- Email: [email protected]
Risk factors and pharmacologic prophylaxis for venous thromboembolism in elective spine surgery
Accepted Manuscript Title: Risk factors and pharmacologic prophylaxis for venous thromboembolism in elective spine surgery Author: Ryan P. McLynn, Pablo J. Diaz-Collado, Taylor D. Ottesen, Nathaniel T. Ondeck, Jonathan J. Cui, Patawut Bovonratwet, Blake N. Shultz, Jonathan N. Grauer PII: DOI: Reference:
S1529-9430(17)31064-1 https://doi.org/doi:10.1016/j.spinee.2017.10.013 SPINEE 57521
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
12-5-2017 1-9-2017 5-10-2017
Please cite this article as: Ryan P. McLynn, Pablo J. Diaz-Collado, Taylor D. Ottesen, Nathaniel T. Ondeck, Jonathan J. Cui, Patawut Bovonratwet, Blake N. Shultz, Jonathan N. Grauer, Risk factors and pharmacologic prophylaxis for venous thromboembolism in elective spine surgery, The Spine Journal (2017), https://doi.org/doi:10.1016/j.spinee.2017.10.013. 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.
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Risk Factors and Pharmacologic Prophylaxis for Venous Thromboembolism in Elective Spine Surgery
34
spine surgery, but prophylaxis guidelines are ambiguous for patients undergoing elective spine
35
surgery.
Ryan P. McLynn, BS a Pablo J. Diaz-Collado, MD a Taylor D. Ottesen, BS a Nathaniel T. Ondeck, BS a Jonathan J. Cui, BS a Patawut Bovonratwet, BS a Blake N. Shultz, BA a Jonathan N. Grauer, MD a a
Department of Orthopaedics and Rehabilitation, Yale School of Medicine, 47 College Street, New Haven, CT 06510, United States of America
Acknowledgements: We acknowledge Marilyn Hirsch and the Yale New Haven Hospital NSQIP Department for their contributions to data collection. Additionally, we acknowledge the James G. Hirsch, MD, Endowed Medical Student Research Fellowship for funding the study.
Corresponding author: Jonathan N. Grauer Department of Orthopaedics and Rehabilitation Yale School of Medicine 47 College Street, 2nd Floor New Haven, CT 06510, USA [email protected] Phone: (203) 737-7463 Fax: (203) 785-7132 Abstract BACKGROUND CONTEXT: Venous thromboembolism (VTE) is a known complication after
36 37
PURPOSE: To characterize the incidence and risk factors for VTE and the association of
38
pharmacologic prophylaxis with VTE and bleeding complications after elective spine surgery.
39
1
Page 1 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
STUDY DESIGN/SETTING: Retrospective cohort study of patients undergoing elective spine
2
surgery in the National Surgical Quality Improvement Program (NSQIP) database and a
3
retrospective cohort analysis at an academic medical center.
4 5
PATIENT SAMPLE: 109,609 patients in the NSQIP database from 2005-2014, and 2,855
6
patients at the authors’ institution from January 2013-March 2016, who underwent elective spine
7
surgery.
8 9
OUTCOME MEASURES: The incidence and risk factors for venous thromboembolism were
10
assessed in both cohorts based on NSQIP criteria. The incidence of bleeding complications
11
requiring reoperation was assessed based upon operative reports in the institutional cohort.
12 13
METHODS: Associations of patient and procedure factors with VTE were characterized in the
14
NSQIP population. In the single institution cohort, in addition to NSQIP variables, chart review
15
was completed to determine use of VTE prophylaxis, history of prior VTE, and incidence of
16
hematoma requiring reoperation. The association of patient and procedure variables, including
17
pharmacologic prophylaxis and history of prior VTE, with VTE and hematoma requiring
18
reoperation were determined with multivariate regression.
19 20
RESULTS: Among 109,609 elective spine surgery patients in NSQIP, independent risk factors
21
for VTE were greater age, male sex, increasing body mass index, dependent functional status,
22
lumbar spine surgery, longer operative time, perioperative blood transfusion, longer length of
23
stay, and other postoperative complications. There were 2,855 patients included in the
2
Page 2 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
institutional cohort. Pharmacologic prophylaxis was received by 56.3% of the institutional
2
patients, of whom 97.1% received unfractionated heparin. When controlling for patient and
3
procedural variables, pharmacologic prophylaxis did not significantly influence the rate of VTE,
4
but was associated with significant increase in hematoma requiring return to operating room
5
(RR=7.37; P=0.048).
6 7
CONCLUSIONS: Pharmacologic prophylaxis, primarily with unfractionated heparin, after
8
elective spine surgery was not associated with a significant reduction in VTE. However, there
9
was a significant increase in postoperative hematoma requiring reoperation among patients
10
receiving prophylaxis. This raises questions about routine use of unfractionated heparin for VTE
11
prophylaxis and supports the need for further consideration of risks and benefits of
12
chemoprophylaxis after elective spine surgery.
13 14
Keywords: Venous thromboembolism; spine surgery; elective; pharmacologic prophylaxis;
15
hematoma; ACS-NSQIP
16
3
Page 3 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Introduction
2
Venous thromboembolism (VTE), including deep vein thrombosis and pulmonary embolism, is a
3
known complication of spine surgery with potentially serious and possibly fatal consequences.
4
The incidence, risk factors, and management of VTE after total joint arthroplasty and
5
orthopaedic trauma have been well characterized, and evidence-based guidelines exist to aid in
6
clinical decision making.[1-6] Despite recent efforts to study VTE after spine surgery, the
7
literature has widely varying estimates of incidence and contradictory messages about the
8
efficacy of VTE prophylaxis for this population.[7-9]
9 10
The reported incidence of VTE after spine surgery has ranged from 0.3% to 31%, varying with
11
patient population, prophylaxis usage, and surveillance methods.[7] The potential benefit of
12
prophylactic anticoagulation in spine surgery patients must be carefully weighed against the
13
potential risk of epidural hematoma, which can be associated with neurological deficits and/or
14
wound drainage that can predispose to infection.[10-12]
15 16
Guidelines from the North American Spine Society and the American College of Chest
17
Physicians note that the balance of benefit and risk is unclear and merits further investigation,
18
often forcing surgeons to make a decision about pharmacologic prophylaxis without strong
19
evidence to guide them in terms of patient selection or choice of medication.[10,11] Overall,
20
they state that the decision is largely based on clinical judgment. Current guidelines recognize a
21
need for more studies to evaluate the efficacy of chemoprophylaxis and the optimal agent,
22
particularly in low-risk patients or elective surgeries.[10]
23
4
Page 4 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
In recent years, several studies have endeavored to use national databases to characterize
2
incidence and risk factors for VTE in large, nationally representative samples.[13-16] The
3
American College of Surgeons National Surgical Quality Improvement Program (NSQIP) is a
4
national database consisting of more than 600 hospitals that catalogs demographic, comorbidity,
5
intraoperative, and postoperative factors up to postoperative day 30 for various surgical
6
procedures.[17] Deep vein thrombosis and pulmonary embolism are postoperative adverse
7
events monitored for the NSQIP dataset. Advantages of NSQIP for the study of VTE include
8
large sample sizes and following patients until postoperative day thirty, regardless of discharge
9
status.[4]
10 11
Past investigations have characterized risk factors for VTE in the NSQIP sample.[15,16]
12
However, by including patients with high-risk profiles (e.g. trauma, emergency cases), these
13
studies may not accurately describe patients undergoing elective surgery, who may have a
14
different risk profile. Another limitation of these prior studies is that NSQIP does not offer data
15
on VTE prophylaxis or prior history of VTE.
16 17
The current study first aims to identify the incidence and risk factors for VTE after elective spine
18
surgery in the NSQIP population. The second aim is to perform a retrospective analysis of
19
NSQIP variables supplemented by chart review of elective spine surgery patients at a single large
20
academic medical center to study the association between inpatient pharmacologic prophylaxis
21
and the occurrence of postoperative VTE and bleeding complications. It was hypothesized that
22
chemoprophylaxis would lead to a reduced incidence of VTE without significantly increasing the
23
incidence of bleeding complications requiring reoperation.
24 5
Page 5 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2 3
Materials and Methods
4
A retrospective cohort study was conducted using NSQIP years 2005-2014. Patients undergoing
5
elective spine surgery were identified based on the following procedures and Current Procedural
6
Terminology (CPT) codes: anterior cervical discectomy and fusion (22551, 22554, 63075),
7
anterior cervical corpectomy (63081, 63085), cervical disc arthroplasty (22856, 22857), cervical
8
laminectomy (63015, 63045, 63265), cervical laminotomy (63020, 63040), posterior cervical
9
fusion (22595, 22600), posterior thoracic fusion (22610), thoracic laminectomy (63046, 63266),
NSQIP Cohort
10
thoracic corpectomy (63085, 63086), anterior lumbar fusion (22558), anterior lumbar
11
corpectomy (63087, 63088), lumbar laminectomy (63047, 63005, 63012, 63267), lumbar
12
laminotomy (63030, 63042), and posterior lumbar fusion (22612, 22630, 22633). Patients were
13
excluded for cases marked emergency or non-elective, fusion of seven or more levels (to exclude
14
deformity cases with increased risks[17,18]), missing data, and primary International
15
Classification of Diseases (ICD) diagnosis codes indicating trauma, tumor, or infection.
16 17
Demographic, comorbidity, intraoperative, and postoperative factors were characterized.
18
Demographic factors included age, sex, and body mass index. Comorbidity factors included
19
smoking status, diabetes, hypertension, chronic obstructive pulmonary disease, congestive heart
20
failure, preoperative renal failure, disseminated cancer, coagulopathy, preoperative anemia,
21
dependent functional status, and American Society of Anesthesiologists’ classification of III or
22
higher.
23
6
Page 6 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Intraoperative factors included operative time, procedure type (characterized as anterior cervical,
2
posterior cervical, thoracic, anterior lumbar, posterior lumbar, or combined anterior/posterior
3
lumbar), undergoing a multi-level procedure, and perioperative blood transfusion. Postoperative
4
factors included the presence of several postoperative complications (venous thromboembolism,
5
superficial surgical site infection, deep surgical site infection, pneumonia, mechanical ventilation
6
> 48 hours, unplanned reintubation, urinary tract infection, postoperative renal insufficiency or
7
failure, sepsis/septic shock, stroke, cardiac arrest, and myocardial infarction) and length of stay.
8
Of note, the database specifies that all thromboembolic events were confirmed with imaging and
9
required treatment with anticoagulation and/or inferior vena cava filter; however, because of the
10
varying practices at the many participating centers, it cannot be definitively known whether all
11
VTEs presented symptomatically or whether some were detected through routine imaging
12
screenings of all patients at some centers.[19]
13 14
Institutional Cohort
15
The elective spine surgery patients recorded for NSQIP at a single large academic medical
16
center, January 2013-March 2016, were identified. A retrospective cohort from this population
17
was identified utilizing the same inclusion and exclusion criteria as noted above for the NSQIP
18
cohort.
