Cervical Nonunion and Kyphosis
This patient is a pleasant 40-year-old female nonsmoker who presented to us for evaluation of neck pain. She had previously undergone C4-C7 laminectomy and fusion with allograft in early 2011 for neck and arm pain. She did well initially from the standpoint of her arm pain, but her neck pain continued virtually unabated.
Eight months after her first operation, routine x-rays at follow-up revealed loose hardware at C4 with nonunion at C4-C5. She underwent revision fusion with extension of the laminectomy and fusion to C3 proximally again with allograft in September 2011. She did note a decrease in her neck pain after the second procedure and tolerated her remaining discomfort for 1 year until October 2012. At that point, her neck pain increased to the point where it was intolerable, and she was sent for a facet injection at C2-C3 for pain control. She developed blurred vision immediately after the injection and was understandably hesitant to undergo further injections for pain control.
She saw her original surgeon, who noted loose hardware at C6-C7 as well as C2-C3 facet arthritis and recommended C2-T2 revision posterior fusion. She presented to our office for a second opinion in this matter. She complained of posterior neck pain radiating down to the shoulders but denied any arm pain, weakness, difficulty with fine motor skills, impaired balance, or bladder or bowel incontinence. She notes that her biggest impairment is difficulty performing procedures and placing lines in her position as an ICU nurse practitioner. Her medical history is positive for anxiety for which she takes Wellbutrin, but otherwise, her history is unremarkable.
Well-developed female in no apparent distress. Height 5'8, weight 128 lbs.
Posterior neck incision is well-healed with a somewhat broad scar. The paraspinal musculature is significantly displaced from midline with a 1cm depression in the midline with palpable hardware under the skin.
Cervical range of motion is decreased in all planes with flexion limited by about 50% from normal and extension limited to neutral due to pain.
- Neurologic exam reveals 5/5 strength in all upper and lower extremity muscle groups
- Sensation is intact to light touch in the C5-T1 and L3-S1 distributions
- Reflexes are 2+ biceps, triceps, and brachioradialis bilaterally
- Hoffmann's reflexes are negative bilaterally
- Negative for clonus in the lower extremities bilaterally
Pre-operative x-rays of the cervical spine (Figures 1, 2) taken in October 2012 show bilateral lateral mass screws placed from C3-C7. There are crosslinks present at C3 and C5 integrated into the screw caps. The C6 and C7 screws are loose. There is significant facet arthritis and erosion at C2-C3 bilaterally. The cervical spine is neutral to slightly kyphotic and does not reconstitute lordosis on extension films.
CT scans of the cervical spine (Figures 3-5) pre-operatively show loose hardware at C6 and C7 bilaterally with solid fusion from C3-C5. The laminae and spinous processes from C3-C7 have been removed completely. The paraspinal musculature is significantly displaced from midline.
Pre-operative MRI from February 2012 shows no central or foraminal stenosis in the cervical spine.
In addition to her previous surgeries in 2011, the patient has done physical therapy prior to her first surgery and in the period between her second surgery and her consultation in our office. She has been seeing a chiropractor weekly for the last 6 months. She has tried a facet injection as noted in her history, but an adverse reaction to the injection led to lack of enthusiasm for subsequent injections.
Posterior cervical nonunion, C2-C3 arthrosis, and paraspinal muscle dysfunction.
Suggest TreatmentIndicate how you would treat this patient by completing the following brief survey. Your response will be added to our survey results below.
The patient did elect to undergo a circumferential fusion with placement of pars screws at C2 bilaterally, replacement of lateral mass screws from C3-C6, and pedicle screws at T1 and T2. Lateral mass screws could not be placed at C6 due to erosion from the prior screws loosening.
We elected to proceed with revision 360-degree fusion with placement of allografts at C6-C7 and C7-T1 anteriorly and posterior fusion from C2-T2 posteriorly with iliac crest autograft. We elected to use anterior grafts at C6-C7 and C7-T1 for two reasons.
First, the fusion rate for a 360-degree fusion is higher than for a posterior fusion alone, especially in a patient in whom two nonunions had already developed with posterior-only surgeries. Additionally, placement of lordotic grafts should help correct the patient's overall neutral to slightly kyphotic alignment. We felt that anterior fusion alone would be insufficient, especially in light of her significant arthrosis at C2-C3 and need to address this level as well. Extending the fusion to T1 was necessary both to help correct alignment and to obtain a solid foundation for the fusion. The C6 lateral mass screws were quite loose and placement of hardware at this level would likely be difficult or impossible. C6 pedicle screws are technically feasible but much more difficult given the very scarred dura. The patient has tried chiropractic care with little benefit, and injections are not an option for her given her adverse reaction. Physical therapy is highly unlikely to be effective in light of her obvious pseudarthroses.
