Patient positioning is critical, but often overlooked, when performing the LLIF procedure for deformity correction. Prior to placing the patient on the operating room (OR) table, an extension piece can be placed at the foot of the table. Once the extension piece is in place, the OR table can be reversed, or turned around, so that the patient’s head is placed on the extension located at the true end of the bed. The extension maximizes OR table length and gives the surgeon more room caudally. The bed can then be maximally translated away from the central metal bedpost (toward anesthesia). The placement of the extension piece, reversing and translating the OR bed, can result in more room under the table caudally and less restriction for the fluoroscopic C-arm relative to the OR table bedpost. This can result in better visualization of the caudal lumbar vertebra and pertinent bony anatomic landmarks (Fig. 20.9).

Patients should be positioned in the lateral decubitus position with their iliac crest at the break of the table with all bony prominences well padded. A true lateral decubitus position is essential to this procedure. This position will allow for the abdominal contents to fall forward and away from the psoas more easily during peritoneal release from the retroperitoneal space, which can decrease the risk of injury to the peritoneum and its contents. Placing the patient’s iliac crest at the break of the OR table allows for flexing (breaking) the table, which can allow better visualization and access to the caudal lumbar levels, particularly L4–L5. It may also aid in the correction of the coronal deformity if the surgical approach is through the concavity of the lumbar curve. Most surgeons stand at the patient’s back when performing LLIF surgery, so the patient’s back should be positioned close to the posterior edge of OR table so that the surgeon does not have to lean considerably over the patient to visualize the operative field. Care has to be taken not to position the patient too far posterior so that the spine overlaps the metal bars on the side of the OR table. This can be particularly problematic when the bed has to be rotated toward the patient’s back due to the scoliosis in order to obtain neutral lateral x-rays. Rotating the patient posteriorly will rotate the metal bar on the side of the bed under the patient, possibly blocking C-arm visualization (Fig. 20.10).

The patient’s hips and knees should be flexed to approximately 60 and 90°, respectively, which takes tension off the psoas muscle and more importantly the lumbar plexus that lies within it. Femoral nerve strain at L4–L5 can increase with breaking the OR table by putting the psoas muscle and lumbar plexus on stretch.

O’Brien and colleagues demonstrated in a cadaveric model that table flexion results in preloading the femoral nerve when approaching L4–L5 [46]. With 40° of table flexion at the pelvis, there was anterior displacement of the nerve by approximately 1.5 mm resulting in the highest nerve strain (average, 6–7 %) compared with 0°. Strain in the femoral nerve decreased with increasing hip flexion for both table flexion angles (40 and 0°). Flexing the hips and knees relaxes the psoas muscle and nerves minimizing the effects of breaking the table when required. Additionally, flexing the hips and knees may permit increased mobilization of the psoas and lumbar plexus allowing more displacement when the retractor is placed and opened, decreasing stretch and possibly neurological injury.

An axillary roll should be placed to prevent brachial plexus injury. Padding should be placed between the OR table and down leg to protect the peroneal nerve and the skin from breakdown over the bony prominences. A pillow should be placed between the patient’s legs. Additionally, an arm board should be placed to support the patient’s down arm. An arm holder or a pillow can be placed to support the patient’s up arm. Once the patient is provisionally positioned, tape should be applied across the patient’s upper chest and lower hips to secure them to the OR table outside of the planned operative field. This prevents the patient from shifting or rotating intraoperatively. Tape should also be applied to the patient’s thighs and legs to secure them in flexion (Fig. 20.9). Tape can be applied multiple times, even circumferentially around the patient and bed to prevent patient migration during the procedure.

Once the patient is secured to the OR table, the table can be flexed at the iliac crest for optimal visualization of L4–L5. The bed should be extended at the patient’s feet to prevent the hips and knees from extending. The bed can be placed in reverse Trendelenburg to compensate for the breaking in the table to level the patient’s torso.

