Gregory C. Dowd, MD * and Fremont P. Wirth, MD, FACS**
* Department of Neurosurgery, Louisiana State University Medical Center, New Orleans, LA
** Savannah Neurological Institute, St. Joseph's Hospital, Savannah, Georgia
[Abstract | Introduction | Methods | Results | Discussion | Conclusions | References]
Abstract:
A prospective, randomized trial was performed comparing anterior cervical
discectomy (ACD) with anterior cervical discectomy and fusion (ACDF) for
the treatment of cervical spondylosis with neurologic compromise. 44 patients
had ACD and 40 underwent ACDF. Operative time, hospital stay and the need
for analgesia were less in the ACD group. It was found that while the incidence
of fusion was greater in the ACDF group (97% vs 70%, p< 0.01), patient
satisfaction and a return to pre operative activity level was similar between
groups. This suggests that the addition of a fusion procedure may be unnecessary.
Introduction:
Cervical spondylosis is a process whereby degeneration and instability in
the spine are compensated by hypertrophy of the supporting ligamentous structures
and bony outgrowths. This process has been described as occurring in three
phases: dysfunction, instability and stabilization (22). The end point of
this process is auto fusion of the involved joint space. This is an adaptive
process in many areas of the body (eg., hip, knee) resulting in increased
structural support and some surgical procedures are designed to facilitate
this endpoint (arthrodesis). However, the proximity of neurologic structures
to the spine can lead to unwanted neurologic side effects of this process(3,12,22).
By narrowing the bony spaces within which neural structures lie, compression
and dysfunction can occur. Radiculopathy as a result of foraminal stenosis
or myelopathy from spinal canal stenosis with spinal cord compression may
occur.
Initially, the posterior approach to the cervical spine was utilized (12). Although this yielded favorable results for soft, accessible disc fragments, it provided limited access to and exposure of midline disc fragments and calcified spurs characteristic of cervical spondylosis. Based on the limitations of the posterior approach, the anterior approach to intervertebral disc pathology was developed at several institutions apparently independent of each other. Since this time, controversy has developed concerning the superiority of the anterior or the posterior approach. It has not been demonstrated that one is better than the other. As such, both still have their proponents and detractors (3 4,6,9,19,20).
The initial descriptions of the anterior approach for cervical discectomy all included a fusion procedure (1,3,4). Concern over the possibility of developing late kyphosis from disc space collapse or radiculopathy from foraminal narrowing supported this philosophy. However, as experience with the procedure mounted, it became clear that some patients had relief of the radicular symptoms but developed complications related to the fusion procedure (5). Some surgeons began to perform simple discectomy without the addition of a fusion procedure for these patients (8,10,11,13).
This has led to contention regarding the necessity of a fusion procedure. Many reports are difficult to interpret as the two techniques were utilized at different times and/ or by different surgeons (11,18). In the absence of a randomized prospective study in patients with cervical spondylosis comparing anterior cervical discectomy with and without fusion, firm recommendations regarding the superiority of either approach are difficult to support.To this end, the current study was undertaken.
Methods:
The series reported in this communication represents a three year experience
from 1986 until 1989. One hundred consecutive patients presenting with cervical
spondylosis of one or two levels and a diagnosis of radiculopathy or radiculo-
myelopathy were evaluated for the study. Sixteen patients either withdrew
from the study due to patient treatment preferences or did not meet inclusion
criteria. The remaining 84 patients were prospectively randomized via closed
envelope technique to anterior cervical discectomy alone or with a fusion
procedure. This protocol was reviewed and approved by the institutional
review board.
All of the patients in this series were managed by the principal investigator (FPW). Additionally, the procedures were performed at a single institution in conjunction with a dedicated neurosurgical operating room staff ensuring a modicum of consistency. The setting was favorable for this study as the patient pool is largely drawn from the local area, which aids in patient follow up.
The surgical technique was quite consistent in all cases with the incision
through an appropriate horizontal skin fold on the right side of the
neck. Dissection was via the usual route medial to the carotid sheath but
lateral
to the tracheo- esophageal bundle. Lateral X ray was used to confirm
vertebral level. The Cloward retractor system was used deep to the longis
colli muscle
group. Discectomy was performed utilizing a Smith- Robinson technique
but with opening the posterior longtitudinal ligament in all cases and removal
of posteriorly directed osteophytes with curette and Kerrison rongeur
(Figure
1B). Cases with radiculopathy had foraminal augmentation. The cartilagenous
endplates were carefully removed.
A fusion procedure was added for those patients randomized to this
arm of the study (Figure 1C). Iliac crest autograft bone via a modified
Cloward
technique (21) was utilized for fusion. The recipient site and graft
were tailored to ensure a snug fit. This was facilitated using cervical
traction.
No screw or plate fixation was used. Tissues and skin were closed with
absorbable sutures. No drains were used. Post operatively, fusion patients
wore a soft
collar for a period of six weeks.
Clinical follow up was performed while in hospital, during two office visits (five and ten weeks) and at delayed phone contact (mean 4.5 years, range 1.5- 8 years).