19 20
For each of these patients, NSQIP-recorded thirty postoperative day follow-up data was
21
available. In addition to the NSQIP-recorded variables, a retrospective chart review was
22
completed to examine for use of pharmacologic and mechanical VTE prophylaxis, history of
23
prior VTE, and incidence of hematoma requiring reoperation. The institutional data included
7
Page 7 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
only symptomatic VTEs (i.e. presenting with lower extremity pain or swelling for deep vein
2
thrombosis, dyspnea or chest pain for pulmonary embolism), consistent with NSQIP criteria
3
stated above,[19] and confirmed with venous duplex ultrasonography (in the case of DVT) or
4
computed tomography angiogram or ventilation/perfusion scan (in the case of PE).
5 6
Statistical Analysis
7
For the NSQIP cohort, multivariate Poisson regression with robust error variance was used to
8
test the association of the noted demographic, comorbidity, intraoperative, and postoperative
9
factors with development of venous thromboembolism within postoperative day 30.
10 11
The national and institutional samples were compared on the basis of age, sex, BMI, ASA class
12
≥ III, procedure type, length of stay, and incidence of VTE. Student’s t-test was used to compare
13
continuous variables, and chi-squared test was used to compare categorical variables.
14 15
To assess for variables that may influence prescription of pharmacologic prophylaxis in the
16
institutional cohort, a multivariate Poisson regression with robust error variance was used to test
17
the association of demographic, comorbidity (including history of VTE), and intraoperative
18
factors with whether or not a patient received prophylaxis.
19 20
Multivariate Poisson regression with robust error variance was used to test the association of
21
NSQIP variables, as well as history of prior VTE and pharmacologic prophylaxis status, with
22
incidence of venous thromboembolism within postoperative day 30, controlling for demographic,
23
comorbidity, intraoperative, and postoperative factors. The same multivariate regression analysis
8
Page 8 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
was repeated within certain subpopulations deemed to be at high risk for VTE based upon the
2
analysis.
3 4
Another multivariate Poisson regression with robust error variance was performed to test
5
association of demographic, comorbidity, and intraoperative variables, as well as pharmacologic
6
prophylaxis, with incidence of hematoma requiring reoperation.
7 8
The level of significance for all tests was set at P<0.05. Statistical tests were performed using
9
Stata® version 13.1 (StataCorp, LP, College Station, Texas, USA).
10 11
The current study was funded by the James G. Hirsch, MD, Endowed Medical Student Research
12
Fellowship ($4,000), provided by the Office of Student Research at the investigators’ institution.
13
There were no study-related conflicts of interest for any authors.
14
9
Page 9 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
Results NSQIP Cohort
3
In the national study, 109,609 patients were included. Initially, 123,320 patients were identified
4
based upon included CPT codes, of which 13,711 were excluded because of non-elective case,
5
fusion of 7+ levels, or diagnosis of tumor, trauma, or infection, or missing data. Demographics
6
of this population are presented in Table 1. The incidence of venous thromboembolism was
7
0.61% (672/109,609 patients).
8 9
The results of the multivariate analysis for independent risk factors for VTE are presented in
10
Table 2. Significant factors with the greatest relative risk were: stroke (relative risk (RR)=4.27,
11
P<0.001), pneumonia (RR=2.92, P<0.001), cardiac arrest (RR=2.62, P=0.015), anterior lumbar
12
procedure (RR=2.59, P<0.001), and superficial surgical site infection (RR=2.28, P=0.001).
13 14
Institutional Cohort
15
There were 2,855 patients included in the institutional study. Demographics of this population
16
are presented in Table 1. Initially, 3,254 patients were identified based upon included CPT codes,
17
of which 399 were excluded because of non-elective case, fusion of 7+ levels, or diagnosis of
18
tumor, trauma, or infection, or missing data. The incidence of venous thromboembolism was
19
1.23% (35/2,855 patients), which was significantly greater than the national cohort (0.61%,
20
P<0.001). There were also differences in BMI (national: 30.2±16.0 kg/m2, institutional 29.5±5.7;
21
P=0.027) and ASA class ≥ III (national: 39.9%, institutional: 34.4%; P<0.001). There were no
22
differences in age, sex, or length of stay.
23
10
Page 10 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
In the institutional study, all patients received mechanical prophylaxis with sequential
2
compression devices and early ambulation. Pharmacologic prophylaxis was received by 56.1%
3
of patients. Among those patients, the three most utilized medications were unfractionated
4
heparin (97.1%), low molecular weight heparin (2.5%), and warfarin (0.3%).
5 6
Variables significantly associated with receiving pharmacologic prophylaxis are delineated in
7
Table 3. The factors most significantly associated with receiving pharmacologic VTE
8
prophylaxis were: thoracic surgery (RR=1.62, P<0.001), anterior lumbar surgery (RR=1.32,
9
P<0.001), history of prior VTE (RR=1.21, P=0.006), ASA class ≥ III (RR=1.14, P<0.001), and
10
multi-level procedure (RR=1.10, P=0.007). Perioperative blood transfusion was associated with
11
decreased frequency of prophylaxis (RR=0.62, P<0.001), but cause and effect of this relationship
12
could not be determined.
13 14
In a univariate comparison, there was no significant association between pharmacologic
15
prophylaxis and incidence of VTE (RR=1.32, P=0.421; Figure 1).
16 17
In the multivariate analysis for associations with venous thromboembolism in the institutional
18
cohort, independent risk factors for VTE are delineated in Table 4. The factors most
19
significantly associated with VTE in order of decreasing relative risk were: urinary tract
20
infection (RR=16.05, P<0.001), history of prior venous thromboembolism (RR=9.48, P<0.001),
21
perioperative blood transfusion (RR=3.56, P=0.003), male sex (RR=2.64, P=0.018), and
22
increasing BMI (RR=1.08 per unit, P=0.030). Even with multivariate analysis, pharmacologic
23
prophylaxis was not found to significantly influence the rate of VTE (RR=0.68, P=0.424).
11
Page 11 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
The following subpopulations were analyzed based upon increased risk for VTE: age ≥ 65 years
3
(N=863), male sex (N=1,491), BMI ≥ 30 (N=1,218), lumbar procedure type (N=2,124), and
4
history of prior VTE (N=84). Following multivariate analysis, pharmacologic prophylaxis was
5
not associated with significant decrease in VTE incidence in any of the subgroups.
6 7
The incidence of postoperative bleeding or hematoma requiring return to the operating room was
8
0.4% (11/2,855). Ten of these were surgical site hematomas, while one patient experienced
9
bilateral cerebellar hematomas. Seven of the hematomas presented with significant neurological
10
deficits (63.6%), three presented with pain or wound drainage (27.3%), and one presented with
11
difficulty breathing (9.1%). Among 10 patients experiencing hematoma who received
12
pharmacologic prophylaxis, nine received unfractionated heparin and one received low
13
molecular weight heparin.
14 15
Incidence of postoperative hematoma requiring reoperation was significantly greater among
16
patients receiving prophylaxis (0.62% receiving prophylaxis versus 0.08% not receiving
17
prophylaxis; RR=7.80, P=0.020; Figure 1). In the multivariate analysis for associations with
18
postoperative hematoma (Table 5), the association with pharmacologic prophylaxis (RR=7.37,
19
P=0.048) remained significant. There was also a significant association with smoking (RR=3.16,
20
P=0.048).
21 22
12
Page 12 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Discussion
2
The incidence of VTE in the NSQIP population undergoing elective spine surgery was found to
3
be 0.61%. This is on the low end of incidences typically reported in the spine surgery literature,
4
including prior studies using NSQIP.[15,16] This relatively low incidence may have been found
5
because of the exclusion of high-risk diagnoses (e.g. trauma, neoplasm) that past studies have
6
shown to be associated with increased risk of VTE.[12,16,17] In contrast, the incidence of VTE
7
in the institutional cohort was significantly greater at 1.23%. It is unclear whether this variation
8
is due to patient factors, differences in management, or vigilance of screening for VTE.
9 10
Independent risk factors for VTE identified in both national and institutional analyses included
11
greater age, male sex, increasing BMI, perioperative blood transfusion, longer length of stay, and
12
urinary tract infection. Several variables—dependent functional status, lumbar surgery, greater
13
operative time, and various postoperative complications—identified as risk factors in the national
14
sample were not identified in the institutional cohort. It is not clear whether these differences are
15
due to actual differences between the populations versus effects that could not be identified due
16
to lesser sample size in the institutional sample. Of note, several postoperative complications
17
(e.g. stroke, cardiac arrest, surgical site infections) could not be assessed at the institutional level
18
because there were no incidences of VTE among patients also experiencing the complication.
19 20
Notably, no medical comorbidity was associated with increased risk of VTE in either the
21
national or institutional population. This contrasts with several past studies that have identified
22
conditions including cancer, hypertension, anemia, and high ASA class as risk factors.[13,15,16]
23
Other notable variables not associated with increased incidence of VTE were multilevel fusions
13
Page 13 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
and circumferential fusions (compared to posterior fusion), whereas some prior studies have
2
identified these as risk factors.[17,20]
3 4
Contrary to our hypothesis, pharmacologic prophylaxis did not significantly decrease the
5
incidence of venous thromboembolism, even after controlling for differences in
6
patient/procedural factors and patterns of prophylaxis usage. Further, in subgroup analysis
7
among populations with increased VTE risk (e.g. patients with prior VTE history), there was no
8
subgroup where prophylaxis was associated with a significant change in the incidence of VTE.
9
This is not out of the spectrum of findings in the literature where widely varying findings have
10
been reported concerning the efficacy of pharmacologic prophylaxis in preventing VTE in the
11
spine population.[7-11]
12 13
In a meta-analysis of VTE following elective surgery, Sansone and colleagues found a
14
significant reduction in deep vein thrombosis with pharmacologic prophylaxis compared to only
15
mechanical or no prophylaxis, though there were not enough pulmonary embolism cases to
16
evaluate.[12] Notably, they could not perform multivariate analysis to assess for potential
17
confounding variables because such data was not available in the original studies. Additionally,
18
the studies included in the meta-analysis utilized active screening for VTE, so many
19
asymptomatic cases were included. It is possible that there are different effects when comparing
20
clinically evident venous thromboembolic disease versus asymptomatic cases detected on
21
screening. In contrast, Schuster and colleagues conducted a systematic review of elective
22
thoracolumbar surgery that found no change in VTE risk with pharmacologic prophylaxis,
23
though the study was limited by small sample size.[9]
14
Page 14 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
Although one could question if greater power might have lead to identifying a significant
3
reduction of VTE in the current study, it is important to note that pharmacologic VTE
4
prophylaxis was associated with significant increase in incidence of bleeding complications
5
requiring return to the operating room, even after controlling for patient factors that may
6
influence bleeding risk or prescribing patterns.
7 8
While epidural hematoma and other bleeding complications are a commonly cited concern with
9
prophylactic anticoagulation after spine surgery, there is a lack of evidence to decisively
10
characterize the level of risk that prophylaxis poses and to identify patients where benefits
11
outweigh risks.[10] The current study suggests that pharmacologic prophylaxis with
12
unfractionated heparin (97.1% of prophylaxis used in the institutional cohort) was significantly
13
associated with increased risk for bleeding complications that require reoperation in elective
14
spine surgery. This supports the frequent assertion that these adverse events are an important
15
consideration before initiating prophylaxis after elective spine surgery.