Given the segmental deformities often seen in scoliosis, the OR table and fluoroscopy often need to be adjusted at each level to ensure optimal radiographic imaging (Fig. 20.11). It is the authors’ recommendation that the bed, not the C-arm, should be adjusted to obtain true AP and lateral x-rays of each individual operative level. With the C-arm locked at 0°, the bed can be rotated until a true AP image is obtained, and the spinous process of each level is in perfect midline position between the two pedicles and the endplates are parallel (Fig. 20.12). With the C-arm locked at 90°, the bed can be inclined or declined (Trendelenburg vs. reverse Trendelenburg) until a true lateral image is obtained with parallel endplates and overlapping pedicles and facet joints (Fig. 20.13). Moving the table and not the C-arm allows the surgeon an easier reference point of a true AP (surgeon’s hand perfectly horizontal and parallel to the floor) and lateral (surgeon’s hand perfectly vertical and perpendicular to the floor). True lateral x-rays can help prevent endplate violation during disk prep and implant placement, which can eliminate the advantage of deformity correction and indirect decompression of the LLIF technique. More worrisome, without true AP x-rays, vertebral rotation can lead to implant encroachment posteriorly in the neuroforamen and nerve injury, or anteriorly resulting in catastrophic vascular injury.

After draping the operative field with 10 × 10 drapes, the skin can be prepped with alcohol. Prior to the incision, each planned LLIF surgical level should be visualized with a true AP and lateral x-ray to ensure all planned operative levels can be visualized and accessed. The bony ribs can be palpated and marked on the skin. A flexible radiopaque guide wire can be used to identify the disks of each planned operative level, along with outlining the iliac crest on the skin to identify any bony restrictions, particularly the iliac crest restricting access to L4–L5 (Fig. 20.14).

Tape can be used to mark the floor relative to the position of the C-arm, which allows the C-arm machine to be moved in and out of the operative field with precision (Fig. 20.15). Additionally, the position of the C-arm gantry identifying true AP and lateral x-rays for each planned operative level can be marked with tape and the levels labeled (Fig. 20.16). This preoperative marking of the skin can minimize the skin incision, and labeling the floor and C-arm gantry with tape can expedite the intraoperative localization process of each planned surgical level and provide a reproducible guide for other radiology technologists who may not have been present at the beginning of the case.

Access to the lateral lumbar spine can be performed using a one- or two-incision technique. Regardless of the technique, the skin incision over the planned operative level is made first, exposing the subcutaneous fat, which is dissected and retracted out of the operative field. The external oblique (EO) muscle is the first muscle encountered, and its muscle fibers run obliquely toward the umbilicus. This muscle layer can be bluntly dissected to expose the internal oblique (IO) muscle, whose muscle fibers run perpendicular to the EO muscle. The IO can be bluntly dissected over each individual lumbar disk level to reveal the transverse abdominal muscle or it can be cut. Care must be taken to bluntly dissect the IO muscle first to avoid injuring the subcostal, iliohypogastric, and ilioinguinal nerves, which originate from the T12 and L1 nerve roots and course anteriorly and inferiorly and pierce and innervate the abdominal wall musculature. Injury to any of these nerves can result in abdominal wall paresis causing an abdominal bulge or droop, which can be permanent or temporary depending on the degree of nerve injury (Fig. 20.17). Cahill and colleagues reported a 4.2 % incidence of postoperative abdominal wall bulge, which they related to permanent injury to the motor nervous supply to the lateral abdominal wall [47]. Once the IO muscle is dissected, the transverse abdominal (TA) muscle is visualized and dissected. As its name contends, the TA fibers run transversely or horizontally. Immediately deep to the TA muscle is the transversalis fascia. This aponeurotic membrane can be distinct and lies between the inner surface of the TA muscle and the parietal peritoneum. Once this is dissected, the retroperitoneal fat can be visualized.

