Introduction to BAK

Low back disorders are the most common musculoskeletal disorder causing tremendous burden on the health care delivery system. Low back pain has been shown to be one of the causes for the ever increasing rise in worker's compensation payments and continues to be a major source of disability for the working population". Despite these sociological considerations, the pathological processes that result in back disorders have always been a source of confusion for both the patient and the healthcare provider.

Degeneration of the intervertebral disc has been suggested both as a normal physiological process as well as a disease process. Degenerative changes include loss of water content of the nucleus pulposus and annular tears leading to disc space narrowing and occasionally disc herniation. Although there is a high occurrence of abnormal magnetic resonance imaging (MRI) findings in the lumbar spines of asymptomatic subjects', radiographically confirmed degenerative disc conditions are often implicated in back-related symptoms.

There is a wide variety of treatment options for patients that complain of low back and/or leg pain. The prevailing attitude is that an aggressive non- operative therapy should be attempted prior to considering the surgical options. Once non-operative treatment has been exhausted and the patient continues to have pain, there are several surgical options at the disposal of the surgeon.

Lumbar interbody fusion is one of many surgical techniques that can be used to treat symptomatic disc disease. The goal, as with any type of fusion attempt, is to stabilize the spine and allow bony growth between vertebral bodies. The advantage afforded by interbody fusion as opposed to posterior intertransverse process or laminar fusions is the ability to address one of the major sources of pain, namely the disc. Indeed, recent research has shown the disc to be the source of pain in many cases". Laboratory studies have also demonstrated the inflammatory nature of disc material.

Interbody devices are a recent development in the quest for the optimal interbody fusion environment to provide structural stability to the interspace during prove bony fusion. A forerunner of the BAK implant was the "Bagby Basket" that was developed to treat a cervical spine instability condition in horses known as Wobbler Syndrome'. A hollow, stainless steel, perforated cylinder was used to successfully fuse the cervical spine" and later, the metacarpophalangeal' joint in horses after distracting the joint space and filling the basket(s) with autogenous bone. In the mid 1980's, investigators began collaborative efforts to develop a modification of the early basket for use in the human lumbar spine. The result of these efforts produced the BAK implant, a hollow, porous, square-threaded, titanium cylinder""," that can be filled with autograft bone and placed in the intervertebral space.

After extensive laboratory, and animal experimentation , a multi-center clinical trial was initiated. The purpose of this clinical study was to demonstrate the safety and efficacy of the BAK interbody fusion system in stabilizing affected vertebrae, thereby reducing the patient's pain, improving function and achieving osseous fusion.

Materials and Methods

The BAK interbody fusion system (Spine-Tech, Inc., Minneapolis MN) (Figure 1) consists of both implants and associated implantation instruments. The BAK implant is a hollow, porous, square-threaded, cylindrical device made of titanium alloy (TI-6AI-4V). Five sizes of implants were available to the surgeons in this study. The choice of BAK size to achieve maximal safe interbody distraction was determined individually for each patient during preoperative planning. In this clinical study, the BAK was packed with morselized autogenous bone graft (Figure 1) taken from the iliac crest or local bone after decompression and, in most cases, two implants were placed bilaterally in each intervertebral space. The associated instrumentation used for BAK implantation was designed to reproducibly prepare the bone bed while protecting the surrounding tissues and safely positioning the implants".

The salient points in the implantation instrumentation include the distraction of the intervertebrat space and the ability to protect the surrounding anatomical structures. The surgical technique involved disc space preparation for bilateral placement of implants 4 mm apart by first drilling an 8 mm hole in the disc and placing an appropriately sized distraction plug in the resulting hole. A second hole drilled laterally from the first accommodated the drill tube guide pin which was followed by a protective drill tube and sheath. Placement of these distraction plugs resulted in tensioning of the annulus fibrosus and ultimately immediate stabilization of the functional spinal unit after subsequent placement of the implants. Teeth on the distal end of the drill tube were anchored into the vertebral body to maintain uniform bilateral distraction and to enable effective reaming and tapping of the drilled hole.

