- Email: [email protected]
Should long-segment cervical fusions be routinely carried into the thoracic spine? A multicenter analysis
Accepted Manuscript Title: Should long segment cervical fusions be routinely carried into the thoracic spine? –multi-center analysis Author: Eeric Truumees, Devender Singh, Matthew J. Geck, John K. Stokes PII: DOI: Reference:
S1529-9430(17)31003-3 https://doi.org/doi:10.1016/j.spinee.2017.09.010 SPINEE 57504
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
20-6-2017 28-8-2017 20-9-2017
Please cite this article as: Eeric Truumees, Devender Singh, Matthew J. Geck, John K. Stokes, Should long segment cervical fusions be routinely carried into the thoracic spine? –multi-center analysis, The Spine Journal (2017), https://doi.org/doi:10.1016/j.spinee.2017.09.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Should Long Segment Cervical Fusions Be Routinely Carried into the Thoracic Spine? –
2
Multi-center Analysis
3
Eeric Truumees, MD1; Devender Singh, PhD2; Matthew J. Geck, MD3, John K. Stokes, MD4
4 5 6
1
7
2
8 9
3
10
4
The University of Texas Dell Medical School, Seton Brain & Spine Institute: Professor of Surgery, CEO, Austin, TX Seton Spine & Scoliosis Center: Research Scientist, Austin, TX
The University of Texas Dell Medical School, Seton Spine & Scoliosis Center: Assistant Professor of Surgery, Austin, TX Seton Spine & Scoliosis Center: Neurosurgeon, Austin, TX
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Corresponding Author: Eeric Truumees, MD Professor of Surgery The University of Texas Dell Medical School CEO-Seton Brain & Spine Institute 1600 West 38th Street, Suite 200, Austin, Texas 78731-6400 Phone: 512 324 3580 Email: [email protected],
Conflicts of Interest and Source of Funding: The pilot phase of this study was funded by Globus Inc. Relevant financial disclosure outside the submitted work: board membership; royalties, stock/shareholder, ownership, consultancy. Abstract
1 Page 1 of 18
1
Background Context: While recommendations for caudal “end level” in posterior cervical
2
reconstruction remain highly variable, the benefits of routine extension of posterior cervical
3
fusions into the thoracic spine remain unclear.
4 5
Purpose: Compare clinical and radiographic outcomes in patients in whom posterior fusions
6
ended in the cervical spine versus those in whom the fusion was extended into the thoracic spine.
7 8
Study design/Setting: Multi-center retrospective analysis of prospectively followed patients.
9
Patient Sample: 177 adult spine patients undergoing 3 or more level posterior cervical fusions
10
for degenerative disease from January 2008 to May 2013.
11 12
Outcome Measures: Cervical lordosis, C2-C7 sagittal plumbline, T1 slope, visual analog scale
13
(VAS), oswestry disability index (ODI), rate of pseudarthrosis, length of hospital stay (LOS),
14
estimated blood loss (EBL) and operative time (OR).
15 16
Methods: We assembled a multicenter (4 sites) radiographic and clinical database of patients
17
that had undergone 3 or more level posterior cervical fusions for degenerative disease from
18
January 2008 to May 2013 with at least 2 years of post-operative (post-op) follow-ups. Patients
19
were divided into two groups: group I (fusion ending in the cervical spine) and group II (fusion
20
extending into the thoracic spine). All radiographic measurements were performed by an
21
independent experienced clinical researcher.
22 23
Results: Group I and Group II had 104 and 73 patients, respectively. Mean EBL for group II was
24
significantly higher than group I. Mean OR time and LOS were comparatively higher for group
25
II than group I but were not statistically significant (p>0.05). Mean cervical lordosis improved
26
post-operatively in both groups. There were no statistically significant differences in change or 2 Page 2 of 18
1
maintenance of mean cervical lordosis (2 wk vs. 2 year post-op) between the two groups
2
(p>0.05). Similarly, the change in mean C2-C7 sagittal plumbline and T1 slope was not
3
statistically significantly different between the two groups or with follow-up(p>0.05). Clinically,
4
significant improvements in VAS and ODI were noted in both groups from pre-op to final
5
follow-up, but the difference between groups was not statistically significant. While the rate of
6
pseudarthrosis was significantly higher in group I (21.2%) than group II (10.96%), there were no
7
statistically significant differences in adjacent segment degeneration or revision surgery rates
8
between the groups.
9 10 11
Conclusion: Both groups had similar clinical and radiographic outcomes. Extension of a
12
posterior cervical fusion into the thoracic spine leads to lower pseudarthrosis rate, while stopping
13
in the cervical spine yields lower EBL, OR and LOS, demonstrating that there are different
14
benefits for each approach. However, while the optimal end-level remains debatable, there are
15
scenarios in which upper thoracic extension should be considered. At this point, we recommend
16
extension of surgery in smokers and other patients at increased risk for pseudarthrosis as well as
17
in patients with anatomical limitations to strong C7 bone anchorage.
18 19
Keywords: Posterior cervical fusion, cervicothoracic fusion, cervicothoracic junction, outcomes,
20
multilevel posterior cervical fusion
21 22 23
Introduction 3 Page 3 of 18
1
While recommendations for caudal “end level” in posterior cervical reconstruction remain highly
2
variable, the benefits of routine extension of posterior cervical fusions into the thoracic spine
3
remain unclear. It has been suggested that long posterior cervical fusions risk subjacent
4
degeneration, spondylolisthesis, fracture, or kyphotic collapse when the fusion ends at C6 or C7.
5
In fact, ongoing neck pain and adjacent segment degeneration are two of the most common
6
sources of post-operative morbidity in patients undergoing long posterior cervical fusions
7
population.
8 9
Recommendations for appropriate caudal end vertebra vary widely (stop at C7 vs. extend to T1
10
through T4). Some biomechanical, but sparse clinical evidence supports these recommendations.
11
In a recently published study (2016), Schroeder and colleagues suggested that multilevel
12
posterior cervical fusions should be extended to T1 as long construct at C7 increases the rate of
13
revision (35.3% vs. 18.3%).1 Authors suggested that the odds of revisions in patients whose
14
construct terminated at C7 were 2.29 times greater than patients whose construct terminated at
15
T1 (p=0.02). Another study examining this research question was presented at various spine
16
meetings between 2009 ~2013 but has not yet been published.2-5 The authors reviewed cases
17
with minimum 3 levels posterior cervical fusion performed between 2000~2006 with at least 18
18
month follow-up. Radiographic and clinical outcomes in 36 patients with end-vertebrae
19
instrumentation in the cervical spine (EIV-C) were compared with 53 patients with a thoracic
20
end vertebra (EIV-T). The groups’ demographics were similar. Interestingly, despite the
21
additional levels exposed and fused in the thoracic group, operative time was the same. The rate
22
of symptomatic adjacent segment degeneration (ASD) requiring intervention was significantly
23
higher at the cranial adjacent level in the EIV-C group (28%) compared with the EIV-T group
24
(9%, p=0.04) at 2 years. Similarly, EIV-C patients had a significantly higher rate of caudal-level 4 Page 4 of 18
1
symptomatic ASD requiring intervention compared with EIV-T patients (39% vs 15%, p=0.01).
