Once the patient is in the prone position, the pathologic level is identified by fluoroscopy. Entry points for a percutaneous pedicle screws insertion are marked on the skin on both side of the spinous process. Small skin incisions are made on the previously identified entry points and carried down to the fascia. With the aid of fluoroscopic vision, pedicles are identified, K-wire is positioned, and cannulated screws are inserted through the pedicle in the vertebral body. Then the screws are attached to the tubular extensions, the K-wire is removed, and a further K-wire is placed in the facet joint ipsilaterally to the side where the cage will be inserted (Fig. 8.2). Sequential dilators were passed over the K-wire until a tubular retractor can be place in site and expanded to provide a 2.5–4.0 cm operative field and a pedicle-to-pedicle exposure. The facet joint and the lateral aspect of the lamina are removed, the corresponding nerve root is identified, and the disc is approached. Once discectomy is performed, the interbody space is widened by screw distraction, and end plates are freed from the cartilaginous tissue using curettes and end plate scrapers. The distraction of the interbody space is an important aspect of the procedure because it allows an adequate nerve root decompression and facilitates cage insertion as well as spondylolisthesis reduction. Once the end plates are properly prepared, the autogenous bone obtained from the resected lamina and facet is mixed with demineralized bone matrix and placed within the interbody space anteriorly and contralaterally to the annulotomy. The cage is then inserted under fluoroscopic vision and placed anteriorly in the disc space where the bone is stronger and the risk of end plate rupture is lower. Rods are positioned bilaterally and tightened to the screws applying a compression to avoid cage displacement. Final fluoroscopic control is done, and then wounds are sutured. The day after the procedure, the patient is able to stand up, and a lumbar X-ray is performed in the orthostatic position.

The mean operative time was 168.3 and 227.6 min for single and multilevel surgery. No patients needed blood transfusions. Mean hospital stay was 4.2 days.

The mean preoperative ODI and VAS scores were 55.6 and 7.3. Follow-up period ranged from 2 to 6 years (mean, 4.1 years). Pain improvement at 1 and 2 years follow-up was observed, respectively, in 175 and 184 of the 221 patients (79.1% and 83.2%) with a mean ODI and VAS reduction of 14.7 and 3.5 at 1 year and 13.9 and 3.1 at 2 years.

Fusion rate was 75.1% at 1 year (166 of 221 patients) and 84.1% at 2 years (186 of 221 patients). Pseudoarthrosis (grades III and IV) was observed in 34 patients (15.3%), but only four patients require a supplemental surgery to repair the nonunion.

Twenty-one of the 221 patients (9.5%) had perioperative adverse events (from the day of surgery to 12 weeks after) consisting of dural tearing (eight cases), cage subsidence (six cases), wound infection (three cases), cage mobilization (two cases), and radiculopathy due to screw malpositioning (two cases). Only six patients (2.7%) required a second surgery (two patients with wound infection, the patients with cage mobilization, and the patients with screw malpositioning). The cases of dural tearing were repaired during the first surgery with muscle, fascia, and fibrin glue, and they did not develop a fluid wound collection. The six cases of cage subsidence were not re-operated on. Two of them developed an interbody fusion; the others ended in pseudoarthrosis that none required a second surgery.

Symptomatic adjacent segment disease was observed in 14 patients (6.3%), and it consists of spinal stenosis (seven cases), herniated lumbar disc (five cases), and spondylolisthesis (two cases). All patients need a second surgery to relieve pain.

There is a growing body of literature demonstrating that Mi-TLIF and open procedures have similar rate of clinical outcome on radiographic fusion [13–15], justifying the use of a minimally invasive approach that can reduce the soft tissue trauma. In our experience, Mi-TLIF has provided good results at 1 and 2 years follow-up in both pain improvement and radiographic fusion (mean ODI and VAS reduction of 14.7 and 3.5 at 1 year and 13.9 and 3.1 at 2 years; fusion rate of 75.1% at 1 year and 84.1% at 2 years) (Fig. 8.3).

The mean operative time and fluoroscopic exposure were superior to that of an open procedure. However these results were affected by our learning curve that has made first procedures more time-consuming to those we perform now [19, 20]. In fact, as the number of performed Mi-TLIF raised, the operative time has become shorter so that actually no clear differences in surgical time exist between minimally invasive and open procedure. Nevertheless fluoroscopic time still remains a problem in Mi-TLIF procedure because the relevant anatomy is not directly visible and fluoroscopic views are frequently required to confirm the safe and accurate placement of screws and cage [16].

As reported by other authors [13–16], also in our experience, blood loss was minimal, with no cases requiring blood transfusion, and the length of hospital stay was short (mean 4.2 days). In fact, the reduced soft tissue trauma offered by the Mi-TLIF not only decreases blood loss but also improves postoperative pain, so allowing the patient to walk the day after the procedure and consequently providing a faster discharge at home.

Recent reviews have reported similar or lower rates of complications in Mi-TLIF compared to open procedures [13, 14, 21], while other studies highlighted a higher rate of revision surgery. In our series only 21 of the 221 patients (9.5%) had perioperative adverse events, and only six patients (2.7%) required a second surgery. The complications we have to deal with were dural tearing (eight cases), cage subsidence (six cases), wound infection (three cases), cage mobilization (two cases), and radiculopathy due to screw malpositioning (two cases). These complications are the same reported in a recent meta-analysis of 513 patients by Wong et al. [22] in which the incidence of durotomy, instrumentation failure, neurologic deficits, and wound infections were, respectively, 5.1%, 2,1%, 0.8%, and 0.2%.