There were two surgical approaches used in this clinical trial to implant the BAK. The anterior retroperitoneal approach required the assistance of a general or vascular surgeon. The posterior approach required a decompressive laminotomy of sufficient size to accommodate the insertion instruments. The choice of approaches was left to the discretion of the attending surgeon.

An investigational device exemption (IDE) study protocol was finalized in December of 1991. The clinical trial was a prospective, multi-center, non-randomized investigation of the BAK interbody fusion system. The first patient was enrolled in the study in April 1992 and the final patient, prior to database closure, was enrolled in November 1995. For follow- up periods up to and including 24 months, the proceeding analysis and summary represent all available data for each time period at the time of database closure. Data presented beyond two years represents that which was available as of February 1997. Data presented and reviewed in support of the Pre- Market Approval (PMA) application did not include information beyond the two year follow-up.

The following forty-two surgeons at nineteen US. study centers contributed surgical cases for this trial:

• David Arnold (Athens GA),
• Bruce Bartic (Minneapolis MN),
• Gail Benson (Sioux Falls SD),
• Walter Carlson (Sioux Falls SD),
• William Carr (Portland OR),
• Ray Cunningham (Columbia MO),
• Guy Danielson (Tyler TX),
• Edgar Dawson (Los Angeles CA),
• John DowcUe (St. Paul MN),
• Randall Dryer (Austin TX),
• Donald Erickson (Minneapolis MN),
• Jeffrey Flemming (Portland OR),
• Kevin Foley (Memphis TN),
• Bruce Fredrickson (Syracuse NY),
• Catherine Gallo (Eugene OR),
• Patrick Golden (Eugene OR),
• Charles Gordon (Tyler TX),
• Joseph Grant (San Rainon CA),
• Kent Grewe (Portland OR),
• Robert Hacker (Eugene OR),
• Paul Hartleben (St. Paul MN),
• Thomas Highland (Columbia MO),
• Samuel Jorgenson (Danville CA),
• William Kane (Minneapolis MN),
• Scott Kitchel (Eugene OR),
• Stephen Kuslich (Stillwater MN),
• Richard Mattert (Eugene OR),
• Christopher Miller (Eugene OR),
• John Peloza (Dallas TX),
• Gaylon Rockswold (Minneapolis MN),
• Robert Rovner (Pleasanton CA),
• Edward Seljeskog (Rapid City SD),
• John Sherman (Edina MN),
• James Smith (Austin TX),
• William Taylor (Austin TX),
• George Tipton (Austin TX),
• Scott Tromankauser (Boston MA),
• Eric Widell (Ketchum ID),
• John Wilhelmy (Danville CA),
• Hansen Yuan (Syracuse NY),
• Thomas Zdeblick (Madison WI),
• James Zucherman (San Francisco CA).

The indication for the BAK device was painful degenerative disc disease, with or without herniation, at one or two contiguous levels of the lumbar spine (L2 to the sacrum) in patients between the age of 21 and 65. Patients must have presented with a history of persistent, chronic, disabling low back pain of discogenic origin confirmed by radiographic methods such as discography and/or MRI. To be included in the study, the minimum time for symptoms was 6 months. Previous attempts at interbody fusion was considered an exclusion criteria. Patients were also excluded if their diagnosis included spondytolisthests of Grade 11 or higher, active infection, metabolic bone disease such as osteoporosis, poor general heath (i.e., significant circulatory or cardiac disease, active malignancy, etc.), obesity greater than 40% of ideal body weight, pregnancy, and multiple problematic psychosocial factors.

The outcome of the procedure was evaluated at three, six, twelve, twenty-four, and thirty-six months post-surgery. Fusion was determined radiographically with flexion-extension x-rays. Fusion was considered solid if there were no lucencies and, in accordance with previously-published literature 12,25,31, less than three (3) degrees of vertebral motion in the sagittal plane as assessed by computer-assisted analysis of digitized landmarks in each pair of bending films. This objective digitizing method was employed since it was recognized that there is a probability of misinterpretation of fusion integrity based simply on a subjective assessment of flexion-extension overlays. Final-],v. an independent radiologist evaluated all cases in which the digitization method indicated sagittal motion between three (3) and seven (7) degrees; this was deemed necessary because of inherent measurement errors that potentially .1 might have contributed to an unreasonable false-negative or false-positive result. The radiologist made the final assessment of fusion. Patients with two-level procedures were considered a fusion success only if both levels were fused.