Access can also be obtained through a two-incision technique. After the first incision located directly over the planned surgical disk space is performed, dissection is carried to the transversalis fascia as described above. A second small incision can then be placed a finger-length posterior to the first incision and can be used as a direct access point to the retroperitoneal space. Once the second incision is made, blunt muscle dissection is performed to the transversalis fascia posteriorly, which can be penetrated with a pointed clamp into the retroperitoneal space (Fig. 20.18). Through the second more posterior incision, blunt finger dissection can sweep the peritoneum, fat, and any adhesions off the transversalis fascia and psoas muscle. The transverse process is a bony landmark that can be palpable for localization confirming the position in the retroperitoneal space. Once blunt dissection is performed, the same finger can be directed to the first incision and used as a visual marker for dissection of the transversalis fascia ensuring entry into the retroperitoneal space. The initial dilator can then be placed through the first incision using the finger in the second incision to guide the dilator to the retroperitoneal space and psoas muscle.

The retroperitoneal fat can be swept anteriorly to visualize the fascia enveloping the psoas muscle. The genitofemoral (GF) nerve lies directly on top of the psoas fascia. The GF nerve originates from the upper L1 to L2 segments of the lumbar plexus and passes caudally and emerges from the anterior surface of the psoas muscle. It can be encountered in the center on the LLIF operative field commonly at L3–L4. The nerve continues downward and divides into two branches, the genital branch and the femoral branch. The genital branch passes through the deep inguinal ring and enters the inguinal canal. The genital branch continues down and supplies the scrotal skin in men and accompanies the round ligament of the uterus terminating in the skin of the mons pubis and labia majora in women. The femoral branch passes underneath the inguinal ligament, traveling adjacent to the external iliac artery, supplying the skin of the upper anterior thigh and groin. Whether each operative level is approached with individual small dissections through the muscle layers and fascia or a mini-open incision is performed for better visualization, the psoas fascia should be released to expose the muscle fibers beneath it. Using a “no-look” approach, moving the initial dilator 2–3 mm cranially and caudally, in-line with the muscle fibers, can release the fascia as the dilator is advanced into the psoas muscle and docked (Fig. 20.19). This can decrease the risk of dragging the fascia and GF nerve into the psoas causing stretch with dilator advancement or compression with retractor opening, both of which can cause injury and anterior thigh symptoms. Wanding the initial dilator more than 2–3 mm cranially and caudally could result in nerve injury.

Sequential dilation of the psoas muscle can then be performed for each planned operative level using biplanar fluoroscopy and directional electromyography (EMG) or mechanomyography (MMG) neuromonitoring. The initial dilator is placed through the psoas and docked at the anterior 1/3 of the disk space where it is secured with a guide wire and confirmed with lateral x-ray. Sequential dilation of the psoas muscle is performed followed by placement of the retractor, which is fixed to the operating table. Each dilator should be monitored with directional EMG or MMG. Neuromonitoring alerts using triggered EMG or MMG can allow repositioning of the dilators and retractor to avoid nerve injury. Most nerve injuries are associated with instrumentation at the L4–L5 level [48–52].

Surgeons can start at increasing thresholds of 15–20 milliamps (mA), which may provide a safer transpsoas working corridor. If a response is obtained, the mA can be lowered until a response is not elicited. Feedback can be provided visually from the patient with muscle jerk of the extremity and from the neuromonitoring technologist, with thresholds below 5 mA usually indicating direct nerve contact [53]. Threshold responses between 5 and 10 mA indicate close proximity, and responses greater than 10 mA usually are indicative of a safe working distance away from the motor nerves [54, 55].

Prior to placing the retractor, the finger of a sterile latex glove can be cut off and placed over the retractor, which can allow retractor opening while preventing retroperitoneal fat and psoas muscle from entering the operative field. Additionally, angling the retractor fixation arm vertically, instead of horizontally, can apply additional downward pressure and stability and may eliminate the need for a bone fixation pin or shim in the disk (Fig. 20.20). This can prevent retractor migration and creep of the psoas muscle under the retractor blades. Vertebral body blade fixation pins can result in unrecognized bony bleeding or segmental artery injury, and disk shims fix the posterior retractor blade, which is in close proximity to the lumbar plexus risking injury. AP and lateral images are then used to confirm position of the working corridor.

Once the retractor is placed and a safe working corridor is established and confirmed with x-ray, the discectomy can be performed. Minimize retractor opening to prevent stretch of the lumbar plexus or injury to the vertebral segmental vessels located at the midportion of the vertebral body (Fig. 20.21).