2
The development of caudal-level ASD was highest at C7 (41%), followed by C6 (40%), C5
3
(25%), T1 (18%), T3 (18%), T2 (17%), and T4 (0%). Despite higher rates of symptomatic ASD,
4
the overall complication rate and surgical revision rates were similar between the groups. Neck
5
disability index (NDI) outcomes at 2 years post-op were significantly better in the EIV-T group
6
(24.5 vs 34.0, p=0.05). The authors concluded that crossing the cervico-thoracic junction affords
7
protection of the adjacent levels without adding significant operative time or morbidity.
8 9
Based on the limited published data available, our primary objective was to compare clinical and
10
radiographic outcomes in patients in whom posterior fusions ended in the cervical spine versus
11
those in whom the fusion was extended into the thoracic spine. Our null hypothesis was that
12
extension of posterior cervical fusions into the upper thoracic spine improves clinical outcomes
13
while decreasing kyphosis.
14 15
Materials and Methods
16
The study was approved by the institutional review board. We assembled a multicenter (4 sites)
17
radiographic and clinical database of patients that had undergone 3 or more level posterior
18
cervical fusions for degenerative disease from January, 2008 to May, 2013 with at least 2 years
19
of post-operative (post-op) follow-ups. Patients were divided into two groups: group I (fusion
20
ending in the cervical spine) and group II (fusion extending into the thoracic spine). All
21
radiographic measurements were performed by an independent experienced clinical researcher.
22
Cervical lordosis was defined as the sagittal Cobb angle between C2 and C7 vertebral bodies. T1
23
slope was defined as the angle formed by the intersection of a line drawn tangential to the
24
superior end plate of T1 and a horizontal reference line. C2-C7 sagittal plumbline was defined as 5 Page 5 of 18
1
the distance between C2 plumbline and C7. For the analysis, bone grafts were divided into four
2
groups: local only; local and allografts; bone morphogenetic protein (BMP) only; and iliac crest
3
only. Current smokers included those patients smoking at the time of or within 6 months of their
4
surgery. The odds ratio was defined as the odds of smoker patients developing pseudarthrosis
5
compared to the odds of non-smoker patients developing pseudarthrosis while undergoing
6
multilevel posterior cervical fusions.
7 8
Statistical analysis
9
Characteristics of subjects were analyzed using means and standard deviations calculations. As
10
needed, two–sample t-test with unequal variances and paired t-tests were used to assess for
11
differences between the two groups. The non-normally distributed data were analyzed via the
12
Mann-Whitney’s U test (nonparametric). Analysis of variance (ANOVA) was used to investigate
13
the effects of type of bone grafts on the rate of pseudarthrosis. The level of significance was set
14
at α=0.05. All the statistical analyses were performed using SPSS software, version 15.0 (SPSS
15
Inc.).
16 17
Results
18
Group I and group II had 104 and 73 patients, respectively. The demography of the two groups
19
was similar as shown in Table 1. The minimum and maximum number of spinal levels treated
20
for group I and group II were 3 & 4 and 3 & 9, respectively. As shown in Table 2, mean total
21
estimated blood loss (EBL) for group II was significantly higher than group I (578.9±518.5 ml
22
vs. 367.7±308.6 ml, p<0.05). Mean operative time (266.9±144.2 min vs. 248.9±133.8 min) and
23
length of hospital stay (5.6±3.7 days vs. 4.94±3 days) were comparatively higher for group II
24
than group I but were not statistically significant (p>0.05). Tables 3 and 4 details the number of
6 Page 6 of 18
1
spinal levels treated for group I and II, respectively. As can be seen from these tables, C3-C7 and
2
C3-T1 were the most frequently treated levels in groups I and II, respectively. 47.6% of group I
3
patients were treated with both anterior and posterior approach. In group II, 37% of the patients
4
were treated with both anterior and posterior approach.
5 6
Mean cervical lordosis at 2 years post-op improved in both groups. As shown in Figure 1, there
7
was no significant statistical difference in change in mean cervical lordosis (2 wks. vs. 2 years
8
post-op)between the two groups (p>0.05).
9 10
Similarly, there were no significant statistical differences in change (2wks. vs. 2 years post-op)
11
in mean C2-C7 sagittal plumbline (group I: 2.9±1.87 mm; group II:2.7± 1.6mm) and T1 slope
12
(group I: 2.2±2.4˚; group II: 3.4±2.8˚) between the two groups (p>0.05).
13 14
There were significant improvements in mean clinical outcomes (i.e. visual analog scale (VAS)
15
and oswestry disability index (ODI) at 2 years follow ups in both groups but there were no
16
statistically significant differences between the two groups (p>0.05). Mean VAS improvements
17
(pre-op vs. 2 years post-op) in group I and group II were 5.7±2.2 vs. 3.5±3.1 and 5.5±2.5 vs.
18
3.2±2.6, respectively. Similarly, mean ODI improvements (pre-op vs. 2 years post-op) in group I
19
and group II were 39.2±22.1 vs. 26.2±20.2 and 44.9±20.1 vs. 29.6±19.7, respectively.
20 21
Revision Surgery
22
The all-cause revision rates at 2 years follow ups for Group I (ending in cervical spine) and
23
Group II (ending in thoracic spine) was 8.9% and 5.5%, respectively. This difference was not
24
statistically significant.
25
Pseudarthrosis 7 Page 7 of 18
1
Rate of pseudarthrosis was higher in group I (21.2%) than group II (10.96%). This difference
2
was statistically significant (p<0.05). Mean age of patients with pseudarthrosis in group I and
3
group II were 56(±9) and 67 (±4) years, respectively. Females had higher numbers of
4
pseudarthrosis than males (group I: 67% vs. 33%; group II: 55% vs. 45%; p<0.05). Overall,
5
53.3% of the patients with pseudarthrosis were current smokers. The rate of smoking in the solid
6
fusion group was 21.9%. The odds ratio of pseudarthrosis for a smoker compared with a non-
7
smoker was 4.071 (95% CI: 1.798-9.221). Mean number of spinal levels treated for patients
8
with pseudarthrosis in group I and group II were 3.6(±0.79) and 6.2(±2.5), respectively. Mean T1
9
slope for patients with pseudarthrosis increased significantly (2 wk vs. 2 year post-op) in both
10
groups (p<0.05). Both groups with pseudarthrosis had significantly higher mean C2-C7 sagittal
11
plumbline at 2 years follow-up (p<0.05). Mean cervical lordosis decreased in both groups with
12
pseudarthrosis (2 wk vs. 2 year post-op). The difference was not statistically significant (p>0.05).