Pain was assessed using a six point analog scale similar to that applied by Prolo et al .21 With 1 representing "no pain" and 6 "disabling pain" Successful pain outcome was determined by improvement in the score from preoperative assessment of at least 1 level. Given the prevalence, incidence, and severity of low back pain in the general population, a "normal" range may be defined on this scale between 1.5 and 2.5.

Functional outcome evaluated the following seven parameters on categorical/numeric scales: ability to stand, Sit, walk, squat, put on socks and shoes, the level of recreational activity, and the level of work. A summation of each functional assessment yielded a best (lowest) score possible of seven (7) and the worst (highest) score of 32. As a point of reference, an asymptomatic, "normal" individual would score between 9 and 12 on this scale.

Statistical analysis was performed using the SAS/STAT and BMDP software. Contingency table analyses were performed on the one year and two year data set with variables of interest including: age, gender, presence of listhesis, compensation issues, duration of symptoms, smoking history, previous open spine surgery, preoperative pain score, preoperative function score, level of implantation, presence of single implants, bone graft augmentation, use of bone growth stimulator, and narcotic use.

RESULTS

Study Cohort. Total enrollment between April 1992 and November 1995 for the study was 947 patients. The anterior approach was used in 591 cases (62.4%) with 305 one-level cases and 286 two-level cases. The posterior approach was used in 356 cases (37.6%) with 272 one-level cases and 84 two-level cases. At the time of database closure for FDA reporting, data collection was complete for 97.2% of the eligible patients (n=680) at 12 months and 94.6% of eligible patients (n=299) at 24 months. As of February 1997, 143 of the enrolled patients were followed at 36 months and 11 patients were followed at 48 months.

Patient Variables. The study cohort consisted of 54% males and 46% females. The average age was 41.5 years. Compensation related injuries were present in 57% of the patients. All patients enrolled had a primary diagnosis of degenerative disc disease and 43% had concomitant disc herniation. Spondylolisthesis or retrolisthests was present in 12% of the group. The average duration of symptoms prior to surgery was 5.5 years with 81% of the patients experiencing symptoms for more than I year. Smokers were 26% of the group. Previous spine surgery (primarily decompressions) was reported in 36% of the cases; 5% of the cases had previous posterior fusion attempts. In 4.8% of patients there were situations, such as the presence of scar tissue or inability to gain adequate exposure, in which the surgeon determined that placement of two BAK devices at a vertebral level was not possible.

Procedural Variables. The duration of surgery (elapsed time from incision to closing of the wound) averaged 174 ± 86 minutes for all procedures. Overall mean blood loss was 382 ± 403 cc with the greatest blood loss in the posterior two level group and least in the anterior one level group (p=0.0001). Overall the posterior approach was associated with higher blood loss. The average length of stay was 4.4 ± 2.4 days. Hospital stay was longest for patients undergoing posterior two-level treatments.

Fusion. Table 1 shows the fusion rates (expressed as a percentage of the patients) for the study cohort according to surgical approach and number of levels. As of February 1997, fusion data was available for 118 patients at 36 months.

Fusion success was secondarily analyzed on the basis of the number of levels operated. This analysis showed that 88.3% of levels operated on anteriorly and 85.3% of levels operated on posteriorly were fused at one year. At two years the fusion rate increased to 93.5% of the levels and 90.6% of the levels for anterior and posterior, respectively.

Statistical analyses showed that fusion success was effected by certain variables. The L4-L5 level had a slightly lower fusion rate at 12 months (p=0.041) and 24 months (p=0.027). Patients with a single implant experienced a lower rate of 12 month f;fusion success (p=0.047), however at 24 months the fusion success rates were comparable to those with two devices per level. Smoking was not found to decrease fusion rates in this study population.

