedures (23,24).

Other aspects of harvest technique may be a significant factor, but this has been difficult to prove. It is unclear whether there is a difference in rates between anterior and posterior harvest site pain (11,18,25). Ahlmann et al. retrospectively compared morbidity between anterior and posterior graft harvest in 88 patients treated for chronic osteomyelitis, and identified higher rates in association with anterior graft harvest (3). In this study population, more severe pain and greater duration of pain were reported following anterior graft harvest. When utilizing the posterior midline lumbar incision for graft harvest, a fascial splitting approach has been associated with a lower complication rate compared with a subcutaneous approach (3% versus 15%), but there is no clear difference in pain rates with harvest through the primary surgical incision as opposed to a separate incision (22,26). Similarly, there is no difference in pain following anterior harvest of corticocancellous bone from the inner table of the ilium as opposed to the outer table (27,28).

To minimize the incidence of chronic harvest site pain, we recommend placing the incision off the most prominent aspect of the iliac crest to minimize postoperative scar sensitivity. The planned incision should be marked preoperatively along with relevant regional landmarks and the skin mobilized to ensure that the graft harvest site is readily accessible. When possible, the iliac crest contour should be preserved, either by harvesting corticocancellous strips from below the crest, or by reflecting a hinged osteoperitoneal cap off the top of the crest and harvesting cancellous bone from between the iliac tables. Anteriorly, an incision 1 to 2 cm medial to the crest will minimize irritation from patient clothing, such as waistbands and belts (29). Rounding off the corners of the remaining crest with a high-speed burr may also decrease the incidence of donor site pain (30).

Local wound infiltration with anesthetic such as bupivacaine can decrease postoperative pain but is only useful for the first 24 hours following surgery (31). For longer periods of anesthetic administration, an optional technique is a percutaneous catheter allowing intermittent or continuous infusion of bupivacaine (32). In order to completely avoid the risk of chronic donor site pain, alternative sources for bone graft harvest have been evaluated, including local graft (33).

Significant blood loss is a potential complication with either anterior or posterior bone graft harvest and is partly related to technique and quantity of bone harvested. Even without direct vascular injury, substantial blood loss can occur from exposed cancellous bone surfaces. If inadequately controlled, persistent bleeding following wound closure can result in sizeable hematoma formation with significant local pain. Overall, rates of 1% to 10% have been reported (21,34,35). Following anterior harvest, blood can track retroperitoneally or subfascially down the anterolateral thigh. Patients may report an expanding mass and local pain and tenderness. The risk appears to be less following posterior graft harvest due to natural compression with supine positioning (36).

Avoidance can be accomplished through meticulous dissection and hemostasis. Generous application of topical agents such as microcrystalline or microfibrillar collagen or thrombin-impregnated collagen sheets appears most effective
when dealing with broad areas of diffusely bleeding cancellous bone. If platelet-rich plasma is being utilized to supplement bone graft, platelet-poor plasma can be introduced into the harvest site (32). For localized bone bleeding, such as with anterior tricortical graft harvest, bone wax is ideal (37). The use of suction drainage varies among surgeons, with no clear support in the literature for or against its use (11,14). When used, drains should be removed within 48 hours to minimize infection risk. Application of a pressure dressing over anterior harvest sites using fluffed gauze and elastic tape is commonly performed by plastic and oral maxillofacial surgeons and may reduce the incidence of hematoma formation (12).

Hematoma occurrence has been associated with increased pain and infection risk (26,38,39). Large hematomas should be considered for aspiration using sterile technique followed by application of a pressure dressing. With recurrence, percutaneous placement of a catheter and continuous suction can be considered (32).

One prospective, randomized study compared morbidity rates for two different posterior graft harvest techniques (2). Subperiosteal exposure of the outer table, followed by harvest of corticocancellous strips with osteotomes and gouges, was compared with intraosseous cancellous bone harvest through a cortical window, leaving the inner and outer tables intact. No difference was found in blood loss or pain, but the intraosseous method provided consistently less bone graft and prolonged operative time. Overall, one-fifth of patients reported persistent moderate to severe pain at 2 years. However, 56% of patients reported the same degree of persistent low back pain, leading the authors to speculate that donor site pain in their study population was “part of a general pain syndrome with high association between back pain, leg pain, and donor site pain.”