16 17
For patients who experienced a bleeding complication, the use of chemoprophylaxis, timing of
18
first dose, and timing that the hematoma was recognized were assessed (Appendix 1). Of note,
19
three of the ten patients who received pharmacologic prophylaxis and subsequently sustained a
20
hematoma received their first dose in the evening of postoperative day zero. It is not clear
21
whether this represents a significant difference from patients who did not experience a
22
hematoma. However, there is some evidence that the risk of bleeding complications is increased
15
Page 15 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
with earlier initiation of chemoprophylaxis, so this finding may be consistent with existing
2
literature.[21]
3 4
While prevention of venous thromboembolism after spine surgery has been previously studied in
5
the literature, the current investigation makes several notable contributions. While prior studies
6
of chemoprophylaxis after elective spine surgery have generally featured sample sizes ranging
7
from 40-400,[7,21-23] the current study’s large sample size of nearly 3,000 enables more
8
accurate assessment of relatively rare events, VTE and postoperative hematoma. This may be
9
particularly relevant concerning hematomas, as the rate of postoperative bleeding complications
10
requiring reoperation was less than 1 per 250 patients, meaning that prior studies likely could not
11
study these events with adequate power. Additionally, the current study featured an internal
12
control group of patients who did not receive pharmacologic prophylaxis, meaning that
13
differences between the control and treatment groups could be identified and controlled for and
14
that both groups featured the same inclusion and exclusion criteria. In contrast, many prior
15
investigations have been case series where all patients received the same treatment and any
16
comparisons were to historical rates in other manuscripts, meaning population differences could
17
not be characterized.[7,21-23] These represent meaningful strengths over prior studies evaluating
18
the risks and benefits of pharmacologic VTE prophylaxis after spine surgery.
19 20
Compared to prior studies relying on the NSQIP database, the current study is novel in
21
combining national data and single-institution data that was supplemented with a targeted chart
22
review to investigate VTE after elective spine surgery. This approach combines the large sample
23
sizes, 30-day follow-up, and high-quality data abstraction by trained NSQIP reviewers with key
16
Page 16 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
variables for studying VTE that are not available in the database. The connection with NSQIP
2
data also provided a large number of other variables in order to control for potential associations
3
with which patients received prophylaxis and which developed a VTE or bleeding
4
complication.[19] To our knowledge, this is also the first study of VTE incidence using NSQIP
5
to focus solely on elective spine surgery, whereas prior studies have included all spine cases,
6
which could affect study results.[15,16] Additionally, compared to other database studies,
7
NSQIP provides specific variables to confirm whether cases are elective, meaning that the
8
current study offers a more accurate sample of elective cases compared to VTE studies that have
9
utilized other databases.
10 11
The current study is limited by several factors, most notably its retrospective, observational
12
nature. Randomized controlled trials are needed to better evaluate the role of prophylactic
13
anticoagulation in venous thromboembolism and its risk for bleeding complications. Another
14
limitation is that the overwhelming majority of patients on prophylaxis received unfractionated
15
heparin. Due to the retrospective design of the study, choice of prophylactic agent was entirely
16
due to individual surgeons’ practices. There is some evidence supporting low-molecular-weight
17
heparin as a safe option for anticoagulation after elective spine surgery, though this study utilized
18
a relatively small sample size and may have been underpowered, [21] and newer studies are
19
beginning to assess novel oral anticoagulants as well.[24] It was not possible to compare the
20
efficacy of various anticoagulants in the current study, and other anticoagulants may potentially
21
demonstrate superior results compared to unfractionated heparin; however they may also be
22
associated with bleeding risks/concerns, necessitating further study of these other agents.
23
17
Page 17 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
While the overall study benefited from a large sample size, some of the high-risk subgroups had
2
small sample sizes that may have limited the power to detect an effect of prophylaxis on VTE
3
incidence. Dedicated investigation of these subpopulations may potentially demonstrate a greater
4
benefit of anticoagulation compared to the entire population. Another limitation was the ability
5
to monitor for VTEs only to postoperative day 30. While the timing of VTE after elective spine
6
surgery has not been conclusively established, it is possible that some cases occurred beyond the
7
monitoring period, which could potentially influence results.
8 9
In conclusion, venous thromboembolism is a well-recognized and potentially dangerous
10
complication of elective spine surgery, creating a need for safe and effective prophylactic
11
measures. The current study found that pharmacologic prophylaxis predominantly with
12
unfractionated heparin did not significantly decrease the risk of VTE but significantly increased
13
the risk of bleeding events requiring reoperation. Current guidelines state that there is generally
14
insufficient evidence to support routine use of chemoprophylaxis in low-risk patients, and the
15
findings of the current study support the need for further risk/benefit considerations.
16
18
Page 18 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
References
2
[1]
Fitzgerald RH, Spiro TE, Trowbridge AA, Gardiner GA, Whitsett TL, O’Connell MB,
3
Ohar JA, Young TR. Prevention of venous thromboembolic disease following primary
4
total knee arthroplasty: A randomized, multicenter, open-label, parallel-group
5
comparison of enoxaparin and warfarin. J Bone Joint Surg Am 2001;83:900-906.
6
[2]
7 8
Geerts WH, Code KI, Jay RM, Chen E, Szalai JP. A prospective study of venous thromboembolism after major trauma. N Engl J Med 1994;331(24):1601-1606.
[3]
9
Sobieraj DM, Lee S, Coleman CI, Tongbram V, Chen W, Colby J, Kluger J, Makanji S, Ashaye AO, White CM. Prolonged versus standard-duration venous thromboprophylaxis
10
in major orthopedic surgery: A systematic review. Ann Intern Med 2012;156(10):720-
11
727.
12
[4]
Kester BS, Merkow RP, Ju MH, Peabody TD, Bentrem DJ, Ko CY, Bilimoria KY. 2014.
13
Effects of post-discharge venous thromboembolism on hospital quality comparisons
14
following hip and knee arthroplasty. J Bone Joint Surg Am 2014;96(17):1676-1684.
15
[5]
Falck-Ytter Y, Francis CW, Johanson NA, Curley C, Dahl OE, Schulman S, Ortel TL,
16
Pauker SG, Colwell CW. Prevention of VTE in orthopedic surgery patients:
17
Antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest
18
Physicians evidence-based clinical practice guidelines. Chest 2012;141(2_suppl):e278S-
19
e325S.
20
[6]
Jacobs JJ, Mont MA, Bozic KJ, Della Valle CJ, Goodman SB, Lewis CG, Yates AJ,
21
Boggio LN. Preventing venous thromboembolic disease in patients undergoing elective
22
hip and knee arthroplasty: Evidence-based guideline and evidence report. Rosemont, IL:
23
American Academy of Orthopaedic Surgeons; 2011.
19
Page 19 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[7]
2 3
Glotzbecker MP, Bono CM, Wood KB, Harris MB. Thromboembolic disease in spinal surgery: A systematic review. Spine (Phila Pa 1976) 2009;34(3):291-303.
[8]
Cox JB, Weaver KJ, Neal DW, Jacob RP, Hoh DJ. Decreased incidence of venous
4
thromboembolism after spine surgery with early multimodal prophylaxis: Clinical article.
5
J Neurosurg Spine 2014;21(4):677-684.
6
[9]
7 8
Schuster JM, Fischer D, Dettori JR. Is chemical antithrombotic prophylaxis effective in effective thoracolumbar spine surgery? Evid Based Spine Care J 2010;1(2):40-45.
[10]
9
Bono CM, Watters WC 3rd, Heggeness MH, Resnick DK, Shaffer WO, Baisden J, BenGalim P, Easa JE, Fernand R, Lamer T, Matz PG, Mendel RC, Patel RK, Reitman CA,
10
Toton JF. An evidence-based clinical guideline for the use of antithrombotic therapies in
11
spine surgery. Spine J 2009;9(12):1046-1051.
12
[11]
Gould MK, Garcia DA, Wren SM, Karanicolas PJ, Arcelus JI, Heit JA, Samama CM.
13
Prevention of VTE in nonorthopedic surgical patients: Antithrombotic therapy and
14
prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based
15
clinical practice guidelines. Chest 2012;141(2):e227S-e277S.
16
[12]
Sansone JM, del Rio AM, Anderson PA. The prevalence of and specific risk factors for
17
venous thromboembolic disease following elective spine surgery. J Bone Joint Surg Am
18
2010;92(2):304-313.
19
[13]
Fineberg SJ, Oglesby M, Patel AA, Pelton MA, Singh K. The incidence and mortality of
20
thromboembolic events in lumbar spine surgery. Spine (Phila Pa 1976)
21
2013;38(13):1154-1159.
20
Page 20 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[14]
Oglesby M, Fineberg SJ, Patel AA, Pelton MA, Singh K. The incidence and mortality of
2
thromboembolic events in cervical spine surgery. Spine (Phila Pa 1976)
3
2013;38(9):E521-527.
4
[15]
Sebastian AS, Currier BL, Kakar S, Nguyen EC, Wagie AE, Habermann ES, Nassr A.
5
Risk factors for venous thromboembolism following thoracolumbar surgery: Analysis of
6
43,777 patients from the American College of Surgeons National Surgical Quality
7
Improvement Program 2005 to 2012. Global Spine J 2016;6(8):738-743.
8
[16]
9
HJ, Vates GE. Risk factors associated with venous thromboembolism in patients
10 11
undergoing spine surgery. J Neurosurg Spine 2017;26(1):90-6. [17]
12 13
[18]
[19]
User Guide for the 2015 ACS NSQIP Participant Use Data File. Chicago: American College of Surgeons; 2016
[20]
18 19
Cheng JS, Arnold PM, Anderson PA, Fischer D, Dettori JR. Anticoagulation risk in spine surgery. Spine (Phila Pa 1976) 2010;35(9 Suppl):S117-24.
16 17
Schairer WW, Pedtke AC, Hu SS. Venous thromboembolism after spine surgery. Spine (Phila Pa 1976) 2014 [Epub ahead of print].
14 15
Piper K, Algattas H, DeAndrea-Lazarus IA, Kimmell KT, Li YM, Walter KA, Silberstein
Brambilla S, Ruosi C, La Maida GA, Caserta S. Prevention of venous thromboembolism in spinal surgery. Eur Spine J 2004;13(1):1-8.
[21]
Strom RG, Frempong-Boadu A. Low-molecular-weight heparin prophylaxis 24 to 36
20
hours after degenerative spine surgery: risk of hemorrhage and venous
21
thromboembolism. Spine (Phila Pa 1976) 2013;38(23):E1498-1502.
22 23
[22]
Al-Dujaili TM, Majer CN, Madhoun TE, Kassis SZ, Saleh AA. Deep venous thrombosis in spine surgery patients: incidence and hematoma formation. Int Surg 2012;97(2):150-4.
21
Page 21 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[23]
Spinal Cord Injury Thromboprophylaxis Investigators. Prevention of venous
2
thromboembolism in the rehabilitation phase after spinal cord injury: prophylaxis with
3
low-dose heparin or enoxaparin. J Trauma 2003;54(6):1111-5.
4
[24]
Du W, Zhao C, Wang J, Liu J, Shen B, Zheng Y. Comparison of rivaroxaban and
5
parnaparin for preventing venous thromboembolism after lumbar spine surgery. J Orthop
6
Surg Res 2015;10:78.
7 8 9 10 11
Figure 1. Relative risk of experiencing venous thromboembolism or hematoma based on
12
prophylaxis status, with or without controlling for other variables. Diamond indicates relative
13
risk; bar indicates 95% confidence interval. Black shading indicates that prophylaxis is
14
significant at P < 0.05; gray shading indicates that variable is not significant. VTE=Venous
15
thromboembolism; OR=operating room.
