Key-words:
Anterior cervical discectomy and fusion - cervical fixation - single-level bracing
Introduction
Anterior cervical discectomy and fusion (ACDF) is performed approximately 132,000
times a year in the United States.[[1]] ACDF has wide-ranging indications, including cervical radiculopathy, myelopathy,
degenerative disc disease, and degenerative spondylolysis with similar or improved
fusion rates and low rates of subsidence compared to noninstrumented techniques.[[2]],[[3]],[[4]],[[5]],[[6]],[[7]],[[8]],[[9]] Frequently, patients are placed in a cervical brace postoperatively for a variable
amount of time. Since the popularization of anterior plating, the utility of postoperative
bracing has been debated, and there is a wide disparity in surgeon use.[[10]]
Proponents of postoperative bracing suggest that using an external orthosis will decrease
axial load on the construct and thereby reduce the risk of pseudarthrosis, graft subsidence,
and adjacent segment disease even in the presence of an anterior plate.[[11]],[[12]] Critics of bracing believe that the internal fixation afforded by an ACDF, in part
due to continuing improvements in hardware, obviates any requirement for external
bracing.[[13]] Biomechanical analyses of cervical orthoses have produced evidence to suggest that
bracing leads to a change in gait affecting patient quality of life or safety,[[14]],[[15]],[[16]] although this finding remains contentious.[[17]],[[18]],[[19]] Other studies have demonstrated that cervical braces are associated with dysphagia,[[20]] pressure ulcers,[[21]] and respiratory compromise.[[22]]
There is evidence to suggest that short-term use of cervical bracing following ACDF
correlates with improvements in pain and short-term patient-reported outcomes.[[23]] Yet other studies report no differences in surgical results.[[24]],[[25]] Although there are conflicting opinions, cervical bracing is regularly used in
the setting of single-level ACDF.[[26]],[[27]] The current data regarding implementation of postoperative cervical bracing are
somewhat equivocal and the current body of literature lacks adequate data for surgeon
decision-making as to the use of cervical bracing subsequent to single-level surgery.
We sought to clarify these uncertainties by analyzing short-term outcomes after single-level
ACDF cases and the associated direct costs related to cervical bracing.
Methods
Study population
In this Institutional Review Board (IRB) approved the study, patients undergoing single-level
ACDF surgical intervention across a university health system were enrolled retrospectively
from July 1, 2013 to June 30, 2017. A waiver of informed consent was granted by the
IRB as this study was considered to be minimal risk to patients. The Neurosurgery
Quality Improvement Initiative (NQII) EpiLog tool provided prospective data acquisition
on consecutive patients (n = 577) undergoing single-level ACDF between 2013 and 2017.
Briefly, the NQII EpiLog tool is a nonproprietary clinical research and quality improvement
architecture that was built and overlaid onto the electronic health record system
and enables prospective data collection.
Data collection
Patients' data were collected through the NQII EpiLog tool from the electronic health
record. Patient age, gender, race, and American Society of Anesthesiologists (ASA)
score that rates perfect health as 1 and moribund as 5, and multiple common medical
comorbidities, including diabetes, chronic obstructive pulmonary disease (COPD), coronary
artery disease (CAD), and total number of preoperative medical comorbidities, which
sums total adjacent diagnoses at the time of index surgery as surrogate for patient
disease severity were recorded [[Table 1]]. Length of stay (LOS) [[Table 2]], discharge disposition, emergency room (ER) visit within 30 days, and readmission
within 30 days were also recorded [[Figure 1]]. Of the total 577 patients included in the final study, a small cohort of patients
(n = 74) prospectively completed the EQ-5D-3 L questionnaire, a validated measure
of health outcomes for cost-utility analysis, to calculate quality-adjusted life years
(QALY). Total cost was calculated as all actual costs directly incurred by the hospital,
retrieved from billing databases [[Table 2]].
Table 1: Patient demographics and comorbidities
Table 2: Comparison of quality and cost of care
Figure 1: Short-term postoperative risk assessment
Statistical analysis
All continuous variables were assessed with the Student t-test or Wilcoxon rank-sum
test where appropriate. All categorical variables were analyzed with Pearson's Chi-square
test or Fisher's exact test. Multinomial logistic regression analyses were used to
determine disposition location based on independent variable of bracing. Significant
results were defined as P < 0.05. Averages are presented as mean ± standard deviation.