13
Overall, ANOVA showed no significant effect of type of bone grafts on the rate of
14
pseudarthrosis (p>0.05).
15 16
Discussion
17
The anatomic and biomechanical complexities of posterior spine surgery crossing the
18
cervicothoracic junction, and other junctional regions, pose particular treatment challenges.6-12
19
While an extensive literature addresses questions of long term construct viability with long
20
fusions to or across the thoraco-lumbar and lumbo-sacral junctions, little supportive data
21
supports recommendations for an optimal caudal “end level” in posterior cervical reconstruction.
22
This study aimed to answer the research question should long segment cervical fusions be
23
routinely carried out into the thoracic spine. We hypothesized that extension of posterior cervical
8 Page 8 of 18
1
fusions into the upper thoracic spine would improve clinical outcome possible through optimized
2
sagittal balance and decreased adjacent segment degeneration and spondylolisthesis rates. Yet,
3
our data suggests that, except for pseudarthrosis rates, both groups had similar clinical and
4
radiographic outcomes. Clinically, significant but similar improvements in VAS and ODI were
5
noted in both groups at 2 years follow-up. The rate of revision surgery and supra- or subjacent
6
degeneration or instability was statistically similar between groups. Mean cervical lordosis, C2-
7
C7 sagittal plumbline and T1 slope improved in both groups at 2 years post-op follow up. As
8
expected, the increased levels incorporated in the reconstructions performed in Group II were
9
associated with greater EBL, OR time and LOS than group I.
10
Interestingly; pseudarthrosis was more frequently encountered in group I (21.2%) than group II
11
(10.96%). Females in both groups had significantly higher rate of pseudarthrosis than males. As
12
expected, majority (53.3%) of patients with pseudoarthrosis were current smokers. Based on
13
odds ratio, we estimated that current smoker patients undergoing multilevel posterior cervical
14
fusions were 4 times more likely to develop pseudarthrosis than non-smoker patients undergoing
15
similar surgical procedures. For both groups with pseudarthrosis, mean T1 slope, C2-C7 sagittal
16
plumbline and cervical lordosis at 2 years follow up differed significantly from those who did
17
not develop pseudarthrosis. We did not find any significant effect of type of bone grafts on the
18
rate of pseudarthrosis. We had anticipated that, given the infrequency with which anterior
19
column procedures are performed at C7-T1 and T1-2, the pseudarthrosis rate would be higher.
20
At this point, possible reasons for the decreased pseudarthrosis rate include detection issues.
21
Especially if the lower cervical anchor was a lateral mass screw, this implant may exhibit more
22
obvious failure in the form of loosening or posterior displacement. Such implant failure, in turn,
23
could alert the treating team to an otherwise radiographically undetectable pseudarthrosis. Other
9 Page 9 of 18
1
possibilities include greater surface area for the fusion mass afforded by extension into the
2
thoracic spine. Alternatively, the larger screws typically employed in the thoracic spine may
3
offer greater construct rigidity and, hence, a more stable mechanical environment in which to
4
achieve fusion when compared to lower cervical anchors, lateral mass screws in particular.
5 6
Schroeder and colleagues (2016) in a single center study with 219 patients favored extending the
7
multilevel cervical fusions to T1.1 They concluded that the multilevel cervical fusions should be
8
extended to T1 as stopping a long construct at C7 increases the rate of revision. Their study
9
divided the cohort into three groups (fusions terminating at C7, T1 and T2-T4). The rate of
10
revisions in C7, T1 and T2-T4 were 35.3%, 18.3% and 40%, respectively (p=0.008). The
11
multivariate analysis showed that the odds of revision in patients whose construct terminated at
12
C7 were 2.29 (1.16-4.61) times more than patients whose construct terminated at T1 (p=0.02).
13
The study did not find any significant difference in revision rate between stopping at T1 and T2-
14
T4.
15 16
In that our findings differ from Schroeder and colleagues, some key differences between the two
17
studies should be described: Our series is a multi-center (4 sites) study with 177 patients while
18
Schroeder’s study is a single center study with 219 patients. Both studies had similar patient
19
demographics and number of minimum posterior spinal levels treated. While Schroeder’s study
20
reported a larger sample size, only 31.5% of their patients completed minimum radiographic
21
follow ups (1 year). In contrast, 95% of our patients completed minimum radiographic follow
22
ups (2 years). Additionally, Schroeder’s study did not report any surgical variables (such as EBL,
23
OR and LOS) and/or patient reported outcomes such as VAS and ODI. We thoroughly
24
investigated these variables in our cohort. 10 Page 10 of 18
1
We also compared our study to that of Auerbach and colleagues. While the abstract has been
2
published in the transactions of spine meetings, we were unable to locate the full paper. Their
3
single center study, with smaller sample size of 89, found that crossing the cervico-thoracic
4
junction protects the adjacent levels without adding significant operative time or morbidity.2-5
5
We used this study to generate our null hypothesis. However, our multi-center series of 177
6
patients found similar clinical and radiographic outcomes in two groups.
7 8
Given the limited data available and the conflicting findings with Schroeder et al. and Auerbach
9
et al., we recommend that these results should be verified in prospective settings with additional
10
patients, longer follow-up and greater statistical power. Certain subgroups of patients may well
11
benefit from extension of their fusion into the upper thoracic spine. For example, those in whom
12
reliable fixation to the C7 pedicle is not available or those with greater need for sagittal balance
13
correction. But, for frailer, myelopathic patients, shorter segment fusion with appropriate bone
14
building and muscle strengthening might be preferred.
15 16
It is important to mention several limitations of our study. Among them are the retrospective
17
nature of the study, limited sample size and heterogeneous patient population. Additionally, we
18
did not conduct a multivariate analysis to account for any confounders (such as patient
19
comorbidities, etc.) in our findings. Nonetheless, we controlled all our radiographic
20
measurements as these were measured by a single experienced clinical researcher. All other
21
clinical data were thoroughly verified with other participating centers. However, the data set had
22
limitations. For example, in the setting of revision surgery, the indication for that revision was
23
not always clearly recorded or multiple indications for surgery were noted. These measures
24
eliminated the possibilities of any inter-observer bias in the data collection and analysis. Finally, 11 Page 11 of 18
1
95% of our cohort completed the minimum 2 years radiographic follow up as required by the
2
study. Of course, many of the potential advantages of extension of a long posterior cervical
3
fusion into the thoracic spine may be better explicated through longer term follow-up. We intend
4
to re-analyze this cohort at a minimum of 5 years post-operatively. Currently, however, this is
5
the longest follow-up interval reported. We must also acknowledge the limitations in the
6
radiographs available in detecting upper thoracic spondylolisthesis. Body habitus and the
7
shoulder girdle often limited fine detail assessment between C7 and T3. Additionally, the minor
8
rotation encountered in swimmer’s lateral views limited detection of minor degrees of subjacent
9
level anterolisthesis. Our prospective study includes use of EOS imaging to better detail these
10
aspects of cervico-thoracic alignment.