16 17
22
Page 22 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 1. Demographics
Number of patients
Age, years
National
Institutional
109,609
2,855
56.4 ± 14.4
56.0 ± 14.7
Gender
0.125
0.815
Female
52,628 (48.0%)
1,364 (47.8%)
Male
56,931 (52.0%)
1,491 (52.2%)
30.2 ± 16.0
29.5 ± 5.7
0.027
39.9%
34.4%
<0.001
(43,758/109,609)
(983/2,855)
Length of stay, days
2.4 ± 5.3
2.4 ± 2.9
0.633
Incidence of venous
0.61%
1.23%
<0.001
(672/109,609)
(35/2,855)
Body Mass Index, kg/m2
ASA Class ≥ III
thromboembolism
2 3
P-value
ASA = American Society of Anesthesiologists
4 5
23
Page 23 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 2. Multivariate relative risk for venous thromboembolism in national cohort Risk Factor
N (VTE)
N (risk
Multivariate Relative Risk#
P-value
factor) Age, years
‡
‡
1.02 (1.01-1.03)
<0.001
Male sex
‡
‡
1.28 (1.09-1.49)
0.002
BMI, per unit
‡
‡
1.001 (1.000-1.001)
0.005
Smoking
83
26,316
0.56 (0.45-0.71)
<0.001
Diabetes mellitus
118
16,767
0.81 (0.63-1.03)
0.089
Insulin-dependent diabetes
40
5,371
0.94 (0.64-1.42)
0.776
Hypertension
423
53,295
1.06 (0.90-1.28)
0.515
COPD
39
4,193
1.11 (0.80-1.57)
0.549
Congestive heart failure
2
220
0.51 (0.14-1.85)
0.320
Preop renal disease
1
243
0.16 (0.02-1.65)
0.124
Disseminated cancer
2
130
0.97 (0.24-4.61)
0.973
Coagulopathy
18
1,431
1.33 (0.83-2.10)
0.227
Preoperative anemia
126
19,676
0.87 (0.71-1.08)
0.200
Dependent functional status
39
2,343
1.71 (1.24-2.43)
<0.001
ASA class ≥ III
364
43,758
1.12 (0.95-1.34)
0.209
Medical Comorbidities
Procedure type
<0.001
Anterior cervical
77
24,499
Reference
Posterior cervical
28
4,672
1.32 (0.88-2.09)
0.207
Thoracic
9
644
0.99 (0.40-2.54)
0.981
Anterior lumbar
38
3,133
2.59 (1.75-3.87)
<0.001
Posterior lumbar
479
73,798
1.46 (1.14-1.89)
0.003
Anterior/posterior lumbar
41
2,863
2.18 (1.47-3.26)
<0.001
24
Page 24 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
Intraoperative Factors ‡
‡
1.003 (1.002-1.003)
<0.001
Multi-level procedure
307
35,670
1.11 (0.95-1.32)
0.228
Periop blood transfusion
124
5,299
1.78 (1.38-2.24)
<0.001
Superficial SSI
19
901
2.28 (1.39-3.63)
0.001
Deep/organ space SSI
25
706
2.19 (1.30-3.65)
0.003
Pneumonia
41
521
2.92 (1.66-5.05)
<0.001
Failure to wean from
30
238
1.41 (0.67-3.23)
0.381
Unplanned reintubation
30
348
1.32 (0.63-2.78)
0.459
Urinary tract infection
44
1,160
2.07 (1.40-3.07)
<0.001
Renal insufficiency/failure
12
155
1.99 (0.94-4.01)
0.063
Sepsis/septic shock
48
612
1.95 (1.17-3.20)
0.009
Stroke
11
98
4.27 (2.00-9.05)
<0.001
Cardiac arrest
11
115
2.62 (1.20-5.60)
0.015
Myocardial infarction
11
193
1.55 (0.70-3.32)
0.267
‡
‡
1.01 (1.01-1.01)
<0.001
Operative time, minutes
Postoperative Complications
ventilator
Length of stay, days
1
# Relative risk (95% confidence interval)
2
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; SSI = Surgical site infection; ASA =
3
American Society of Anesthesiologists. Shading indicates statistical significance at P < 0.05.
4
N(VTE) indicates number of patients with each risk factor who experienced a VTE.
5
N(risk factor) indicates total number of patients with given risk factor, out of total N = 109,609.
6
‡ indicates a continuous variable, so all patients are included.
25
Page 25 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 3. Multivariate regression for significant associations with receiving pharmacologic prophylaxis Multivariate Relative Risk#
P-value
Age, years
1.01 (1.006-1.012)
<0.001
Female sex
1.09 (1.03-1.17)
0.004
0.99 (0.988-0.999)
0.044
Smoking
1.09 (1.001-1.17)
0.048
Diabetes mellitus
0.93 (0.84-1.03)
0.172
Insulin-dependent diabetes
1.04 (0.88-1.22)
0.680
Hypertension
0.97 (0.90-1.04)
0.347
COPD
0.95 (0.80-1.12)
0.536
Congestive heart failure
1.08 (0.46-2.51)
0.861
Preop renal disease
1.10 (0.71-1.69)
0.669
Disseminated cancer
0.90 (0.49-1.66)
0.732
Coagulopathy
0.87 (0.66-1.14)
0.304
Preoperative anemia
1.00 (0.89-1.13)
0.966
status
1.14 (0.88-1.48)
0.332
ASA class ≥ III
1.14 (1.06-1.22)
<0.001
History of prior VTE&
1.21 (1.06-1.39)
0.006
Risk Factor
BMI, per unit
Medical Comorbidities
Dependent functional
Procedure type Anterior cervical
Reference
Posterior cervical
1.19 (0.99-1.43)
0.063
Thoracic
1.62 (1.26-2.09)
<0.001
Anterior lumbar
1.32 (1.16-1.50)
<0.001
Posterior lumbar
1.08 (0.99-1.18)
0.094
26
Page 26 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism A/P lumbar
0.97 (0.82-1.14)
0.689
1.003 (1.002-1.003)
<0.001
Multi-level procedure
1.10 (1.02-1.17)
0.007
Periop blood transfusion
0.62 (0.54-0.72)
<0.001
Intraoperative Factors Operative time, minutes
1
#
2
&
3
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; ASA = American Society of
4
Anesthesiologists.
5
Shading indicates statistical significance at P < 0.05.
Relative risk (95% confidence interval) Chart review variable
6 7 8
27
Page 27 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 4. Multivariate relative risk for venous thromboembolism in institutional cohort, including chart review N (VTE)
N (risk factor)
Multivariate Relative Risk#
P-value
Age, years
‡
‡
1.07 (1.03-1.11)
<0.001
Male sex
‡
‡
2.64 (1.18-5.90)
0.018
BMI, per unit
‡
‡
1.08 (1.01-1.15)
0.030
Smoking
5
540
1.12 (0.45-3.02)
0.813
Diabetes mellitus
4
403
0.41 (0.12-1.23)
0.178
Insulin-dependent diabetes
2
127
1.62 (0.33-11.08)
0.604
Hypertension
20
1,238
0.45 (0.21-0.94)
0.037
COPD
2
79
1.68 (0.50-5.30)
0.369
Congestive heart failure
0
3
N/A
Preop renal disease
0
8
N/A
Disseminated cancer
0
5
N/A
Coagulopathy
2
36
0.88 (0.27-3.18)
0.852
Preoperative anemia
3
182
0.80 (0.27-2.30)
0.690
Dependent functional status
0
35
N/A
ASA class ≥ III
15
983
0.74 (0.29-1.76)
Risk Factor
Medical Comorbidities
Procedure type
0.483
0.242
Anterior cervical
3
632
Reference
Posterior cervical
1
85
1.65 (0.15-18.43)
Thoracic
0
14
N/A
Anterior lumbar
5
158
3.28 (0.68-15.74)
0.138
Posterior lumbar
22
1,858
1.21 (0.37-3.93)
0.757
A/P lumbar
4
108
3.04 (0.70-13.16)
0.138
0.684
28
Page 28 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism Intraoperative Factors Operative time, minutes
‡
‡
1.004 (0.999-1.008)
0.059
Multi-level procedure
10
798
0.98 (0.43-2.42)
0.963
Periop blood transfusion
10
137
3.56 (1.74-7.94)
0.003
Superficial SSI
0
27
N/A
Deep/organ space SSI
0
16
N/A
Pneumonia
3
26
3.29 (0.87-15.59)
0.162
Failure to wean from
2
5 N/A
N/A
Postoperative Complications
ventilator Unplanned reintubation
1
6
5.04 (0.32-137.15)
0.353
Urinary tract infection
3
28
16.05 (3.48-49.74)
<0.001
Renal insufficiency/failure
1
5
7.27 (0.86-51.35)
0.070
Sepsis/septic shock
1
16
7.16 (0.16-56.78)
0.178
Stroke
0
3
N/A
Cardiac arrest
0
5
N/A
Myocardial infarction
0
5
N/A
Length of stay, days
‡
‡
1.07 (1.04-1.09)
<0.001
History of prior VTE&
9
84
9.48 (3.07-28.23)
<0.001
Pharmacologic prophylaxis&
22
1,603
0.68 (0.25-1.71)
0.424
1
N(VTE) indicates number of patients with each risk factor who experienced a VTE. N(risk factor) indicates total
2
number of patients with given risk factor, out of total N = 2,855. # indicates relative risk (95% confidence interval).
3
‡ indicates continuous variable, so all patients are included. &Chart review variables. N/A indicates variable could
4
not be analyzed because N(VTE) = 0 or due to covariance. BMI = Body mass index; COPD = Chronic obstructive
5
pulmonary disease; SSI = Surgical site infection; ASA = American Society of Anesthesiologists.
6
29
Page 29 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 5. Multivariate relative risk for postoperative hematoma in institutional cohort Multivariate Relative Risk#
P value
Age
1.02 (0.99-1.05)
0.188
Male sex
2.89 (0.71-11.85)
0.140
BMI
0.93 (0.84-1.03)
0.154
N/A
N/A
Smoking
3.16 (1.01-9.88)
0.048
Diabetes mellitus
0.93 (0.17-5.18)
0.937
Hypertension
0.54 (0.18-1.65)
0.282
N/A
N/A
Coagulopathy
5.69 (0.50-64.39)
0.160
Preoperative anemia
0.87 (0.12-6.48)
0.894
ASA class ≥ III
3.05 (0.96-9.71)
0.059
Operative time
1.00 (0.99-1.01)
0.301
Multi-level procedure
2.58 (0.71-9.42)
0.151
Perioperative blood transfusion
2.53 (0.36-17.96)
0.353
Pharmacologic prophyaxis&
7.37 (1.01-53.46)
0.048
Medical Comorbidities Dependent functional status
COPD
Intraoperative Factors
2
#
3
&
4
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; ASA = American Society of
5
Anesthesiologists.
6
Shading indicates statistical significance at P < 0.05.