The overall rate of postoperative wound infection following graft harvest is less than 1% (38,40). Several studies have suggested a higher risk with posterior harvest sites. To minimize risk, preoperative antibiotics should be administered approximately an hour prior to skin incision. Cephazolin, a broad-spectrum, first-generation cephalosporin, is ideal due to its long half-life and effective antistaphylococcal coverage. Cephazolin is inconsistently effective against Gramnegative bacilli, and additional antibiotic coverage should be considered if Gram-negative septicemia is a concern. For long procedures, interval redosing, typically every 4 hours for cephazolin, should be performed.

When grafting to a potentially infected tissue bed that has been debrided prior to graft harvest, entirely separate sterile instruments — and possibly draping — should be utilized. Rigorous hemostasis to avoid local hematoma formation will also reduce the likelihood of local infection. Thorough saline irrigation prior to wound closure should be performed.

Disruption of normal gait patterns can occur following either anterior or posterior graft harvest. So-called gluteal gait abnormality has been reported in approximately 3% of patients (41,42,43). Excessive stripping or inadequate repair of the gluteus medius attachment to the iliac crest has been blamed following anterior graft harvest. Affected patients may demonstrate hip abductor weakness and an abductor lurch. The tensor fascia lata functions as a lower extremity stabilizer and hip flexor, and inflammation or weakness of this muscle may result from incomplete reattachment (44). Following posterior graft harvest, patients may report weakness with stair-climbing or rising from a seated position (18). Goulet et al. reported a 12.6% rate of ambulation difficulty at 6 months in a group of predominantly spinal surgery patients undergoing posterior graft harvest. This figure declined to 5.7% at 2 years (16).

Injury to the lateral femoral cutaneous nerve (LFCN) results in meralgia paresthetica, characterized by hypesthesia, anesthesia, or dysesthesia along the anterolateral aspect of the proximal to middle thigh. Two major injury mechanisms have been suggested during anterior iliac crest bone graft harvest: (a) direct laceration or electrocautery of an anatomic variant nerve overlying the anterior superior iliac spine; or (b) neurapraxia from excessive tension during retraction of the iliacus muscles while exposing the inner table of the ilium. Again, the literature reports widely varying rates of this complication (from 4.5% to 37%), depending largely on study design (3,4,9,14,16,18,22,45).

Multiple anatomic variations in the course of the lateral femoral cutaneous nerve make it vulnerable to iatrogenic injury during the course of anterior iliac crest bone graft harvest. Cadaveric dissections have revealed a 30% incidence of significant anatomic variation (46). The LFCN is a purely sensory nerve that arises from various contributions of the L1, L2, and L3 nerve roots. After emerging from the lateral psoas, the LFCN travels retroperitoneally subjacent to the iliac fascia and above the iliacus toward the junction of the inguinal ligament and the anterior superior iliac spine. Anatomic relationships of the nerve as it passes from the pelvis into the thigh are highly variable. Aszmann et al. reported 5 anatomic variations in 52 cadaveric dissections and devised a classification system (Fig. 16.1) (47). Type A (4%) nerves cross the iliac crest; type B (27%) pierce the inguinal ligament; type C (23%) travel beneath the inguinal ligament and through the sartorius tendon; type D (26%) travel beneath the inguinal ligament medial to the sartorius origin; and type E (20%) travel medially over the
iliopsoas. Other studies have demonstrated variable location with respect to distance from the anterior superior iliacspine, ranging from 6 cm medial to 2 cm lateral (48). Overall, it has been estimated that 9.9% of nerves are anatomically vulnerable traveling over or near the anterior iliac crest (49). Due to the angulated course of the LFCN across the inguinal region, extreme passive hip extension increases traction on the nerve and should be avoided if possible (50).