11 12
Conclusion
13
We conclude routine extension of long posterior cervical fusions into the thoracic spine does not
14
confer significant benefits in terms of cervical alignment or patient reported outcomes at two
15
year follow-up. Given that such extension is associated with higher EBL, OR and LOS, we
16
suggest that its use be limited in medically frail patients. On the other hand, crossing the cervico-
17
thoracic junction may be justified in smokers or other patients at increased risk for pseudarthrosis
18
as well as in patient in whom adequate lower cervical bone anchorage is not available.
19 20
Acknowledgement
21
We would like to thank the other sites for contributing their data for this study. We thank Globus
22
Inc. for funding the pilot phase of this study.
23
12 Page 12 of 18
1 2 3
References 1. Schroeder GD, Kepler CK, Kurd MF, Mead L, Millhouse PW, Kumar P, Nicholson K,
4
Stawicki C, Helber A, Fasciano D, Patel AA,
5
Anderson DG, Hilibrand AS, Vaccaro AR. Is it necessary to extend a multilevel
6
posterior cervical decompression and fusion to the upper thoracic spine? Spine 2016;
7
41(23); 1845-9
Woods BI, Radcliff KE, Rihn JA,
8
2. Auerbach J, Cho W, Riew KD. Crossing the cervicothoracic junction in multilevel
9
posterior cervical fusions reduces the rate of symptomatic adjacent segment breakdown.
10
37th Annual Meeting, poster, Cervical Spine Research Society (CSRS), 2009
11
3. Auerbach JD, Sehn JK, Cho W, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
12
MS, Riew KD. Crossing the cervicothoracic junction in multilevel posterior cervical
13
fusions reduces the rate of symptomatic adjacent segment breakdown. 17th Annual
14
Meeting, podium, International Meeting on Advanced Spine Techniques (IMAST), 2010
15
4. Auerbach JD, Cho W, Sehn JK, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
16
MS, Riew KD. Crossing the cervicothoracic junction in multilevel posterior cervical
17
fusions reduces the rate of symptomatic adjacent segment breakdown. 26th Annual
18
Meeting, podium, North American Spine Society (NASS), 2011
19
5. Cho W, Auerbach JD, Sehn JK, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
20
MS, Riew KD: Crossing the Cervicothoracic Junction in Multilevel Posterior Cervical
21
Fusions Reduces the Rate of Symptomatic Adjacent Segment Breakdown. Poster,
22
American Academy of Orthopedic Surgeons (AAOS), 2013
23 24
6. An HS, Vaccaro A, Cotler JM, Lin S. Spinal disorders at the cervicothoracic junction. Spine 1994; 19:2557-64 13 Page 13 of 18
1
7. Inoue A, Ikata T, Katoh S. Spinal deformity following surgery for spinal cord tumors
2
and tumorous lesions: analysis based on an assessment of the spinal functional curve.
3
Spinal Cord 1996; 34:536-42
4 5 6 7 8 9
8. Stanescu S, Ebraheim NA, Yeasting R, Bailey AS, Jackson WT, Morphometric evaluation of the cervico-thoracic junction. Spine 1994; 19:2082-88 9. Bailey AS, Stanescu S, Yeasting RA, Ebraheim NA, Jackson T. Anatomic relationships of the cervicothoracic junction. Spine 1995; 20:1431-39 10. Chapman JR, Anderson PA, Pepin C, Toomey S, Newell DW, Grady MS. Posterior instrumentation of the unstable cervicothoracic spine. J Neurosurg 1996; 84;552-58
10
11. Albert TJ, Klein GR, Joffe D, Vaccaro AR. Use of cervicothoracic junction pedicle
11
screws for reconstruction of complex cervical spine pathology. Spine 1998; 23:1596-99
12
12. White AA III, Panjabi MM, Clinical Biomechanics of the Spine, ed 2. Philadelphia, PA,
13
JB Lippincott, 1990
14 15 16 17
14 Page 14 of 18
1 2
Figure 1: Mean (Standard deviation) cervical lordosis for the two groups at 2 week and 2 years post-op 20
Mean Cervical Lordosis (˚)
16 12
10.62 9.27
10.32
10.7
Group I Group II
8 4 0 3 4 5
2 wk post-op
2 years post-op
Note: There was no significant statistical difference in change in cervical lordosis (2 wk vs. 2 year post-op) between the two groups (p>0.05)
6 7 8
15 Page 15 of 18
1
Table 1: Demographics
2
Group I Group II 104 73 N 52/52 40/33 Male/Female 59.74 (±11.2) 59.16 (±13.6) Age (years) 28.3 (±6.04) 29.64 (±15.4)* BMI Smoking Status 30 (28.8%) 18 (24.7%) Smokers 60 (57.7%) 43 (58.9%) Non-Smokers 14 (13.5%) 12 (16.4%) Former Smokers (> 6 months) Ethnicity 65 63 Caucasian 22 4 Hispanic 10 3 African-American 7 3 Others * Significant statistical difference between the two groups (p<0.05)
3 4 5 6
Table 2: Surgical details
7
Group I Group II Min= 3, Max=4 Min=3, Max=9 Number of spinal levels treated 4.94 (±3) 5.6 (±3.7) Length of hospital stay (days) 248.9 (±133.8) 266.9 (±144.2) OR time (skin-to-skin) (min) Estimated blood loss (ml) 120.8 (±45.4) 181.8 (±99.45)* Anterior approach 321.7 (±307.6) 525.7 (±516.98)* Posterior approach 367.7 (±308.6) 578.9(±518.5)* Total blood loss *Significant statistical difference between the two groups (p<0.05)
8 9 10
16 Page 16 of 18
1
Table 3: Spinal Level Treatment Details: Group I Group I # Posterior Levels Treated (Min; Max) Spinal Levels Treated (posterior approach) C1-C4 C1-C5 C1-C6 C1-C7 C2-C5 C2-C6 C2-C7 C3-C6 C3-C7 C4-C7 # Patients treated with both anterior/posterior approach
Min=3; Max=4 1 1 3 4 3 5 10 8 54 15 49
2 3 4
17 Page 17 of 18
1
Table 4: Spinal Level Treatment Details: Group II Group II # Posterior Levels Treated (Min; Max) Min=3; Max=9 Spinal Levels Treated (posterior approach) C1-T2 1 C1-T3 1 C2-T1 10 C2-T2 6 C2-T4 3 C3-T1 14 C3-T2 8 C3-T3 4 C4-T1 5 C4-T2 6 C5-T1 4 C5-T2 6 C5-T3 2 C5-T4 1 C6-T4 1 C7-T3 1 # Patients treated with both anterior/posterior approach 27
2 3
18 Page 18 of 18
S1529-9430(17)31003-3 https://doi.org/doi:10.1016/j.spinee.2017.09.010 SPINEE 57504
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
20-6-2017 28-8-2017 20-9-2017
Please cite this article as: Eeric Truumees, Devender Singh, Matthew J. Geck, John K. Stokes, Should long segment cervical fusions be routinely carried into the thoracic spine? –multi-center analysis, The Spine Journal (2017), https://doi.org/doi:10.1016/j.spinee.2017.09.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Should Long Segment Cervical Fusions Be Routinely Carried into the Thoracic Spine? –
2
Multi-center Analysis
3
Eeric Truumees, MD1; Devender Singh, PhD2; Matthew J. Geck, MD3, John K. Stokes, MD4
4 5 6
1
7
2
8 9
3
10
4
The University of Texas Dell Medical School, Seton Brain & Spine Institute: Professor of Surgery, CEO, Austin, TX Seton Spine & Scoliosis Center: Research Scientist, Austin, TX
The University of Texas Dell Medical School, Seton Spine & Scoliosis Center: Assistant Professor of Surgery, Austin, TX Seton Spine & Scoliosis Center: Neurosurgeon, Austin, TX
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Corresponding Author: Eeric Truumees, MD Professor of Surgery The University of Texas Dell Medical School CEO-Seton Brain & Spine Institute 1600 West 38th Street, Suite 200, Austin, Texas 78731-6400 Phone: 512 324 3580 Email: [email protected],
Conflicts of Interest and Source of Funding: The pilot phase of this study was funded by Globus Inc. Relevant financial disclosure outside the submitted work: board membership; royalties, stock/shareholder, ownership, consultancy. Abstract
1 Page 1 of 18
1
Background Context: While recommendations for caudal “end level” in posterior cervical
2
reconstruction remain highly variable, the benefits of routine extension of posterior cervical
3
fusions into the thoracic spine remain unclear.