Relative risk (95% confidence interval) Chart review variable
7
30
Page 30 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
31
Page 31 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2 3
Affirmation of Authorship - VTE Spine JPEG.jpg
32
Page 32 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2 3
VTE Prophylaxis Spine - Figure 1 jpeg.jpg
33
Page 33 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2
VTE Prophylaxis Spine - Flow Diagram Figure.jpg
34
Page 34 of 34
S1529-9430(17)31064-1 https://doi.org/doi:10.1016/j.spinee.2017.10.013 SPINEE 57521
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
12-5-2017 1-9-2017 5-10-2017
Please cite this article as: Ryan P. McLynn, Pablo J. Diaz-Collado, Taylor D. Ottesen, Nathaniel T. Ondeck, Jonathan J. Cui, Patawut Bovonratwet, Blake N. Shultz, Jonathan N. Grauer, Risk factors and pharmacologic prophylaxis for venous thromboembolism in elective spine surgery, The Spine Journal (2017), https://doi.org/doi:10.1016/j.spinee.2017.10.013. 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.
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Risk Factors and Pharmacologic Prophylaxis for Venous Thromboembolism in Elective Spine Surgery
34
spine surgery, but prophylaxis guidelines are ambiguous for patients undergoing elective spine
35
surgery.
Ryan P. McLynn, BS a Pablo J. Diaz-Collado, MD a Taylor D. Ottesen, BS a Nathaniel T. Ondeck, BS a Jonathan J. Cui, BS a Patawut Bovonratwet, BS a Blake N. Shultz, BA a Jonathan N. Grauer, MD a a
Department of Orthopaedics and Rehabilitation, Yale School of Medicine, 47 College Street, New Haven, CT 06510, United States of America
Acknowledgements: We acknowledge Marilyn Hirsch and the Yale New Haven Hospital NSQIP Department for their contributions to data collection. Additionally, we acknowledge the James G. Hirsch, MD, Endowed Medical Student Research Fellowship for funding the study.
Corresponding author: Jonathan N. Grauer Department of Orthopaedics and Rehabilitation Yale School of Medicine 47 College Street, 2nd Floor New Haven, CT 06510, USA [email protected] Phone: (203) 737-7463 Fax: (203) 785-7132 Abstract BACKGROUND CONTEXT: Venous thromboembolism (VTE) is a known complication after
36 37
PURPOSE: To characterize the incidence and risk factors for VTE and the association of
38
pharmacologic prophylaxis with VTE and bleeding complications after elective spine surgery.
39
1
Page 1 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
STUDY DESIGN/SETTING: Retrospective cohort study of patients undergoing elective spine
2
surgery in the National Surgical Quality Improvement Program (NSQIP) database and a
3
retrospective cohort analysis at an academic medical center.
4 5
PATIENT SAMPLE: 109,609 patients in the NSQIP database from 2005-2014, and 2,855
6
patients at the authors’ institution from January 2013-March 2016, who underwent elective spine
7
surgery.
8 9
OUTCOME MEASURES: The incidence and risk factors for venous thromboembolism were
10
assessed in both cohorts based on NSQIP criteria. The incidence of bleeding complications
11
requiring reoperation was assessed based upon operative reports in the institutional cohort.
12 13
METHODS: Associations of patient and procedure factors with VTE were characterized in the
14
NSQIP population. In the single institution cohort, in addition to NSQIP variables, chart review
15
was completed to determine use of VTE prophylaxis, history of prior VTE, and incidence of
16
hematoma requiring reoperation. The association of patient and procedure variables, including
17
pharmacologic prophylaxis and history of prior VTE, with VTE and hematoma requiring
18
reoperation were determined with multivariate regression.
19 20
RESULTS: Among 109,609 elective spine surgery patients in NSQIP, independent risk factors
21
for VTE were greater age, male sex, increasing body mass index, dependent functional status,
22
lumbar spine surgery, longer operative time, perioperative blood transfusion, longer length of
23
stay, and other postoperative complications. There were 2,855 patients included in the
2
Page 2 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
institutional cohort. Pharmacologic prophylaxis was received by 56.3% of the institutional
2
patients, of whom 97.1% received unfractionated heparin. When controlling for patient and
3
procedural variables, pharmacologic prophylaxis did not significantly influence the rate of VTE,
4
but was associated with significant increase in hematoma requiring return to operating room
5
(RR=7.37; P=0.048).
6 7
CONCLUSIONS: Pharmacologic prophylaxis, primarily with unfractionated heparin, after
8
elective spine surgery was not associated with a significant reduction in VTE. However, there
9
was a significant increase in postoperative hematoma requiring reoperation among patients
10
receiving prophylaxis. This raises questions about routine use of unfractionated heparin for VTE
11
prophylaxis and supports the need for further consideration of risks and benefits of
12
chemoprophylaxis after elective spine surgery.
13 14
Keywords: Venous thromboembolism; spine surgery; elective; pharmacologic prophylaxis;
15
hematoma; ACS-NSQIP
16
3
Page 3 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Introduction
2
Venous thromboembolism (VTE), including deep vein thrombosis and pulmonary embolism, is a
3
known complication of spine surgery with potentially serious and possibly fatal consequences.
4
The incidence, risk factors, and management of VTE after total joint arthroplasty and
5
orthopaedic trauma have been well characterized, and evidence-based guidelines exist to aid in
6
clinical decision making.[1-6] Despite recent efforts to study VTE after spine surgery, the
7
literature has widely varying estimates of incidence and contradictory messages about the
8
efficacy of VTE prophylaxis for this population.[7-9]
9 10
The reported incidence of VTE after spine surgery has ranged from 0.3% to 31%, varying with
11
patient population, prophylaxis usage, and surveillance methods.[7] The potential benefit of
12
prophylactic anticoagulation in spine surgery patients must be carefully weighed against the
13
potential risk of epidural hematoma, which can be associated with neurological deficits and/or
14
wound drainage that can predispose to infection.[10-12]
15 16
Guidelines from the North American Spine Society and the American College of Chest
17
Physicians note that the balance of benefit and risk is unclear and merits further investigation,
18
often forcing surgeons to make a decision about pharmacologic prophylaxis without strong
19
evidence to guide them in terms of patient selection or choice of medication.[10,11] Overall,
20
they state that the decision is largely based on clinical judgment. Current guidelines recognize a
21
need for more studies to evaluate the efficacy of chemoprophylaxis and the optimal agent,
22
particularly in low-risk patients or elective surgeries.[10]
23
4
Page 4 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
In recent years, several studies have endeavored to use national databases to characterize
2
incidence and risk factors for VTE in large, nationally representative samples.[13-16] The
3
American College of Surgeons National Surgical Quality Improvement Program (NSQIP) is a
4
national database consisting of more than 600 hospitals that catalogs demographic, comorbidity,
5
intraoperative, and postoperative factors up to postoperative day 30 for various surgical
6
procedures.[17] Deep vein thrombosis and pulmonary embolism are postoperative adverse
7
events monitored for the NSQIP dataset. Advantages of NSQIP for the study of VTE include
8
large sample sizes and following patients until postoperative day thirty, regardless of discharge
9
status.[4]
10 11
Past investigations have characterized risk factors for VTE in the NSQIP sample.[15,16]
12
However, by including patients with high-risk profiles (e.g. trauma, emergency cases), these
13
studies may not accurately describe patients undergoing elective surgery, who may have a
14
different risk profile. Another limitation of these prior studies is that NSQIP does not offer data
15
on VTE prophylaxis or prior history of VTE.
16 17
The current study first aims to identify the incidence and risk factors for VTE after elective spine
18
surgery in the NSQIP population. The second aim is to perform a retrospective analysis of
19
NSQIP variables supplemented by chart review of elective spine surgery patients at a single large
20
academic medical center to study the association between inpatient pharmacologic prophylaxis
21
and the occurrence of postoperative VTE and bleeding complications. It was hypothesized that
22
chemoprophylaxis would lead to a reduced incidence of VTE without significantly increasing the
23
incidence of bleeding complications requiring reoperation.
24 5
Page 5 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2 3
Materials and Methods
4
A retrospective cohort study was conducted using NSQIP years 2005-2014. Patients undergoing
5
elective spine surgery were identified based on the following procedures and Current Procedural
6
Terminology (CPT) codes: anterior cervical discectomy and fusion (22551, 22554, 63075),
7
anterior cervical corpectomy (63081, 63085), cervical disc arthroplasty (22856, 22857), cervical
8
laminectomy (63015, 63045, 63265), cervical laminotomy (63020, 63040), posterior cervical
9
fusion (22595, 22600), posterior thoracic fusion (22610), thoracic laminectomy (63046, 63266),
NSQIP Cohort
10
thoracic corpectomy (63085, 63086), anterior lumbar fusion (22558), anterior lumbar
11
corpectomy (63087, 63088), lumbar laminectomy (63047, 63005, 63012, 63267), lumbar
12
laminotomy (63030, 63042), and posterior lumbar fusion (22612, 22630, 22633). Patients were
13
excluded for cases marked emergency or non-elective, fusion of seven or more levels (to exclude
14
deformity cases with increased risks[17,18]), missing data, and primary International
15
Classification of Diseases (ICD) diagnosis codes indicating trauma, tumor, or infection.
16 17
Demographic, comorbidity, intraoperative, and postoperative factors were characterized.
18
Demographic factors included age, sex, and body mass index. Comorbidity factors included
19
smoking status, diabetes, hypertension, chronic obstructive pulmonary disease, congestive heart
20
failure, preoperative renal failure, disseminated cancer, coagulopathy, preoperative anemia,
21
dependent functional status, and American Society of Anesthesiologists’ classification of III or
22
higher.
23
6
Page 6 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Intraoperative factors included operative time, procedure type (characterized as anterior cervical,
2
posterior cervical, thoracic, anterior lumbar, posterior lumbar, or combined anterior/posterior
3
lumbar), undergoing a multi-level procedure, and perioperative blood transfusion. Postoperative
4
factors included the presence of several postoperative complications (venous thromboembolism,
5
superficial surgical site infection, deep surgical site infection, pneumonia, mechanical ventilation
6
> 48 hours, unplanned reintubation, urinary tract infection, postoperative renal insufficiency or
7
failure, sepsis/septic shock, stroke, cardiac arrest, and myocardial infarction) and length of stay.
8
Of note, the database specifies that all thromboembolic events were confirmed with imaging and
9
required treatment with anticoagulation and/or inferior vena cava filter; however, because of the
10
varying practices at the many participating centers, it cannot be definitively known whether all
11
VTEs presented symptomatically or whether some were detected through routine imaging
12
screenings of all patients at some centers.[19]
13 14
Institutional Cohort
15
The elective spine surgery patients recorded for NSQIP at a single large academic medical
16
center, January 2013-March 2016, were identified. A retrospective cohort from this population
17
was identified utilizing the same inclusion and exclusion criteria as noted above for the NSQIP
18
cohort.
19 20
For each of these patients, NSQIP-recorded thirty postoperative day follow-up data was
21
available. In addition to the NSQIP-recorded variables, a retrospective chart review was
22
completed to examine for use of pharmacologic and mechanical VTE prophylaxis, history of
23
prior VTE, and incidence of hematoma requiring reoperation. The institutional data included
7
Page 7 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
only symptomatic VTEs (i.e. presenting with lower extremity pain or swelling for deep vein
2
thrombosis, dyspnea or chest pain for pulmonary embolism), consistent with NSQIP criteria
3
stated above,[19] and confirmed with venous duplex ultrasonography (in the case of DVT) or
4
computed tomography angiogram or ventilation/perfusion scan (in the case of PE).