Figure 1:Diagramatic representations of the cervical spondylosis in a preoperative state- A; after an anterior cervical discectomy (ACD)-B; after an anterior cervical discectomy and fusion (ACDF)- C.
Results:
Following randomization, 44 patients underwent ACD while 40 patients received
ACDF. These groups were found to be comparable on the basis of their demographic
data (figure 2),
Figure 2: Demographic data of the 84 patients randomized in this study. The two groups (ACD and ACDF) are comparable. ACD: anterior cervical discectomy; ACDF: anterior cervical discectomy and fusion.
symptoms per patient history (figure 3),
Figure 3: Preoperative symptomatology in the two (ACD v. ACDF) randomized groups. Note that the two groups are quite comparable except for arm weakness- this was a more significant complaint in the ACDF group. However, on neurological examination, there was no significant difference in objective weakness between the two groups.ACD: anterior cervical discectomy; ACDF: anterior cervical discectomy and fusion.
findings on physical exam (figure 4)
Figure 4: Physical examination findings in the two (ACD v. ACDF) randomized groups. Once again, the two groups are comparable.ACD: anterior cervical discectomy; ACDF: anterior cervical discectomy and fusion.
and the form of preoperative therapy that each received (figure 5).
Figure 5: Preoperative therapy and treatment (Rx) duration in two groups (ACD v. ACDF). The two groups are very similar.ACD: anterior cervical discectomy; ACDF: anterior cervical discectomy and fusion.
Importantly, the number of patients with problems caused by motor vehicle accident or involved in compensation claims was similar between groups. More patients in the ACDF group (10/40 vs. 1/44, p< 0.005) complained of arm weakness. However, physical examination (18/40 vs. 18/44, ns) did not support this difference between groups.
The distribution of operative levels and ratio of single/ double level procedures was similar between groups (figure 6).
Figure 6: The distribution of operative levels and the ratio of single/ double procedures is similar in the two groups (ACD v. ACDF).ACD: anterior cervical discectomy; ACDF: anterior cervical discectomy and fusion.
Two level procedures were performed in 59% of the ACD group compared
with 50% of the ACDF group.
The presence of myelopathy was determined by evidence of hand intrinsic
musculature weakness and/ or pathologic long tract findings. Gait, bowel
and bladder function were not recorded. By these criteria,12 patients in
the ACD group and 14 patients in the ACDF group demonstrated signs of myelopathy.
The operative time was approximately 29 minutes longer in the ACDF group (131 vs 102 minutes, p< 0.001). This difference and significance were maintained when single and double levels were compared separately. The length of time in post operative unit was not different between groups (103 vs. 98 minutes, ns)
Post operative results were quite different between the two groups. There was no perioperative mortality and no patients deteriorated neurologically. The ACD group was significantly more satisfied with their pain relief (34/44 v 20/40, p<0.01) on the morning following surgery. Patients with ACD requested fewer narcotic shots (3.5 vs. 3.6, ns), narcotic pills (3.6 vs. 4.4, p< 0.05) and non-narcotic analgesics (0.4 vs. 1.0, p< 0.01). Two patients developed transient hoarseness in the ACDF group compared with none in the ACD group. Medical complications (urinary tract infections, atelectasis, elevated leukocyte count, etc.) were significantly fewer in the ACD group (4/44 vs 10/40, p<0.05). Finally, the hospital stay was shorter for the ACD group (3.6 vs. 5.0 days, p< 0.005). Taken together, these data suggest that a fusion procedure adds a significant length of time to the operative procedure. ACDF patients had more pain, developed more medical complications and stayed in the hospital longer.
The two groups were evaluated for rdiographic fusion. Thirty-one patients in each group had complete flexion- extension cervical X ray series post operatively. Patients with two level procedures that showed fusion at only a single level were classified as failures. ACDF patients had a statistically greater fusion rate (30/31 v 22/31, p<0.01). Stratification of patients between one and two level procedures, smoking history or age did not demonstrate any difference in fusion success with multi- variate analyses. However, women were found to have a lower rate of successful fusion (28/37 v 24/25, p<0.05) when compared with men regardless of procedure type.
Post hospital follow up revealed that the discrepancy in pain relief abated by the first office visit ("improved": ACD 40/44 vs. ACDF 37/40, ns). At the second office visit there was near universal improvement in pain (ACD 31/32 vs. ACDF 35/35, ns) in both groups. Phone follow up demonstrated some recrudescence of pain although this was similar between groups. Severe pain was present in some patients ( ACD 4/33 vs ACDF 5/23, ns) at the time of phone follow up.
The number of patients returning to work was greater in the ACD group at both the first (10/44 vs. 4/37, ns) and second (12/32 vs. 10/33, ns) office visits. Phone follow up revealed that more patients in the ACD group had maintained their pre-operative level of activity (24/33 v 15/23, ns). Also, when asked whether the procedure "helped and would be worth repeating given a second chance", patients from both groups were largely satisfied ( ACD 33/33 vs ACDF 22/23, ns). However, these differences do not attain statistical significance.