Results
Patient demographics
The retrospective cohort analysis consisted of patients who underwent single-level
ACDF (n = 577), where 509 patients were braced and 68 were unbraced [[Table 1]]. There was a significant difference in gender, as men were less frequently braced
than women (P = 0.017). Among the study population, there was no difference between
the braced and unbraced cohorts with regard to race (P = 0.299) or age (52 vs. 51
years, P = 0.540). There were no differences in patient comorbidities including diabetes,
COPD, CAD, obesity, smoking, body mass index, or total comorbidities. There were significant
differences in the overall physical status of patients as defined by the ASA score
with more ASA two patients in the braced cohort and more ASA three patients in the
unbraced cohort (P = 0.010). Among the study population, there was a difference in
graft type where unbraced patients received allograft more often than the braced patients
(P < 0.0001).
Safety of care
LOS was extended for the unbraced cohort compared to the braced cohort (152.7 ± 209.3
vs. 72.63 ± 112.5 h, P < 0.0001). Discharge disposition for the two populations differed,
with the braced cohort being 4.05 times more likely to be discharged home than to
an assisted rehabilitation facility or a skilled nursing facility in comparison to
the unbraced cohort (P < 0.001, confidence interval = 2.2771–7.2169). Readmissions
within 30 days (P = 0.828) and ER visits within 30 days were no differences between
the two groups (P = 1.000), as shown in [[Figure 1]].
Quality and cost-effectiveness
Patient QALY gain, at 3.7 months, was no difference between the two groups (P = 0.080).
Assessment of total costs also showed no difference between bracing and not bracing
patients following ACDF (P = 0.709) [[Table 2]].
Discussion
The present analyses show decreased LOS and increased the frequency of home discharge
for braced patients, but otherwise, indicate no difference in short-term outcomes
or cost between the braced and unbraced cohorts. The use of cervical orthoses remains
commonplace after cervical spine operation. Following early forms of ACDF and other
cervical spine surgeries where internal stabilization could not convincingly be achieved,
cervical bracing was commonly used.[[28]] With improvements in hardware, surgical technique, and efficacy of ACDF, one might
have expected the rate of postoperative bracing to decline. However, the paucity of
decisive literature has made informed decision-making challenging in the setting of
single-level ACDF.
The recent literature includes several studies that assess the utility of bracing
following instrumented or noninstrumented ACDF for multiple clinical indications.
Campbell et al. demonstrated that bracing following single-level allograft ACDF did
not confer any additional benefit in radiographic fusion or patient-reported quality
of life.[[29]] A meta-analysis performed by Camara et al. analyzed seven previous studies of which
only one showed that bracing was advantageous. The conclusions of the other six studies
were divided between no benefit or unknown benefit.[[30]] Another meta-analysis by Zhu et al. asked whether bracing was effective as measured
by patient-reported efficacy, radiographic outcomes, safety, and cost-effectiveness.
From five analyzed studies, there was no evidence to support bracing in improving
patient outcome scores, radiographic fusion rates, or lowering complication rates.[[31]]
Recent studies describe equivocal fusion rates in comparisons of braced and unbraced
patients following ACDF, but due to the historically mixed data and few prospective
studies, there is not yet sufficient evidence for a complete change in practice. In
an analysis of cervical motion following ACDF, it was found that motion was detec[[Table 2]]-week post-ACDF and that motion was independent of the number of levels fused.[[32]] As an example of how these biomechanics may affect outcomes, it is well established
that smoking has negative consequences on bone health and healing.[[33]] This translates into lower fusion rates, increased rates of pseudarthrosis, and
overall worse outcomes for these patients.[[34]],[[35]],[[36]],[[37]] For these patients, the above described postoperative motion may be a significant
contributor to poor surgical outcomes given a poor protoplasm, and these patients
may require the additional support of a cervical brace. The decision to brace may
thus be better approached in an individualized manner with consideration of specific
comorbidities and corresponding likelihood of complications for a given patient.