5 6
Statistical Analysis
7
For the NSQIP cohort, multivariate Poisson regression with robust error variance was used to
8
test the association of the noted demographic, comorbidity, intraoperative, and postoperative
9
factors with development of venous thromboembolism within postoperative day 30.
10 11
The national and institutional samples were compared on the basis of age, sex, BMI, ASA class
12
≥ III, procedure type, length of stay, and incidence of VTE. Student’s t-test was used to compare
13
continuous variables, and chi-squared test was used to compare categorical variables.
14 15
To assess for variables that may influence prescription of pharmacologic prophylaxis in the
16
institutional cohort, a multivariate Poisson regression with robust error variance was used to test
17
the association of demographic, comorbidity (including history of VTE), and intraoperative
18
factors with whether or not a patient received prophylaxis.
19 20
Multivariate Poisson regression with robust error variance was used to test the association of
21
NSQIP variables, as well as history of prior VTE and pharmacologic prophylaxis status, with
22
incidence of venous thromboembolism within postoperative day 30, controlling for demographic,
23
comorbidity, intraoperative, and postoperative factors. The same multivariate regression analysis
8
Page 8 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
was repeated within certain subpopulations deemed to be at high risk for VTE based upon the
2
analysis.
3 4
Another multivariate Poisson regression with robust error variance was performed to test
5
association of demographic, comorbidity, and intraoperative variables, as well as pharmacologic
6
prophylaxis, with incidence of hematoma requiring reoperation.
7 8
The level of significance for all tests was set at P<0.05. Statistical tests were performed using
9
Stata® version 13.1 (StataCorp, LP, College Station, Texas, USA).
10 11
The current study was funded by the James G. Hirsch, MD, Endowed Medical Student Research
12
Fellowship ($4,000), provided by the Office of Student Research at the investigators’ institution.
13
There were no study-related conflicts of interest for any authors.
14
9
Page 9 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
Results NSQIP Cohort
3
In the national study, 109,609 patients were included. Initially, 123,320 patients were identified
4
based upon included CPT codes, of which 13,711 were excluded because of non-elective case,
5
fusion of 7+ levels, or diagnosis of tumor, trauma, or infection, or missing data. Demographics
6
of this population are presented in Table 1. The incidence of venous thromboembolism was
7
0.61% (672/109,609 patients).
8 9
The results of the multivariate analysis for independent risk factors for VTE are presented in
10
Table 2. Significant factors with the greatest relative risk were: stroke (relative risk (RR)=4.27,
11
P<0.001), pneumonia (RR=2.92, P<0.001), cardiac arrest (RR=2.62, P=0.015), anterior lumbar
12
procedure (RR=2.59, P<0.001), and superficial surgical site infection (RR=2.28, P=0.001).
13 14
Institutional Cohort
15
There were 2,855 patients included in the institutional study. Demographics of this population
16
are presented in Table 1. Initially, 3,254 patients were identified based upon included CPT codes,
17
of which 399 were excluded because of non-elective case, fusion of 7+ levels, or diagnosis of
18
tumor, trauma, or infection, or missing data. The incidence of venous thromboembolism was
19
1.23% (35/2,855 patients), which was significantly greater than the national cohort (0.61%,
20
P<0.001). There were also differences in BMI (national: 30.2±16.0 kg/m2, institutional 29.5±5.7;
21
P=0.027) and ASA class ≥ III (national: 39.9%, institutional: 34.4%; P<0.001). There were no
22
differences in age, sex, or length of stay.
23
10
Page 10 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
In the institutional study, all patients received mechanical prophylaxis with sequential
2
compression devices and early ambulation. Pharmacologic prophylaxis was received by 56.1%
3
of patients. Among those patients, the three most utilized medications were unfractionated
4
heparin (97.1%), low molecular weight heparin (2.5%), and warfarin (0.3%).
5 6
Variables significantly associated with receiving pharmacologic prophylaxis are delineated in
7
Table 3. The factors most significantly associated with receiving pharmacologic VTE
8
prophylaxis were: thoracic surgery (RR=1.62, P<0.001), anterior lumbar surgery (RR=1.32,
9
P<0.001), history of prior VTE (RR=1.21, P=0.006), ASA class ≥ III (RR=1.14, P<0.001), and
10
multi-level procedure (RR=1.10, P=0.007). Perioperative blood transfusion was associated with
11
decreased frequency of prophylaxis (RR=0.62, P<0.001), but cause and effect of this relationship
12
could not be determined.
13 14
In a univariate comparison, there was no significant association between pharmacologic
15
prophylaxis and incidence of VTE (RR=1.32, P=0.421; Figure 1).
16 17
In the multivariate analysis for associations with venous thromboembolism in the institutional
18
cohort, independent risk factors for VTE are delineated in Table 4. The factors most
19
significantly associated with VTE in order of decreasing relative risk were: urinary tract
20
infection (RR=16.05, P<0.001), history of prior venous thromboembolism (RR=9.48, P<0.001),
21
perioperative blood transfusion (RR=3.56, P=0.003), male sex (RR=2.64, P=0.018), and
22
increasing BMI (RR=1.08 per unit, P=0.030). Even with multivariate analysis, pharmacologic
23
prophylaxis was not found to significantly influence the rate of VTE (RR=0.68, P=0.424).
11
Page 11 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
The following subpopulations were analyzed based upon increased risk for VTE: age ≥ 65 years
3
(N=863), male sex (N=1,491), BMI ≥ 30 (N=1,218), lumbar procedure type (N=2,124), and
4
history of prior VTE (N=84). Following multivariate analysis, pharmacologic prophylaxis was
5
not associated with significant decrease in VTE incidence in any of the subgroups.
6 7
The incidence of postoperative bleeding or hematoma requiring return to the operating room was
8
0.4% (11/2,855). Ten of these were surgical site hematomas, while one patient experienced
9
bilateral cerebellar hematomas. Seven of the hematomas presented with significant neurological
10
deficits (63.6%), three presented with pain or wound drainage (27.3%), and one presented with
11
difficulty breathing (9.1%). Among 10 patients experiencing hematoma who received
12
pharmacologic prophylaxis, nine received unfractionated heparin and one received low
13
molecular weight heparin.
14 15
Incidence of postoperative hematoma requiring reoperation was significantly greater among
16
patients receiving prophylaxis (0.62% receiving prophylaxis versus 0.08% not receiving
17
prophylaxis; RR=7.80, P=0.020; Figure 1). In the multivariate analysis for associations with
18
postoperative hematoma (Table 5), the association with pharmacologic prophylaxis (RR=7.37,
19
P=0.048) remained significant. There was also a significant association with smoking (RR=3.16,
20
P=0.048).
21 22
12
Page 12 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Discussion
2
The incidence of VTE in the NSQIP population undergoing elective spine surgery was found to
3
be 0.61%. This is on the low end of incidences typically reported in the spine surgery literature,
4
including prior studies using NSQIP.[15,16] This relatively low incidence may have been found
5
because of the exclusion of high-risk diagnoses (e.g. trauma, neoplasm) that past studies have
6
shown to be associated with increased risk of VTE.[12,16,17] In contrast, the incidence of VTE
7
in the institutional cohort was significantly greater at 1.23%. It is unclear whether this variation
8
is due to patient factors, differences in management, or vigilance of screening for VTE.
9 10
Independent risk factors for VTE identified in both national and institutional analyses included
11
greater age, male sex, increasing BMI, perioperative blood transfusion, longer length of stay, and
12
urinary tract infection. Several variables—dependent functional status, lumbar surgery, greater
13
operative time, and various postoperative complications—identified as risk factors in the national
14
sample were not identified in the institutional cohort. It is not clear whether these differences are
15
due to actual differences between the populations versus effects that could not be identified due
16
to lesser sample size in the institutional sample. Of note, several postoperative complications
17
(e.g. stroke, cardiac arrest, surgical site infections) could not be assessed at the institutional level
18
because there were no incidences of VTE among patients also experiencing the complication.
19 20
Notably, no medical comorbidity was associated with increased risk of VTE in either the
21
national or institutional population. This contrasts with several past studies that have identified
22
conditions including cancer, hypertension, anemia, and high ASA class as risk factors.[13,15,16]
23
Other notable variables not associated with increased incidence of VTE were multilevel fusions
13
Page 13 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
and circumferential fusions (compared to posterior fusion), whereas some prior studies have
2
identified these as risk factors.[17,20]
3 4
Contrary to our hypothesis, pharmacologic prophylaxis did not significantly decrease the
5
incidence of venous thromboembolism, even after controlling for differences in
6
patient/procedural factors and patterns of prophylaxis usage. Further, in subgroup analysis
7
among populations with increased VTE risk (e.g. patients with prior VTE history), there was no
8
subgroup where prophylaxis was associated with a significant change in the incidence of VTE.
9
This is not out of the spectrum of findings in the literature where widely varying findings have
10
been reported concerning the efficacy of pharmacologic prophylaxis in preventing VTE in the
11
spine population.[7-11]
12 13
In a meta-analysis of VTE following elective surgery, Sansone and colleagues found a
14
significant reduction in deep vein thrombosis with pharmacologic prophylaxis compared to only
15
mechanical or no prophylaxis, though there were not enough pulmonary embolism cases to
16
evaluate.[12] Notably, they could not perform multivariate analysis to assess for potential
17
confounding variables because such data was not available in the original studies. Additionally,
18
the studies included in the meta-analysis utilized active screening for VTE, so many
19
asymptomatic cases were included. It is possible that there are different effects when comparing
20
clinically evident venous thromboembolic disease versus asymptomatic cases detected on
21
screening. In contrast, Schuster and colleagues conducted a systematic review of elective
22
thoracolumbar surgery that found no change in VTE risk with pharmacologic prophylaxis,
23
though the study was limited by small sample size.[9]
14
Page 14 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1 2
Although one could question if greater power might have lead to identifying a significant
3
reduction of VTE in the current study, it is important to note that pharmacologic VTE
4
prophylaxis was associated with significant increase in incidence of bleeding complications
5
requiring return to the operating room, even after controlling for patient factors that may
6
influence bleeding risk or prescribing patterns.
7 8
While epidural hematoma and other bleeding complications are a commonly cited concern with
9
prophylactic anticoagulation after spine surgery, there is a lack of evidence to decisively
10
characterize the level of risk that prophylaxis poses and to identify patients where benefits
11
outweigh risks.[10] The current study suggests that pharmacologic prophylaxis with
12
unfractionated heparin (97.1% of prophylaxis used in the institutional cohort) was significantly
13
associated with increased risk for bleeding complications that require reoperation in elective
14
spine surgery. This supports the frequent assertion that these adverse events are an important
15
consideration before initiating prophylaxis after elective spine surgery.