Two patients required a second operative procedure, both from the ACD group. The first patient had persistence of radicular symptoms following a C5-6 and C6-7 ACD. At four months postoperatively, posterior foraminotomy was performed at these levels with relief of symptoms. The second patient had a C5-6 ACD with immediate relief of symptoms and radiological evidence of successful fusion. She gradually became symptomatic again and underwent C4-5 and C6-7 ACDF eight years after her initial operation.
Discussion:
Cervical spondylosis is a common condition that has been recognized as a
major cause of arm and neck pain (13,17). Dysfunction at the level of the
intervertebral disc and joint complex is the earliest form of the disorder.
This may occur in response to a specific injury or as a result of cumulative
wear and tear of an active lifestyle (7,22). Hypertrophy of the supporting
tissues in response to this subtle instability occurs. This is a heterogeneous
response with ligamentous hypertrophy, joint narrowing, reactive growth
from the bony endplates and calcification of the joint capsule. These changes
perform the teleologically useful function of stabilizing worn joints with
a stable osseous bridge or auto fusion of that joint space (22). Procedures
to facilitate arthrodesis have been developed for other joints (eg, knee)
to achieve such stability. However, the proximity of neurologic structures
to the affected joint complexes may result in adverse consequences (2,7,17,22).
This is often the case in the cervical spine. Joint narrowing, ligamentous
thickening and bony spurring can cause compression at the root level (ie,
radiculopathy) or the spinal cord level (ie, myelopathy). Also, pain related
to joint instability may preceed neurologic injury (type III syndrome) (17).
The relative importance of each pathologic change will vary between patients. As such, treatment strategies need to be tailored to match the specific changes present in a given patient. Cervical traction will facilitate foraminal opening and cervical collars aid in minimizing dynamic injury. Strengthening exercises can shift some of the support function away from the spine. However, cases with true mechanical compression of neurologic structures often require surgical intervention (17).
Both anterior and posterior approaches have been used for the treatment of this condition (12,19,20). However, incomplete recovery for pathology situated anterior to the spinal cord has been found following posterior decompression alone (6,19). The treatment of radiculopathy is largely successful with either approach. The anterior approach offers maximal exposure to the pathology centered around the disc space (1,3,14). Following decompression of the involved neurologic structures, many authors favor the use of a fusion procedure (1,3,9,14,17,23). Arguments in favor of this approach include the maintenance of disc space height which avoids vertebral settling and minimizes the potential to develop foraminal stenosis. Also, fusion will remove the instability component of the disorder which may cause progressive deterioration (22).
The necessity of adding a fusion procedure is
not universally accepted (5,10,11,13, 15). In the cervical area, fusion
has its own set of complications
in addition
to those of ACD alone (5,16). Most obvious are autogenous donor
site complications such as hematoma and neurologic injury to the lateral
femoral cutaneous
nerve (meralgia paresthetica). Patients often complain of more
pain
in this area than at the cervical surgical site. The use of allograft
material has
the potential for transmission of viral particles and other infectious
agents. Placing the graft in the recipient site may entail further
neck dissection
for exposure than discectomy alone. Late complications of cervical
fusion include graft extrusion causing spinal cord injury or dysphagia.
The results of this study suggest that a fusion procedure adds
little to the ACD alone in the surgical management of cervical
spondylosis
at one
or two levels. The operative time was significantly longer for
ACDF patients. The ACDF patients had more pain as measured by
initial satisfaction
and
requests for analgesic medications. The incidence of medical complications,
such as atelectasis and urinary tract infection, related to the
procedure was greater in the ACDF group and hospital stay was
significantly prolonged.
As expected, the rate of successful fusion was statistically different between groups. The ACDF group had a 97% fusion rate compared with 70% in the ACD group. Connolly et al. have found that using cine radiographs, motion was detectable at operative levels which appeared fused on plain films (5). Fusion has been evaluated in a goat model (23). The investigators in this study found only 25% osseous union by histology in levels exhibiting radiological fusion. This suggests that bony fusion is poorly evaluated by radiographic measures and may be inappropriate as an endpoint. A better goal is relief of symptoms and patient satisfaction. This was well achieved with both ACD and ACDF.
Rosenorn reported a series of patients with cervical herniated disc prospectively randomized to anterior cervical discectomy with or without fusion (15). Five separate surgeons participated in this study. Patients undergoing ACD had a significantly greater return to work for the first 9 weeks and superior clinical outcome at both 3 and 12 months. This is especially significant considering that cadaveric donor bone was utilized for fusion procedures in that series.
Conclusions:
The results of the study demonstrate a distinct difference in early patient
satisfaction. Operative times were shorter in the ACD group. Post operative
pain relief was greater in the ACD group. Post operative requirements for
analgesia were less in the ACD group. This resulted in a decreased length
of hospitalization in the ACD group. Two patients did require further surgery
in the ACD group, but the return to preoperative status and overall satisfaction
were equal between groups.
As expected the rate of successful fusion was greater for the ACDF group (97% vs. 70%, p, 0.01). However, this did not correlate with an improved level of satisfaction with the procedure or less long term pain. This would suggest that the addition of a fusion procedure is not absolutely necessary. Procedure selection should be at the discretion of the surgeon and based on individual patient factors.