4 5
Purpose: Compare clinical and radiographic outcomes in patients in whom posterior fusions
6
ended in the cervical spine versus those in whom the fusion was extended into the thoracic spine.
7 8
Study design/Setting: Multi-center retrospective analysis of prospectively followed patients.
9
Patient Sample: 177 adult spine patients undergoing 3 or more level posterior cervical fusions
10
for degenerative disease from January 2008 to May 2013.
11 12
Outcome Measures: Cervical lordosis, C2-C7 sagittal plumbline, T1 slope, visual analog scale
13
(VAS), oswestry disability index (ODI), rate of pseudarthrosis, length of hospital stay (LOS),
14
estimated blood loss (EBL) and operative time (OR).
15 16
Methods: We assembled a multicenter (4 sites) radiographic and clinical database of patients
17
that had undergone 3 or more level posterior cervical fusions for degenerative disease from
18
January 2008 to May 2013 with at least 2 years of post-operative (post-op) follow-ups. Patients
19
were divided into two groups: group I (fusion ending in the cervical spine) and group II (fusion
20
extending into the thoracic spine). All radiographic measurements were performed by an
21
independent experienced clinical researcher.
22 23
Results: Group I and Group II had 104 and 73 patients, respectively. Mean EBL for group II was
24
significantly higher than group I. Mean OR time and LOS were comparatively higher for group
25
II than group I but were not statistically significant (p>0.05). Mean cervical lordosis improved
26
post-operatively in both groups. There were no statistically significant differences in change or 2 Page 2 of 18
1
maintenance of mean cervical lordosis (2 wk vs. 2 year post-op) between the two groups
2
(p>0.05). Similarly, the change in mean C2-C7 sagittal plumbline and T1 slope was not
3
statistically significantly different between the two groups or with follow-up(p>0.05). Clinically,
4
significant improvements in VAS and ODI were noted in both groups from pre-op to final
5
follow-up, but the difference between groups was not statistically significant. While the rate of
6
pseudarthrosis was significantly higher in group I (21.2%) than group II (10.96%), there were no
7
statistically significant differences in adjacent segment degeneration or revision surgery rates
8
between the groups.
9 10 11
Conclusion: Both groups had similar clinical and radiographic outcomes. Extension of a
12
posterior cervical fusion into the thoracic spine leads to lower pseudarthrosis rate, while stopping
13
in the cervical spine yields lower EBL, OR and LOS, demonstrating that there are different
14
benefits for each approach. However, while the optimal end-level remains debatable, there are
15
scenarios in which upper thoracic extension should be considered. At this point, we recommend
16
extension of surgery in smokers and other patients at increased risk for pseudarthrosis as well as
17
in patients with anatomical limitations to strong C7 bone anchorage.
18 19
Keywords: Posterior cervical fusion, cervicothoracic fusion, cervicothoracic junction, outcomes,
20
multilevel posterior cervical fusion
21 22 23
Introduction 3 Page 3 of 18
1
While recommendations for caudal “end level” in posterior cervical reconstruction remain highly
2
variable, the benefits of routine extension of posterior cervical fusions into the thoracic spine
3
remain unclear. It has been suggested that long posterior cervical fusions risk subjacent
4
degeneration, spondylolisthesis, fracture, or kyphotic collapse when the fusion ends at C6 or C7.
5
In fact, ongoing neck pain and adjacent segment degeneration are two of the most common
6
sources of post-operative morbidity in patients undergoing long posterior cervical fusions
7
population.
8 9
Recommendations for appropriate caudal end vertebra vary widely (stop at C7 vs. extend to T1
10
through T4). Some biomechanical, but sparse clinical evidence supports these recommendations.
11
In a recently published study (2016), Schroeder and colleagues suggested that multilevel
12
posterior cervical fusions should be extended to T1 as long construct at C7 increases the rate of
13
revision (35.3% vs. 18.3%).1 Authors suggested that the odds of revisions in patients whose
14
construct terminated at C7 were 2.29 times greater than patients whose construct terminated at
15
T1 (p=0.02). Another study examining this research question was presented at various spine
16
meetings between 2009 ~2013 but has not yet been published.2-5 The authors reviewed cases
17
with minimum 3 levels posterior cervical fusion performed between 2000~2006 with at least 18
18
month follow-up. Radiographic and clinical outcomes in 36 patients with end-vertebrae
19
instrumentation in the cervical spine (EIV-C) were compared with 53 patients with a thoracic
20
end vertebra (EIV-T). The groups’ demographics were similar. Interestingly, despite the
21
additional levels exposed and fused in the thoracic group, operative time was the same. The rate
22
of symptomatic adjacent segment degeneration (ASD) requiring intervention was significantly
23
higher at the cranial adjacent level in the EIV-C group (28%) compared with the EIV-T group
24
(9%, p=0.04) at 2 years. Similarly, EIV-C patients had a significantly higher rate of caudal-level 4 Page 4 of 18
1
symptomatic ASD requiring intervention compared with EIV-T patients (39% vs 15%, p=0.01).