16 17
For patients who experienced a bleeding complication, the use of chemoprophylaxis, timing of
18
first dose, and timing that the hematoma was recognized were assessed (Appendix 1). Of note,
19
three of the ten patients who received pharmacologic prophylaxis and subsequently sustained a
20
hematoma received their first dose in the evening of postoperative day zero. It is not clear
21
whether this represents a significant difference from patients who did not experience a
22
hematoma. However, there is some evidence that the risk of bleeding complications is increased
15
Page 15 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
with earlier initiation of chemoprophylaxis, so this finding may be consistent with existing
2
literature.[21]
3 4
While prevention of venous thromboembolism after spine surgery has been previously studied in
5
the literature, the current investigation makes several notable contributions. While prior studies
6
of chemoprophylaxis after elective spine surgery have generally featured sample sizes ranging
7
from 40-400,[7,21-23] the current study’s large sample size of nearly 3,000 enables more
8
accurate assessment of relatively rare events, VTE and postoperative hematoma. This may be
9
particularly relevant concerning hematomas, as the rate of postoperative bleeding complications
10
requiring reoperation was less than 1 per 250 patients, meaning that prior studies likely could not
11
study these events with adequate power. Additionally, the current study featured an internal
12
control group of patients who did not receive pharmacologic prophylaxis, meaning that
13
differences between the control and treatment groups could be identified and controlled for and
14
that both groups featured the same inclusion and exclusion criteria. In contrast, many prior
15
investigations have been case series where all patients received the same treatment and any
16
comparisons were to historical rates in other manuscripts, meaning population differences could
17
not be characterized.[7,21-23] These represent meaningful strengths over prior studies evaluating
18
the risks and benefits of pharmacologic VTE prophylaxis after spine surgery.
19 20
Compared to prior studies relying on the NSQIP database, the current study is novel in
21
combining national data and single-institution data that was supplemented with a targeted chart
22
review to investigate VTE after elective spine surgery. This approach combines the large sample
23
sizes, 30-day follow-up, and high-quality data abstraction by trained NSQIP reviewers with key
16
Page 16 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
variables for studying VTE that are not available in the database. The connection with NSQIP
2
data also provided a large number of other variables in order to control for potential associations
3
with which patients received prophylaxis and which developed a VTE or bleeding
4
complication.[19] To our knowledge, this is also the first study of VTE incidence using NSQIP
5
to focus solely on elective spine surgery, whereas prior studies have included all spine cases,
6
which could affect study results.[15,16] Additionally, compared to other database studies,
7
NSQIP provides specific variables to confirm whether cases are elective, meaning that the
8
current study offers a more accurate sample of elective cases compared to VTE studies that have
9
utilized other databases.
10 11
The current study is limited by several factors, most notably its retrospective, observational
12
nature. Randomized controlled trials are needed to better evaluate the role of prophylactic
13
anticoagulation in venous thromboembolism and its risk for bleeding complications. Another
14
limitation is that the overwhelming majority of patients on prophylaxis received unfractionated
15
heparin. Due to the retrospective design of the study, choice of prophylactic agent was entirely
16
due to individual surgeons’ practices. There is some evidence supporting low-molecular-weight
17
heparin as a safe option for anticoagulation after elective spine surgery, though this study utilized
18
a relatively small sample size and may have been underpowered, [21] and newer studies are
19
beginning to assess novel oral anticoagulants as well.[24] It was not possible to compare the
20
efficacy of various anticoagulants in the current study, and other anticoagulants may potentially
21
demonstrate superior results compared to unfractionated heparin; however they may also be
22
associated with bleeding risks/concerns, necessitating further study of these other agents.
23
17
Page 17 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
While the overall study benefited from a large sample size, some of the high-risk subgroups had
2
small sample sizes that may have limited the power to detect an effect of prophylaxis on VTE
3
incidence. Dedicated investigation of these subpopulations may potentially demonstrate a greater
4
benefit of anticoagulation compared to the entire population. Another limitation was the ability
5
to monitor for VTEs only to postoperative day 30. While the timing of VTE after elective spine
6
surgery has not been conclusively established, it is possible that some cases occurred beyond the
7
monitoring period, which could potentially influence results.
8 9
In conclusion, venous thromboembolism is a well-recognized and potentially dangerous
10
complication of elective spine surgery, creating a need for safe and effective prophylactic
11
measures. The current study found that pharmacologic prophylaxis predominantly with
12
unfractionated heparin did not significantly decrease the risk of VTE but significantly increased
13
the risk of bleeding events requiring reoperation. Current guidelines state that there is generally
14
insufficient evidence to support routine use of chemoprophylaxis in low-risk patients, and the
15
findings of the current study support the need for further risk/benefit considerations.
16
18
Page 18 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
References
2
[1]
Fitzgerald RH, Spiro TE, Trowbridge AA, Gardiner GA, Whitsett TL, O’Connell MB,
3
Ohar JA, Young TR. Prevention of venous thromboembolic disease following primary
4
total knee arthroplasty: A randomized, multicenter, open-label, parallel-group
5
comparison of enoxaparin and warfarin. J Bone Joint Surg Am 2001;83:900-906.
6
[2]
7 8
Geerts WH, Code KI, Jay RM, Chen E, Szalai JP. A prospective study of venous thromboembolism after major trauma. N Engl J Med 1994;331(24):1601-1606.
[3]
9
Sobieraj DM, Lee S, Coleman CI, Tongbram V, Chen W, Colby J, Kluger J, Makanji S, Ashaye AO, White CM. Prolonged versus standard-duration venous thromboprophylaxis
10
in major orthopedic surgery: A systematic review. Ann Intern Med 2012;156(10):720-
11
727.
12
[4]
Kester BS, Merkow RP, Ju MH, Peabody TD, Bentrem DJ, Ko CY, Bilimoria KY. 2014.
13
Effects of post-discharge venous thromboembolism on hospital quality comparisons
14
following hip and knee arthroplasty. J Bone Joint Surg Am 2014;96(17):1676-1684.
15
[5]
Falck-Ytter Y, Francis CW, Johanson NA, Curley C, Dahl OE, Schulman S, Ortel TL,
16
Pauker SG, Colwell CW. Prevention of VTE in orthopedic surgery patients:
17
Antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest
18
Physicians evidence-based clinical practice guidelines. Chest 2012;141(2_suppl):e278S-
19
e325S.
20
[6]
Jacobs JJ, Mont MA, Bozic KJ, Della Valle CJ, Goodman SB, Lewis CG, Yates AJ,
21
Boggio LN. Preventing venous thromboembolic disease in patients undergoing elective
22
hip and knee arthroplasty: Evidence-based guideline and evidence report. Rosemont, IL:
23
American Academy of Orthopaedic Surgeons; 2011.
19
Page 19 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[7]
2 3
Glotzbecker MP, Bono CM, Wood KB, Harris MB. Thromboembolic disease in spinal surgery: A systematic review. Spine (Phila Pa 1976) 2009;34(3):291-303.
[8]
Cox JB, Weaver KJ, Neal DW, Jacob RP, Hoh DJ. Decreased incidence of venous
4
thromboembolism after spine surgery with early multimodal prophylaxis: Clinical article.
5
J Neurosurg Spine 2014;21(4):677-684.
6
[9]
7 8
Schuster JM, Fischer D, Dettori JR. Is chemical antithrombotic prophylaxis effective in effective thoracolumbar spine surgery? Evid Based Spine Care J 2010;1(2):40-45.
[10]
9
Bono CM, Watters WC 3rd, Heggeness MH, Resnick DK, Shaffer WO, Baisden J, BenGalim P, Easa JE, Fernand R, Lamer T, Matz PG, Mendel RC, Patel RK, Reitman CA,
10
Toton JF. An evidence-based clinical guideline for the use of antithrombotic therapies in
11
spine surgery. Spine J 2009;9(12):1046-1051.
12
[11]
Gould MK, Garcia DA, Wren SM, Karanicolas PJ, Arcelus JI, Heit JA, Samama CM.
13
Prevention of VTE in nonorthopedic surgical patients: Antithrombotic therapy and
14
prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based
15
clinical practice guidelines. Chest 2012;141(2):e227S-e277S.
16
[12]
Sansone JM, del Rio AM, Anderson PA. The prevalence of and specific risk factors for
17
venous thromboembolic disease following elective spine surgery. J Bone Joint Surg Am
18
2010;92(2):304-313.
19
[13]
Fineberg SJ, Oglesby M, Patel AA, Pelton MA, Singh K. The incidence and mortality of
20
thromboembolic events in lumbar spine surgery. Spine (Phila Pa 1976)
21
2013;38(13):1154-1159.
20
Page 20 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[14]
Oglesby M, Fineberg SJ, Patel AA, Pelton MA, Singh K. The incidence and mortality of
2
thromboembolic events in cervical spine surgery. Spine (Phila Pa 1976)
3
2013;38(9):E521-527.
4
[15]
Sebastian AS, Currier BL, Kakar S, Nguyen EC, Wagie AE, Habermann ES, Nassr A.
5
Risk factors for venous thromboembolism following thoracolumbar surgery: Analysis of
6
43,777 patients from the American College of Surgeons National Surgical Quality
7
Improvement Program 2005 to 2012. Global Spine J 2016;6(8):738-743.
8
[16]
9
HJ, Vates GE. Risk factors associated with venous thromboembolism in patients
10 11
undergoing spine surgery. J Neurosurg Spine 2017;26(1):90-6. [17]
12 13
[18]
[19]
User Guide for the 2015 ACS NSQIP Participant Use Data File. Chicago: American College of Surgeons; 2016
[20]
18 19
Cheng JS, Arnold PM, Anderson PA, Fischer D, Dettori JR. Anticoagulation risk in spine surgery. Spine (Phila Pa 1976) 2010;35(9 Suppl):S117-24.
16 17
Schairer WW, Pedtke AC, Hu SS. Venous thromboembolism after spine surgery. Spine (Phila Pa 1976) 2014 [Epub ahead of print].
14 15
Piper K, Algattas H, DeAndrea-Lazarus IA, Kimmell KT, Li YM, Walter KA, Silberstein
Brambilla S, Ruosi C, La Maida GA, Caserta S. Prevention of venous thromboembolism in spinal surgery. Eur Spine J 2004;13(1):1-8.
[21]
Strom RG, Frempong-Boadu A. Low-molecular-weight heparin prophylaxis 24 to 36
20
hours after degenerative spine surgery: risk of hemorrhage and venous
21
thromboembolism. Spine (Phila Pa 1976) 2013;38(23):E1498-1502.
22 23
[22]
Al-Dujaili TM, Majer CN, Madhoun TE, Kassis SZ, Saleh AA. Deep venous thrombosis in spine surgery patients: incidence and hematoma formation. Int Surg 2012;97(2):150-4.
21
Page 21 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
[23]
Spinal Cord Injury Thromboprophylaxis Investigators. Prevention of venous
2
thromboembolism in the rehabilitation phase after spinal cord injury: prophylaxis with
3
low-dose heparin or enoxaparin. J Trauma 2003;54(6):1111-5.
4
[24]
Du W, Zhao C, Wang J, Liu J, Shen B, Zheng Y. Comparison of rivaroxaban and
5
parnaparin for preventing venous thromboembolism after lumbar spine surgery. J Orthop
6
Surg Res 2015;10:78.
7 8 9 10 11
Figure 1. Relative risk of experiencing venous thromboembolism or hematoma based on
12
prophylaxis status, with or without controlling for other variables. Diamond indicates relative
13
risk; bar indicates 95% confidence interval. Black shading indicates that prophylaxis is
14
significant at P < 0.05; gray shading indicates that variable is not significant. VTE=Venous
15
thromboembolism; OR=operating room.