During the exposure, an incidental durotomy was made between C1 and C2. The leak was well-contained with overlying muscle and no CSF leakage was apparent after the durotomy was made, so we elected not to further expose and repair it. The remainder of the procedure was uneventful.
The patient had some minor headaches post-operatively but was able to leave the hospital on post-op day 4. Four days later, she presented to the emergency room after an apparent seizure at home with worsening headaches. MRI at that time (Figure 6) showed a large CSF collection in the subfascial space coming from the durotomy indicating a persistent leak. EEG showed no seizure activity and she was therefore taken back to surgery for dural repair.
She recovered well from this and the headaches stopped after 2-3 days of bed rest. She was discharged home without further complications. Six months post-op, she continues to have some neck pain, but it is far better than prior to her last procedure, and she has not experienced any more headaches or seizures. She is pleased with her outcome thus far despite lack of complete pain resolution (Figures 7-8).
This case illustrates several important teaching points. In order to get a good surgical result, one must first diagnose the etiology of the problem, then choose the right operation and perform the operation well without complications.
Correctly diagnose the problem and choose appropriate surgical levels.
The patient was noted to have "significant facet arthritis and erosion at C2-C bilaterally" in 2012. Given that this takes years to develop, it was very likely present in 2011, when the initial operation was performed. Careful examination of the plain radiographs and MRI prior to the initial operation would have revealed this. A careful exam, possibly along with facet blocks prior to the index operation would have revealed if this was a symptomatic level. If it was, it either should have been included, or the patient should have been informed as to why it should not be included. Informing the patient about the options and prognosis for this level might have improved patient satisfaction.
With the second operation, the original surgeon fused to C3, increasing the stress to C2-C3. Had he instead done ACDFs from C4-C7 to fix the nonunions, the patient likely would have healed without adding the C3-C4 level. This is assuming that the neck pain was from nonunions.
Other possibilities include the C2-C3 facet arthrosis and inadvertent placement of the C7 screw into the C7-T1 joint, as discussed below. It is usually not difficult to determine if the neck pain is at the top (C2-C3) or base (C6-C7) of the neck on exam. This is assuming that he did not put the original C7 screws into the C7-T1 joint.
Listening to the patient's complaints regarding the location of the pain, correctly evaluating the pre-operative images, obtaining post-operative radiographs, checking for screw misplacement, properly diagnosing the source of pain, and then choosing the appropriate levels might have solved her problem with one and at most, two operations.
Put screws into lateral masses, not into the facet joints.
The CT scan after the second operation shows that the C6 and C7 lateral mass screws violate the C6-C7 and C7-T1 facet joints, respectively. This often causes excruciating neck pain with minimal motion. Since I don't see any other screw tracks on the CT, it is likely that the surgeon used the same trajectory for the index operation. That could explain part or all of the persistent neck pain after the index operation. Immediate removal/replacement of the screws with pedicle screws or adding anterior arthrodesis and fixation after the index operation might have saved this patient the final extensive procedure that resulted in a C2-T1 arthrodesis.
Obtain adequate distal fixation.
For the revision surgery, the original surgeon placed short lateral mass screws into the C7-T1 joint, with a long fusion from C3-C7. With a laminectomy of C3-C7, the posterior tension band has also been removed. This increases the risk of distal screw pullout. C7 pedicle screws or extension to T1 or T2 would have been preferable. When performing long operations from C2 or C3 to C7, it often makes sense to take the fusion down to the upper thoracic spine. Since this is an extensive operation in a 40-year-old, it was not unreasonable to attempt a shorter segment operation. But at the very least, it makes sense to notify her that that she will likely require further surgery in the future. This helps to prevent patient dissatisfaction later, when the surgeon has to tell her that she needs to extend her surgical levels.
Pay attention to soft tissues.
Too often, spine surgeons pay inadequate attention to soft tissues. When dissecting the muscles off of the spine, one should very carefully dissect in the avascular and amuscular midline raphe. This minimizes bleeding and muscle trauma. We typically perform this under microscope visualization to prevent cutting into (and injuring) the muscles. For closure, 3 layers of muscle closure, followed by closure of the fascia, 3 layers of subcutaneous closure, and then the skin will help prevent the type of dehiscence this patient had following her second operation. It also results in a cosmetically pleasing wound.
Few patients trust a surgeon after two failures. No one wants a third operation for the same problem. When performing a revision operation, do everything possible to ensure that there is no need for a third operation. After the index operation failed, the surgeon should have considered circumferential fusion and proximal and distal extension – exactly what was done by Drs. Regan and Kang.
Drs. Regan and Kang have expertly done exactly what was needed to address this poor lady's problem.