2
The development of caudal-level ASD was highest at C7 (41%), followed by C6 (40%), C5
3
(25%), T1 (18%), T3 (18%), T2 (17%), and T4 (0%). Despite higher rates of symptomatic ASD,
4
the overall complication rate and surgical revision rates were similar between the groups. Neck
5
disability index (NDI) outcomes at 2 years post-op were significantly better in the EIV-T group
6
(24.5 vs 34.0, p=0.05). The authors concluded that crossing the cervico-thoracic junction affords
7
protection of the adjacent levels without adding significant operative time or morbidity.
8 9
Based on the limited published data available, our primary objective was to compare clinical and
10
radiographic outcomes in patients in whom posterior fusions ended in the cervical spine versus
11
those in whom the fusion was extended into the thoracic spine. Our null hypothesis was that
12
extension of posterior cervical fusions into the upper thoracic spine improves clinical outcomes
13
while decreasing kyphosis.
14 15
Materials and Methods
16
The study was approved by the institutional review board. We assembled a multicenter (4 sites)
17
radiographic and clinical database of patients that had undergone 3 or more level posterior
18
cervical fusions for degenerative disease from January, 2008 to May, 2013 with at least 2 years
19
of post-operative (post-op) follow-ups. Patients were divided into two groups: group I (fusion
20
ending in the cervical spine) and group II (fusion extending into the thoracic spine). All
21
radiographic measurements were performed by an independent experienced clinical researcher.
22
Cervical lordosis was defined as the sagittal Cobb angle between C2 and C7 vertebral bodies. T1
23
slope was defined as the angle formed by the intersection of a line drawn tangential to the
24
superior end plate of T1 and a horizontal reference line. C2-C7 sagittal plumbline was defined as 5 Page 5 of 18
1
the distance between C2 plumbline and C7. For the analysis, bone grafts were divided into four
2
groups: local only; local and allografts; bone morphogenetic protein (BMP) only; and iliac crest
3
only. Current smokers included those patients smoking at the time of or within 6 months of their
4
surgery. The odds ratio was defined as the odds of smoker patients developing pseudarthrosis
5
compared to the odds of non-smoker patients developing pseudarthrosis while undergoing
6
multilevel posterior cervical fusions.
7 8
Statistical analysis
9
Characteristics of subjects were analyzed using means and standard deviations calculations. As
10
needed, two–sample t-test with unequal variances and paired t-tests were used to assess for
11
differences between the two groups. The non-normally distributed data were analyzed via the
12
Mann-Whitney’s U test (nonparametric). Analysis of variance (ANOVA) was used to investigate
13
the effects of type of bone grafts on the rate of pseudarthrosis. The level of significance was set
14
at α=0.05. All the statistical analyses were performed using SPSS software, version 15.0 (SPSS
15
Inc.).
16 17
Results
18
Group I and group II had 104 and 73 patients, respectively. The demography of the two groups
19
was similar as shown in Table 1. The minimum and maximum number of spinal levels treated
20
for group I and group II were 3 & 4 and 3 & 9, respectively. As shown in Table 2, mean total
21
estimated blood loss (EBL) for group II was significantly higher than group I (578.9±518.5 ml
22
vs. 367.7±308.6 ml, p<0.05). Mean operative time (266.9±144.2 min vs. 248.9±133.8 min) and
23
length of hospital stay (5.6±3.7 days vs. 4.94±3 days) were comparatively higher for group II
24
than group I but were not statistically significant (p>0.05). Tables 3 and 4 details the number of
6 Page 6 of 18
1
spinal levels treated for group I and II, respectively. As can be seen from these tables, C3-C7 and
2
C3-T1 were the most frequently treated levels in groups I and II, respectively. 47.6% of group I
3
patients were treated with both anterior and posterior approach. In group II, 37% of the patients
4
were treated with both anterior and posterior approach.
5 6
Mean cervical lordosis at 2 years post-op improved in both groups. As shown in Figure 1, there
7
was no significant statistical difference in change in mean cervical lordosis (2 wks. vs. 2 years
8
post-op)between the two groups (p>0.05).
9 10
Similarly, there were no significant statistical differences in change (2wks. vs. 2 years post-op)
11
in mean C2-C7 sagittal plumbline (group I: 2.9±1.87 mm; group II:2.7± 1.6mm) and T1 slope
12
(group I: 2.2±2.4˚; group II: 3.4±2.8˚) between the two groups (p>0.05).
13 14
There were significant improvements in mean clinical outcomes (i.e. visual analog scale (VAS)
15
and oswestry disability index (ODI) at 2 years follow ups in both groups but there were no
16
statistically significant differences between the two groups (p>0.05). Mean VAS improvements
17
(pre-op vs. 2 years post-op) in group I and group II were 5.7±2.2 vs. 3.5±3.1 and 5.5±2.5 vs.
18
3.2±2.6, respectively. Similarly, mean ODI improvements (pre-op vs. 2 years post-op) in group I
19
and group II were 39.2±22.1 vs. 26.2±20.2 and 44.9±20.1 vs. 29.6±19.7, respectively.
20 21
Revision Surgery
22
The all-cause revision rates at 2 years follow ups for Group I (ending in cervical spine) and
23
Group II (ending in thoracic spine) was 8.9% and 5.5%, respectively. This difference was not
24
statistically significant.
25
Pseudarthrosis 7 Page 7 of 18
1
Rate of pseudarthrosis was higher in group I (21.2%) than group II (10.96%). This difference
2
was statistically significant (p<0.05). Mean age of patients with pseudarthrosis in group I and
3
group II were 56(±9) and 67 (±4) years, respectively. Females had higher numbers of
4
pseudarthrosis than males (group I: 67% vs. 33%; group II: 55% vs. 45%; p<0.05). Overall,
5
53.3% of the patients with pseudarthrosis were current smokers. The rate of smoking in the solid
6
fusion group was 21.9%. The odds ratio of pseudarthrosis for a smoker compared with a non-
7
smoker was 4.071 (95% CI: 1.798-9.221). Mean number of spinal levels treated for patients
8
with pseudarthrosis in group I and group II were 3.6(±0.79) and 6.2(±2.5), respectively. Mean T1
9
slope for patients with pseudarthrosis increased significantly (2 wk vs. 2 year post-op) in both
10
groups (p<0.05). Both groups with pseudarthrosis had significantly higher mean C2-C7 sagittal
11
plumbline at 2 years follow-up (p<0.05). Mean cervical lordosis decreased in both groups with
12
pseudarthrosis (2 wk vs. 2 year post-op). The difference was not statistically significant (p>0.05).
13
Overall, ANOVA showed no significant effect of type of bone grafts on the rate of
14
pseudarthrosis (p>0.05).