For a majority of the population, however, the current study is consistent with prior
work, indicating that bracing does not afford superior outcomes in the setting of
ACDF.[[24]],[[25]],[[29]],[[30]],[[31]] Although a previous trial demonstrated no radiographic or self-reported health
differences between braced and unbraced cohorts, the study assessed a subset of single-level
ACDF patients and did not investigate differences in follow-up care.[[29]] Here, we aimed to delineate the utility of postoperative bracing following single-level
ACDF and found no difference in short-term complication rates, ER visits, or 30-day
readmissions. These short-term results suggest that foregoing bracing after single-level
ACDF would not affect short-term outcomes. In this study, the variation in cervical
instrumentation and lack of specificity for bracing procedures are purposefully ill-defined.
Actual surgical practice includes inherent differences in surgical planning for each
case and the data used reflect this variability in practice for appropriate external
validity. Therefore, the results of this study further support the nonsuperiority
of bracing after single-level ACDF is consistent across varied operative choices.[[38]]
Interestingly, the braced patients were 4.05 times more frequently discharged to home
than the unbraced cohort. While the readmission and ER visit data do not suggest any
difference in complications, the difference in discharge disposition may suggest an
underlying difference in the early postoperative course of single-level ACDF patients.
The data also show that unbraced patients had an increased LOS as compared to the
braced cohort. Despite these differences, the total cost analysis suggests that this
increase in LOS did not result in any increase in direct cost to the hospital. The
costs of braces themselves are not included in the cost analysis. It is worth noting
that the popular cervical orthoses can range from $25 to $750 and restricting the
use of these devices can help decrease costs to patients and insurers.[[39]] Comparable, or improved, outcomes for patients while decreasing cost is a primary
goal of high-value care and safely removing bracing from postoperative care may provide
that value.[[40]]
The prospective data collection made possible by NQII EpiLog allowed for the analysis
of cervical bracing's effect on QALY gains. Bracing had no benefit on QALY gain, which
provides a secondary measure of operative success that demonstrates the noninferiority
of not bracing after single-level ACDF. These findings suggest that bracing may alter
the discharge disposition and time to discharge without changing the total cost of
care or sacrificing patient quality of life.
This study aimed to assess the efficacy of bracing in a broad and heterogeneous single-level
ACDF population. The study population was constructed such that surgeon-specific practices
of strictly bracing or not bracing patients following single-level ACDF removed bias
of patient selection, which was confirmed in the similarity between the two cohorts.
All patients received anterior plating, but graft type was not controlled for and
showed significant differences between the two cohorts. Yet despite this difference,
there was no effect on 30-day readmissions or ER visits between the two groups. This
analysis was not powered to determine if differences in graft-type or surgical choice
generated differential short-term outcomes in these patients, but again this study
intended to includes variability in patient presentation and surgical practice to
generalize results to all elective single-level ACDF cases.
Limitations of this study are related to the retrospective nature of the cohort analysis
despite the prospective data gathering tool used. There is potential bias in the data
recorded in the electronic health record that cannot be corrected. To this end, we
report data as means and standard deviations without removing outliers to prevent
selection bias. All patients included in the study are reflective of the spectrum
of complexity in ACDF cases seen at this institution. The QALY gain calculations consisted
of only 71 braced patients and 2 unbraced patients from the entire study population
of 577 patients. While this is a small sample size, we do not believe that it disqualifies
the results; the collection of QALY gains data began as a small pilot as a secondary
measure of clinical safety and cost-effectiveness that substantiates the primary outcome
measures mentioned above. The QALY gain evidence lends support to continuing this
data acquisition in a future prospective trial, which would provide the best way to
confirm these results.
Due to the imbalance in sample size between the two cohorts, we were unable to analyze
the dataset with the strictest statistical measures. The univariate analyses reported
above are representative of the relationship between bracing and patient outcomes
but are not able to incorporate preoperative variables in the analysis. We aimed to
expand the population to generate a propensity score-matched trial design to control
for the differing demographic variables and in preparation for a future randomized
control trial.
Conclusions
The use of cervical bracing following single-level ACDF remains a widespread practice
despite the growing evidence base, suggesting equivocal outcomes in this setting.
Here, we demonstrate that patients achieve similar outcomes with no difference in
direct hospital costs and a reduction in costs to the patient by removing the cervical
brace from their postoperative care. Future studies are needed to disentangle findings
regarding discharge disposition and LOS differences between these populations.