Pain. The average preoperative pain score was 5.0 (marked pain), with 77% of the study cohort reporting marked (5) or disabling (6) pain. Moderate 4) pain was reported by 22% of the patients ' preoperatively The average 12 month pa' preoperative in score was 3.1 (mild), and at 24 months was 2.9. At 12 months, 83.3% of patients had improvement of at least one scoring level in pain compared to preoperative scores. Similarly, at 24 months 85.0% of patients reported pain improvement. There was no significant difference seen in pain relief when the data was analyzed according to surgical approach or number of levels implanted. Longitudinal analysis demonstrated pain reduction from preoperative to 3 months (p<0.0001) with the trend continuing past 2 years. table 2 shows the pain scores for the study cohort at the follow-up periods.

As would be expected, pain success was correlated with certain variables. Compensation patients had a lower rate of pain success (p=0.001) at 12 months, however this difference was not evident at 2 years. Patients who had a previous surgery at the operative BAK level had a lower rate of pain success at 12 months (p=0.004) and at 24 months (p=0.071).

Function. The mean preoperative function score was 20.9. The average 12 month score was IS.2 and at 24 month was 14.4. The function success rates were 87.0% at 12 months and 90.7% at 24 months. There was no significant difference noted in functional improvement when the data was analyzed according to surgical approach or number of levels implanted. Longitudinal analysis demonstrated improvement (p<0.0001) in the preoperative function score compared to the 3 month score, with the trend continuing past 2 years. table 3 shows the function scores for the study cohort at the follow-up periods. females demonstrated a higher level of function improvement at 12 months (p-:.':0.008) and at 24 months (p=0.17).

Return to Work. At the time of BAK implantation, 36% of patients were working, 53% were not working due to disability, and 11% were not employed (homemakers, students, retired, etc.). Of those patients who were not working before surgery due to disability, 38% were working at 12 months and 48% returned to work within 24 months. Of those patients considered "work eligible" (i.e., either working or on disability before surgery), 68% were working at 12 months and 78% were working at 24 months. At 3 years, 91% of the work-eligible patients were working.

Worker's compensation (WC) claims were involved in 62.6% of the patients deemed work-eligible preoperatively. Only 28.4% of these patients were working prior to their surgery. The percentage of work-eligible WC patients that were, indeed, working increased to 38.0% at 6 months, 53.1% at one year, 69.8% at two years, and 88.2% at three years. In contrast, 61.1% of the patients without a worker's compensation overlay were working preoperatively and the percentage of these patients that were working increased to 71.1% at 6 months, 79.5% at one year, 83.1% at two years and 100% at three years.

Safety. No unanticipated events or device related deaths were reported in this series. Major complications (i.e., life threatening disability, or extreme treatment measures required) were reported in 2.0% of patients. Major complications reported included: implant migration requiring reoperation, great vessel damage, pulmonary embolism, life-threatening pneumonia, fractured sacrum with instability, and penetration of anterior ligament. There was no significant difference between the anterior versus posterior approach in the incidence of major complications.

Implant migration requiring reoperation, considered one of the major complications, occurred in 1.2% of patients and tended to be an early finding (within 3 months). A retrospective review of the migration cases showed the potential mitigating factors included 1) small implant size relative to disc space height, 2) implant not placed deep enough, and 3) bone threads stripped at the time of surgery.

Intraoperative or perioperative complications occurred more frequently during posterior implantation of the BAK. The majority of these events were dural tears and/or neurologic complications. Postoperative complications (up to 8 weeks), such as ileus or atelectasis, occurred more frequently after anterior placement of the BAK.

Patients with spondylolisthesis tended to have a higher rate of complications than did those patients without.

Migration and neurologic complications were the most commonly reported events in these patients. Most spondylolisthesis patients had the device implanted through the posterior approach. Implanting the BAK in the presence of spondylo-listhesis was found to be more technically challenging due to the inherent instability of the affected motion segment. However, at 12 months the pain, function, and fusion success rates in the spondylolisthesis subgroup were similar to the study group as a whole.

The overall rate of device-related reoperations was low at 4.4%. The device-related reoperations fell into two distinct time periods. Early reoperations (i.e., within 100 days) accounted for one-third of the reoperations and were typically for removal or repositioning of the device due to migration. Late reoperations were performed in two-thirds of those requiring a second procedure with the majority being additional stabilization to relieve ongoing pain. There was no incidence of failure of the structural integrity of the implant.

Discussion