15 16
Discussion
17
The anatomic and biomechanical complexities of posterior spine surgery crossing the
18
cervicothoracic junction, and other junctional regions, pose particular treatment challenges.6-12
19
While an extensive literature addresses questions of long term construct viability with long
20
fusions to or across the thoraco-lumbar and lumbo-sacral junctions, little supportive data
21
supports recommendations for an optimal caudal “end level” in posterior cervical reconstruction.
22
This study aimed to answer the research question should long segment cervical fusions be
23
routinely carried out into the thoracic spine. We hypothesized that extension of posterior cervical
8 Page 8 of 18
1
fusions into the upper thoracic spine would improve clinical outcome possible through optimized
2
sagittal balance and decreased adjacent segment degeneration and spondylolisthesis rates. Yet,
3
our data suggests that, except for pseudarthrosis rates, both groups had similar clinical and
4
radiographic outcomes. Clinically, significant but similar improvements in VAS and ODI were
5
noted in both groups at 2 years follow-up. The rate of revision surgery and supra- or subjacent
6
degeneration or instability was statistically similar between groups. Mean cervical lordosis, C2-
7
C7 sagittal plumbline and T1 slope improved in both groups at 2 years post-op follow up. As
8
expected, the increased levels incorporated in the reconstructions performed in Group II were
9
associated with greater EBL, OR time and LOS than group I.
10
Interestingly; pseudarthrosis was more frequently encountered in group I (21.2%) than group II
11
(10.96%). Females in both groups had significantly higher rate of pseudarthrosis than males. As
12
expected, majority (53.3%) of patients with pseudoarthrosis were current smokers. Based on
13
odds ratio, we estimated that current smoker patients undergoing multilevel posterior cervical
14
fusions were 4 times more likely to develop pseudarthrosis than non-smoker patients undergoing
15
similar surgical procedures. For both groups with pseudarthrosis, mean T1 slope, C2-C7 sagittal
16
plumbline and cervical lordosis at 2 years follow up differed significantly from those who did
17
not develop pseudarthrosis. We did not find any significant effect of type of bone grafts on the
18
rate of pseudarthrosis. We had anticipated that, given the infrequency with which anterior
19
column procedures are performed at C7-T1 and T1-2, the pseudarthrosis rate would be higher.
20
At this point, possible reasons for the decreased pseudarthrosis rate include detection issues.
21
Especially if the lower cervical anchor was a lateral mass screw, this implant may exhibit more
22
obvious failure in the form of loosening or posterior displacement. Such implant failure, in turn,
23
could alert the treating team to an otherwise radiographically undetectable pseudarthrosis. Other
9 Page 9 of 18
1
possibilities include greater surface area for the fusion mass afforded by extension into the
2
thoracic spine. Alternatively, the larger screws typically employed in the thoracic spine may
3
offer greater construct rigidity and, hence, a more stable mechanical environment in which to
4
achieve fusion when compared to lower cervical anchors, lateral mass screws in particular.
5 6
Schroeder and colleagues (2016) in a single center study with 219 patients favored extending the
7
multilevel cervical fusions to T1.1 They concluded that the multilevel cervical fusions should be
8
extended to T1 as stopping a long construct at C7 increases the rate of revision. Their study
9
divided the cohort into three groups (fusions terminating at C7, T1 and T2-T4). The rate of
10
revisions in C7, T1 and T2-T4 were 35.3%, 18.3% and 40%, respectively (p=0.008). The
11
multivariate analysis showed that the odds of revision in patients whose construct terminated at
12
C7 were 2.29 (1.16-4.61) times more than patients whose construct terminated at T1 (p=0.02).
13
The study did not find any significant difference in revision rate between stopping at T1 and T2-
14
T4.
15 16
In that our findings differ from Schroeder and colleagues, some key differences between the two
17
studies should be described: Our series is a multi-center (4 sites) study with 177 patients while
18
Schroeder’s study is a single center study with 219 patients. Both studies had similar patient
19
demographics and number of minimum posterior spinal levels treated. While Schroeder’s study
20
reported a larger sample size, only 31.5% of their patients completed minimum radiographic
21
follow ups (1 year). In contrast, 95% of our patients completed minimum radiographic follow
22
ups (2 years). Additionally, Schroeder’s study did not report any surgical variables (such as EBL,
23
OR and LOS) and/or patient reported outcomes such as VAS and ODI. We thoroughly
24
investigated these variables in our cohort. 10 Page 10 of 18
1
We also compared our study to that of Auerbach and colleagues. While the abstract has been
2
published in the transactions of spine meetings, we were unable to locate the full paper. Their
3
single center study, with smaller sample size of 89, found that crossing the cervico-thoracic
4
junction protects the adjacent levels without adding significant operative time or morbidity.2-5
5
We used this study to generate our null hypothesis. However, our multi-center series of 177
6
patients found similar clinical and radiographic outcomes in two groups.
7 8
Given the limited data available and the conflicting findings with Schroeder et al. and Auerbach
9
et al., we recommend that these results should be verified in prospective settings with additional
10
patients, longer follow-up and greater statistical power. Certain subgroups of patients may well
11
benefit from extension of their fusion into the upper thoracic spine. For example, those in whom
12
reliable fixation to the C7 pedicle is not available or those with greater need for sagittal balance
13
correction. But, for frailer, myelopathic patients, shorter segment fusion with appropriate bone
14
building and muscle strengthening might be preferred.
15 16
It is important to mention several limitations of our study. Among them are the retrospective
17
nature of the study, limited sample size and heterogeneous patient population. Additionally, we
18
did not conduct a multivariate analysis to account for any confounders (such as patient
19
comorbidities, etc.) in our findings. Nonetheless, we controlled all our radiographic
20
measurements as these were measured by a single experienced clinical researcher. All other
21
clinical data were thoroughly verified with other participating centers. However, the data set had
22
limitations. For example, in the setting of revision surgery, the indication for that revision was
23
not always clearly recorded or multiple indications for surgery were noted. These measures
24
eliminated the possibilities of any inter-observer bias in the data collection and analysis. Finally, 11 Page 11 of 18
1
95% of our cohort completed the minimum 2 years radiographic follow up as required by the
2
study. Of course, many of the potential advantages of extension of a long posterior cervical
3
fusion into the thoracic spine may be better explicated through longer term follow-up. We intend
4
to re-analyze this cohort at a minimum of 5 years post-operatively. Currently, however, this is
5
the longest follow-up interval reported. We must also acknowledge the limitations in the
6
radiographs available in detecting upper thoracic spondylolisthesis. Body habitus and the
7
shoulder girdle often limited fine detail assessment between C7 and T3. Additionally, the minor
8
rotation encountered in swimmer’s lateral views limited detection of minor degrees of subjacent
9
level anterolisthesis. Our prospective study includes use of EOS imaging to better detail these
10
aspects of cervico-thoracic alignment.