16 17
22
Page 22 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 1. Demographics
Number of patients
Age, years
National
Institutional
109,609
2,855
56.4 ± 14.4
56.0 ± 14.7
Gender
0.125
0.815
Female
52,628 (48.0%)
1,364 (47.8%)
Male
56,931 (52.0%)
1,491 (52.2%)
30.2 ± 16.0
29.5 ± 5.7
0.027
39.9%
34.4%
<0.001
(43,758/109,609)
(983/2,855)
Length of stay, days
2.4 ± 5.3
2.4 ± 2.9
0.633
Incidence of venous
0.61%
1.23%
<0.001
(672/109,609)
(35/2,855)
Body Mass Index, kg/m2
ASA Class ≥ III
thromboembolism
2 3
P-value
ASA = American Society of Anesthesiologists
4 5
23
Page 23 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 2. Multivariate relative risk for venous thromboembolism in national cohort Risk Factor
N (VTE)
N (risk
Multivariate Relative Risk#
P-value
factor) Age, years
‡
‡
1.02 (1.01-1.03)
<0.001
Male sex
‡
‡
1.28 (1.09-1.49)
0.002
BMI, per unit
‡
‡
1.001 (1.000-1.001)
0.005
Smoking
83
26,316
0.56 (0.45-0.71)
<0.001
Diabetes mellitus
118
16,767
0.81 (0.63-1.03)
0.089
Insulin-dependent diabetes
40
5,371
0.94 (0.64-1.42)
0.776
Hypertension
423
53,295
1.06 (0.90-1.28)
0.515
COPD
39
4,193
1.11 (0.80-1.57)
0.549
Congestive heart failure
2
220
0.51 (0.14-1.85)
0.320
Preop renal disease
1
243
0.16 (0.02-1.65)
0.124
Disseminated cancer
2
130
0.97 (0.24-4.61)
0.973
Coagulopathy
18
1,431
1.33 (0.83-2.10)
0.227
Preoperative anemia
126
19,676
0.87 (0.71-1.08)
0.200
Dependent functional status
39
2,343
1.71 (1.24-2.43)
<0.001
ASA class ≥ III
364
43,758
1.12 (0.95-1.34)
0.209
Medical Comorbidities
Procedure type
<0.001
Anterior cervical
77
24,499
Reference
Posterior cervical
28
4,672
1.32 (0.88-2.09)
0.207
Thoracic
9
644
0.99 (0.40-2.54)
0.981
Anterior lumbar
38
3,133
2.59 (1.75-3.87)
<0.001
Posterior lumbar
479
73,798
1.46 (1.14-1.89)
0.003
Anterior/posterior lumbar
41
2,863
2.18 (1.47-3.26)
<0.001
24
Page 24 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
Intraoperative Factors ‡
‡
1.003 (1.002-1.003)
<0.001
Multi-level procedure
307
35,670
1.11 (0.95-1.32)
0.228
Periop blood transfusion
124
5,299
1.78 (1.38-2.24)
<0.001
Superficial SSI
19
901
2.28 (1.39-3.63)
0.001
Deep/organ space SSI
25
706
2.19 (1.30-3.65)
0.003
Pneumonia
41
521
2.92 (1.66-5.05)
<0.001
Failure to wean from
30
238
1.41 (0.67-3.23)
0.381
Unplanned reintubation
30
348
1.32 (0.63-2.78)
0.459
Urinary tract infection
44
1,160
2.07 (1.40-3.07)
<0.001
Renal insufficiency/failure
12
155
1.99 (0.94-4.01)
0.063
Sepsis/septic shock
48
612
1.95 (1.17-3.20)
0.009
Stroke
11
98
4.27 (2.00-9.05)
<0.001
Cardiac arrest
11
115
2.62 (1.20-5.60)
0.015
Myocardial infarction
11
193
1.55 (0.70-3.32)
0.267
‡
‡
1.01 (1.01-1.01)
<0.001
Operative time, minutes
Postoperative Complications
ventilator
Length of stay, days
1
# Relative risk (95% confidence interval)
2
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; SSI = Surgical site infection; ASA =
3
American Society of Anesthesiologists. Shading indicates statistical significance at P < 0.05.
4
N(VTE) indicates number of patients with each risk factor who experienced a VTE.
5
N(risk factor) indicates total number of patients with given risk factor, out of total N = 109,609.
6
‡ indicates a continuous variable, so all patients are included.
25
Page 25 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 3. Multivariate regression for significant associations with receiving pharmacologic prophylaxis Multivariate Relative Risk#
P-value
Age, years
1.01 (1.006-1.012)
<0.001
Female sex
1.09 (1.03-1.17)
0.004
0.99 (0.988-0.999)
0.044
Smoking
1.09 (1.001-1.17)
0.048
Diabetes mellitus
0.93 (0.84-1.03)
0.172
Insulin-dependent diabetes
1.04 (0.88-1.22)
0.680
Hypertension
0.97 (0.90-1.04)
0.347
COPD
0.95 (0.80-1.12)
0.536
Congestive heart failure
1.08 (0.46-2.51)
0.861
Preop renal disease
1.10 (0.71-1.69)
0.669
Disseminated cancer
0.90 (0.49-1.66)
0.732
Coagulopathy
0.87 (0.66-1.14)
0.304
Preoperative anemia
1.00 (0.89-1.13)
0.966
status
1.14 (0.88-1.48)
0.332
ASA class ≥ III
1.14 (1.06-1.22)
<0.001
History of prior VTE&
1.21 (1.06-1.39)
0.006
Risk Factor
BMI, per unit
Medical Comorbidities
Dependent functional
Procedure type Anterior cervical
Reference
Posterior cervical
1.19 (0.99-1.43)
0.063
Thoracic
1.62 (1.26-2.09)
<0.001
Anterior lumbar
1.32 (1.16-1.50)
<0.001
Posterior lumbar
1.08 (0.99-1.18)
0.094
26
Page 26 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism A/P lumbar
0.97 (0.82-1.14)
0.689
1.003 (1.002-1.003)
<0.001
Multi-level procedure
1.10 (1.02-1.17)
0.007
Periop blood transfusion
0.62 (0.54-0.72)
<0.001
Intraoperative Factors Operative time, minutes
1
#
2
&
3
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; ASA = American Society of
4
Anesthesiologists.
5
Shading indicates statistical significance at P < 0.05.
Relative risk (95% confidence interval) Chart review variable
6 7 8
27
Page 27 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 4. Multivariate relative risk for venous thromboembolism in institutional cohort, including chart review N (VTE)
N (risk factor)
Multivariate Relative Risk#
P-value
Age, years
‡
‡
1.07 (1.03-1.11)
<0.001
Male sex
‡
‡
2.64 (1.18-5.90)
0.018
BMI, per unit
‡
‡
1.08 (1.01-1.15)
0.030
Smoking
5
540
1.12 (0.45-3.02)
0.813
Diabetes mellitus
4
403
0.41 (0.12-1.23)
0.178
Insulin-dependent diabetes
2
127
1.62 (0.33-11.08)
0.604
Hypertension
20
1,238
0.45 (0.21-0.94)
0.037
COPD
2
79
1.68 (0.50-5.30)
0.369
Congestive heart failure
0
3
N/A
Preop renal disease
0
8
N/A
Disseminated cancer
0
5
N/A
Coagulopathy
2
36
0.88 (0.27-3.18)
0.852
Preoperative anemia
3
182
0.80 (0.27-2.30)
0.690
Dependent functional status
0
35
N/A
ASA class ≥ III
15
983
0.74 (0.29-1.76)
Risk Factor
Medical Comorbidities
Procedure type
0.483
0.242
Anterior cervical
3
632
Reference
Posterior cervical
1
85
1.65 (0.15-18.43)
Thoracic
0
14
N/A
Anterior lumbar
5
158
3.28 (0.68-15.74)
0.138
Posterior lumbar
22
1,858
1.21 (0.37-3.93)
0.757
A/P lumbar
4
108
3.04 (0.70-13.16)
0.138
0.684
28
Page 28 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism Intraoperative Factors Operative time, minutes
‡
‡
1.004 (0.999-1.008)
0.059
Multi-level procedure
10
798
0.98 (0.43-2.42)
0.963
Periop blood transfusion
10
137
3.56 (1.74-7.94)
0.003
Superficial SSI
0
27
N/A
Deep/organ space SSI
0
16
N/A
Pneumonia
3
26
3.29 (0.87-15.59)
0.162
Failure to wean from
2
5 N/A
N/A
Postoperative Complications
ventilator Unplanned reintubation
1
6
5.04 (0.32-137.15)
0.353
Urinary tract infection
3
28
16.05 (3.48-49.74)
<0.001
Renal insufficiency/failure
1
5
7.27 (0.86-51.35)
0.070
Sepsis/septic shock
1
16
7.16 (0.16-56.78)
0.178
Stroke
0
3
N/A
Cardiac arrest
0
5
N/A
Myocardial infarction
0
5
N/A
Length of stay, days
‡
‡
1.07 (1.04-1.09)
<0.001
History of prior VTE&
9
84
9.48 (3.07-28.23)
<0.001
Pharmacologic prophylaxis&
22
1,603
0.68 (0.25-1.71)
0.424
1
N(VTE) indicates number of patients with each risk factor who experienced a VTE. N(risk factor) indicates total
2
number of patients with given risk factor, out of total N = 2,855. # indicates relative risk (95% confidence interval).
3
‡ indicates continuous variable, so all patients are included. &Chart review variables. N/A indicates variable could
4
not be analyzed because N(VTE) = 0 or due to covariance. BMI = Body mass index; COPD = Chronic obstructive
5
pulmonary disease; SSI = Surgical site infection; ASA = American Society of Anesthesiologists.
6
29
Page 29 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
Table 5. Multivariate relative risk for postoperative hematoma in institutional cohort Multivariate Relative Risk#
P value
Age
1.02 (0.99-1.05)
0.188
Male sex
2.89 (0.71-11.85)
0.140
BMI
0.93 (0.84-1.03)
0.154
N/A
N/A
Smoking
3.16 (1.01-9.88)
0.048
Diabetes mellitus
0.93 (0.17-5.18)
0.937
Hypertension
0.54 (0.18-1.65)
0.282
N/A
N/A
Coagulopathy
5.69 (0.50-64.39)
0.160
Preoperative anemia
0.87 (0.12-6.48)
0.894
ASA class ≥ III
3.05 (0.96-9.71)
0.059
Operative time
1.00 (0.99-1.01)
0.301
Multi-level procedure
2.58 (0.71-9.42)
0.151
Perioperative blood transfusion
2.53 (0.36-17.96)
0.353
Pharmacologic prophyaxis&
7.37 (1.01-53.46)
0.048
Medical Comorbidities Dependent functional status
COPD
Intraoperative Factors
2
#
3
&
4
BMI = Body mass index; COPD = Chronic obstructive pulmonary disease; ASA = American Society of
5
Anesthesiologists.
6
Shading indicates statistical significance at P < 0.05.
Relative risk (95% confidence interval) Chart review variable
7
30
Page 30 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism 1
31
Page 31 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2 3
Affirmation of Authorship - VTE Spine JPEG.jpg
32
Page 32 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2 3
VTE Prophylaxis Spine - Figure 1 jpeg.jpg
33
Page 33 of 34
Pharmacologic Prophylaxis for Venous Thromboembolism
1 2
VTE Prophylaxis Spine - Flow Diagram Figure.jpg
34
Page 34 of 34