The retractor is opened just enough for adequate visualization and cage placement, the limits of which are usually just cranial and caudal to the endplates. Additionally, be conscious that excessive posterior blade retraction can compress the nerve between the retractor blade and the transverse process. Work quickly to decrease the risk of compressive or ischemic injury to the nerve due to prolonged psoas retraction. In a prospective multicenter trial, Uribe and colleagues evaluated whether triggered EMG monitoring could predict postoperative symptomatic neuropraxia (SN) throughout the retraction process during LLIF procedures. Postoperatively, 13 of 323 (4.03 %) patients had a new motor weakness that was consistent with SN of the lumbar plexus on the approach side. Retraction time was significantly longer in those patients with SN versus those without (32.3 vs. 22.6 min, p = 0.031) [56]. Chaudhary et al. observed diminished motor evoked potentials (MEPs), not at the time of initial retractor placement, but after prolonged retractor opening in patients with postoperative motor nerve injuries. Prolonged mechanical compression and stretch were felt to be the mechanisms of injury [57]. This proposition is supported by studies that have reported a higher likelihood of nerve injury with increasing surgical times [31].

Even with meticulous conscientious dilator and retractor placement in relation to the lumbar plexus, nerve and psoas muscle irritation is a potential side effect even in single-level LLIF procedures. Suggestions to reduce the risk of nerve injury include preoperative administration of gabapentin or Lyrica, along with 10 mg IV dexamethasone may prophylactically combat the inflammatory cascade in nervous and muscular tissue limiting the extent of injury [58, 59]. In cases where extended psoas retraction is required, such as in multilevel deformity cases, releasing the retractor and allowing for the muscle and soft tissue to relax may decrease the likelihood of such injury. Lastly, shallow docking above the psoas muscle allows dissection of the psoas with a Penfield dissector allowing direct visualization of lumbar plexus prior to dilator placement [60].

As with most minimally invasive techniques, there is a learning curve to overcome. We have found, as have others, that the risk of nerve injury declines steadily with greater experience [61]. Le et al. reported a significant reduction in the incidence of postoperative numbness of nearly 60 % (26.1–10.7 %), with their refined technique over a 3-year period [62]. Experience and evolution of a surgeon’s technique can minimize the risk of iatrogenic nerve injury resulting in dramatic changes in patient outcome.

With the retractor in place and opened, the dilators are removed leaving the guide wire in place. The guide wire can provide a reference point for the annulotomy. Using an EMG or MMG probe, the annulus and remaining strands of psoas muscle around the guide wire and edges of the retractor blades are stimulated to ensure there are no nerves crossing the operative field. A bipolar cautery and Penfiled dissector can be used to facilitate removing the remaining strands of muscle providing clear identification of the annulus. The annulus can then be incised vertically with a scalpel both dorsal and ventral to the guide wire delineating the AP working space for cage insertion (Fig. 20.22).

Annulotomy size is based on the selection of the width of the LLIF cage. Cage widths range from 18 to 27 mm, but more common widths include 18, 21, and 22 mm. Marchi et al. reviewed the incidence and effect of subsidence in patients with stand-alone short-segment 1- or 2-level lateral lumbar interbody fusions with two different width cages, 22 and 18 mm [63]. Patients who had wider 22 mm cages had greater lordosis correction, along with lower rates and grades of subsidence. At 12 months, 70 % in the standard group (18 mm) and 89 % in the wide group (22 mm) had Grade 0 or I subsidence, and 30 % in the standard group and 11 % in wide group had Grade II or III subsidence (Fig. 20.23). Subsidence was detected early at 6 weeks postoperatively and correlated with transient clinical worsening in VAS scores. Progression of subsidence was not observed after the 6-week time point. Additionally, subsidence occurred predominantly (68 %) at the inferior endplate. Although fusion rate was not affected by cage dimension (p > 0.999) or by the incidence of subsidence (p = 0.383), most patients requiring secondary revision spinal procedures experienced Grade II and III subsidence (six of ten patients). This illustrates the importance of not violating the endplates and limiting subsidence.