11 12
Conclusion
13
We conclude routine extension of long posterior cervical fusions into the thoracic spine does not
14
confer significant benefits in terms of cervical alignment or patient reported outcomes at two
15
year follow-up. Given that such extension is associated with higher EBL, OR and LOS, we
16
suggest that its use be limited in medically frail patients. On the other hand, crossing the cervico-
17
thoracic junction may be justified in smokers or other patients at increased risk for pseudarthrosis
18
as well as in patient in whom adequate lower cervical bone anchorage is not available.
19 20
Acknowledgement
21
We would like to thank the other sites for contributing their data for this study. We thank Globus
22
Inc. for funding the pilot phase of this study.
23
12 Page 12 of 18
1 2 3
References 1. Schroeder GD, Kepler CK, Kurd MF, Mead L, Millhouse PW, Kumar P, Nicholson K,
4
Stawicki C, Helber A, Fasciano D, Patel AA,
5
Anderson DG, Hilibrand AS, Vaccaro AR. Is it necessary to extend a multilevel
6
posterior cervical decompression and fusion to the upper thoracic spine? Spine 2016;
7
41(23); 1845-9
Woods BI, Radcliff KE, Rihn JA,
8
2. Auerbach J, Cho W, Riew KD. Crossing the cervicothoracic junction in multilevel
9
posterior cervical fusions reduces the rate of symptomatic adjacent segment breakdown.
10
37th Annual Meeting, poster, Cervical Spine Research Society (CSRS), 2009
11
3. Auerbach JD, Sehn JK, Cho W, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
12
MS, Riew KD. Crossing the cervicothoracic junction in multilevel posterior cervical
13
fusions reduces the rate of symptomatic adjacent segment breakdown. 17th Annual
14
Meeting, podium, International Meeting on Advanced Spine Techniques (IMAST), 2010
15
4. Auerbach JD, Cho W, Sehn JK, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
16
MS, Riew KD. Crossing the cervicothoracic junction in multilevel posterior cervical
17
fusions reduces the rate of symptomatic adjacent segment breakdown. 26th Annual
18
Meeting, podium, North American Spine Society (NASS), 2011
19
5. Cho W, Auerbach JD, Sehn JK, Milby AH, Crawford CH, O’Shaughnessy BA, Chang
20
MS, Riew KD: Crossing the Cervicothoracic Junction in Multilevel Posterior Cervical
21
Fusions Reduces the Rate of Symptomatic Adjacent Segment Breakdown. Poster,
22
American Academy of Orthopedic Surgeons (AAOS), 2013
23 24
6. An HS, Vaccaro A, Cotler JM, Lin S. Spinal disorders at the cervicothoracic junction. Spine 1994; 19:2557-64 13 Page 13 of 18
1
7. Inoue A, Ikata T, Katoh S. Spinal deformity following surgery for spinal cord tumors
2
and tumorous lesions: analysis based on an assessment of the spinal functional curve.
3
Spinal Cord 1996; 34:536-42
4 5 6 7 8 9
8. Stanescu S, Ebraheim NA, Yeasting R, Bailey AS, Jackson WT, Morphometric evaluation of the cervico-thoracic junction. Spine 1994; 19:2082-88 9. Bailey AS, Stanescu S, Yeasting RA, Ebraheim NA, Jackson T. Anatomic relationships of the cervicothoracic junction. Spine 1995; 20:1431-39 10. Chapman JR, Anderson PA, Pepin C, Toomey S, Newell DW, Grady MS. Posterior instrumentation of the unstable cervicothoracic spine. J Neurosurg 1996; 84;552-58
10
11. Albert TJ, Klein GR, Joffe D, Vaccaro AR. Use of cervicothoracic junction pedicle
11
screws for reconstruction of complex cervical spine pathology. Spine 1998; 23:1596-99
12
12. White AA III, Panjabi MM, Clinical Biomechanics of the Spine, ed 2. Philadelphia, PA,
13
JB Lippincott, 1990
14 15 16 17
14 Page 14 of 18
1 2
Figure 1: Mean (Standard deviation) cervical lordosis for the two groups at 2 week and 2 years post-op 20
Mean Cervical Lordosis (˚)
16 12
10.62 9.27
10.32
10.7
Group I Group II
8 4 0 3 4 5
2 wk post-op
2 years post-op
Note: There was no significant statistical difference in change in cervical lordosis (2 wk vs. 2 year post-op) between the two groups (p>0.05)
6 7 8
15 Page 15 of 18
1
Table 1: Demographics
2
Group I Group II 104 73 N 52/52 40/33 Male/Female 59.74 (±11.2) 59.16 (±13.6) Age (years) 28.3 (±6.04) 29.64 (±15.4)* BMI Smoking Status 30 (28.8%) 18 (24.7%) Smokers 60 (57.7%) 43 (58.9%) Non-Smokers 14 (13.5%) 12 (16.4%) Former Smokers (> 6 months) Ethnicity 65 63 Caucasian 22 4 Hispanic 10 3 African-American 7 3 Others * Significant statistical difference between the two groups (p<0.05)
3 4 5 6
Table 2: Surgical details
7
Group I Group II Min= 3, Max=4 Min=3, Max=9 Number of spinal levels treated 4.94 (±3) 5.6 (±3.7) Length of hospital stay (days) 248.9 (±133.8) 266.9 (±144.2) OR time (skin-to-skin) (min) Estimated blood loss (ml) 120.8 (±45.4) 181.8 (±99.45)* Anterior approach 321.7 (±307.6) 525.7 (±516.98)* Posterior approach 367.7 (±308.6) 578.9(±518.5)* Total blood loss *Significant statistical difference between the two groups (p<0.05)
8 9 10
16 Page 16 of 18
1
Table 3: Spinal Level Treatment Details: Group I Group I # Posterior Levels Treated (Min; Max) Spinal Levels Treated (posterior approach) C1-C4 C1-C5 C1-C6 C1-C7 C2-C5 C2-C6 C2-C7 C3-C6 C3-C7 C4-C7 # Patients treated with both anterior/posterior approach
Min=3; Max=4 1 1 3 4 3 5 10 8 54 15 49
2 3 4
17 Page 17 of 18
1
Table 4: Spinal Level Treatment Details: Group II Group II # Posterior Levels Treated (Min; Max) Min=3; Max=9 Spinal Levels Treated (posterior approach) C1-T2 1 C1-T3 1 C2-T1 10 C2-T2 6 C2-T4 3 C3-T1 14 C3-T2 8 C3-T3 4 C4-T1 5 C4-T2 6 C5-T1 4 C5-T2 6 C5-T3 2 C5-T4 1 C6-T4 1 C7-T3 1 # Patients treated with both anterior/posterior approach 27
2 3
18 Page 18 of 18