Use and outcomes of venous thromboembolism prophylaxis after spinal fusion surgery


Margaret C. Fang, University of California, San Francisco, 505 Parnassus Ave., Box 0131, San Francisco, CA 94143, USA.
Tel.: +1 415 502 7100; fax: +1 415 514 2094.


Summary. Background: The number of spinal fusion operations in the USA is rapidly rising, but little is known about optimal venous thromboembolism prophylaxis after spinal surgery. Objectives: To examine the use of and outcomes associated with venous thromboembolism prophylaxis after spinal fusion surgery in a cohort of 244 US hospitals. Patients/Methods: We identified all patients with a principal procedure code for spinal fusion surgery in hospitals participating in the Premier Perspective database from 2003 to 2005, and searched for receipt of pharmacologic prophylaxis (subcutaneous unfractionated heparin, low molecular weight heparin, or fondaparinux) and/or mechanical prophylaxis (compression devices and elastic stockings) within the first 7 days after surgery. We also searched for discharge diagnosis codes for venous thromboembolism and postoperative hemorrhage during the index hospitalization and within 30 days after surgery. Results: Among 80 183 spinal fusions performed during the time period, cervical fusions were the most common (49.0%), followed by lumbar fusions (47.8%). Thromboembolism prophylaxis was administered to 60.6% of patients within the first week postoperatively, with the most frequent form being mechanical prophylaxis alone (47.6%). Of the 244 hospitals, 26.2% provided prophylaxis to ≥ 90% of their patients undergoing spinal fusion; however, 33.2% provided prophylaxis to fewer than 50% of their patients. The rate of diagnosed venous thromboembolism within 30 days after surgery was 0.45%, and the rate of postoperative hemorrhage was 1.1%. Conclusions: Substantial variation exists in the use of thromboembolism prophylaxis after spinal fusion surgery in the USA. Nevertheless, overall rates of diagnosed thromboembolism after spinal fusion appear to be low.


The number of spinal fusions performed in the USA has rapidly increased, approaching the rate of joint arthroplasties [1–3]. Although pharmacologic prophylaxis for venous thromboembolism (VTE) is strongly recommended after hip and knee arthroplasty and is supported by randomized clinical trial data [4], the optimal mode of prophylaxis for patients undergoing spinal fusion surgery is unknown. Therefore, guidelines have been vague regarding the optimal choice of antithrombotic therapy after spinal fusion. Although a number of risk factors appear to increase the risk of thromboembolism after neurosurgical procedures, including the type of surgery, older age, and underlying malignancy, it is not clear how individual risk factors should be incorporated into decision-making regarding VTE prophylaxis [5].

Various guidelines have suggested subcutaneous unfractionated heparin, low molecular weight heparins and mechanical means of prophylaxis as reasonable options for patients undergoing spinal fusion surgery. The American College of Chest Physicians 8th Guidelines on Antithrombotic and Thrombolytic Therapy recommend administering VTE prophylaxis after spinal fusion surgery, but do not specify whether mechanical or pharmacologic modalities are preferred [4]. The National Institute of Health and Clinical Excellence guidelines recommend that mechanical prophylaxis plus low molecular weight heparin be provided to patients with at least one risk factor for thromboembolism [6]. Finally, the North American Spine Society recommends using mechanical prophylaxis after spine surgery, but recommends withholding low molecular weight heparin for routine procedures unless there are additional risk factors for VTE [7]. Perhaps related to the lack of evidence, surveys of real-world surgical practice suggest variability among individual surgeons [8].

Because of the lack of evidence to support specific recommendations for VTE prophylaxis after spine surgery, we conducted an analysis of a large, retrospective cohort of patients undergoing spine fusion to describe the prevalence and type of VTE prophylaxis use after surgery as well as the rates of postoperative VTE.


We conducted a retrospective analysis using data from 244 hospitals that participated in Premier Incorporated’s Perspective database. The Perspective database is a voluntary, fee-supported database developed to measure the quality and utilization of healthcare, and is composed of mostly small to midsize non-teaching hospitals in urban locations. The database contains a date-stamped log of all items and services billed during a hospitalization, including medications, laboratory tests, and therapeutic services, as well as information about patient and hospital characteristics, primary and secondary discharge International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9) diagnostic codes, and disposition status.

We included in our analysis patients aged 18 years or older admitted to participating hospitals between 1 October 2003 and 30 September 2005 with a principal procedure code for spinal fusion surgery, subdivided into cervical fusion (codes 81.01–03 and 81.31–33), lumbar fusion (81.06–08 and 81.36–38), and other/unspecified spinal fusion (81.05, 81.34–35, 81.00, 81.30, and 81.39). Surgical approach was categorized into anterior, posterior, circumferential, and other/unknown, on the basis of ICD-9 codes. We also searched for secondary codes that indicated the number of vertebrae fused: fusion of two or three vertebrae (81.62) and fusion of four or more vertebrae (81.62 and 81.64).

Demographic and hospital data, including patient age, sex, race/ethnicity, insurance status, admission type (emergency, outpatient, or transfer from another acute hospital), and hospital region, were obtained from the database, in addition to a general measure of illness severity calculated with apr-drg software (version 15.0, 3M™, Salt Lake City, UT, USA), which is used to predict mortality. We searched for diagnosis codes for major medical comorbidities using software developed by the Agency of Healthcare Research and Quality [9], as major medical illnesses increase overall VTE risk [10]. Finally, we searched for individual comorbid conditions that were part of a validated VTE risk score described by Caprini et al. [11]. The VTE risk score included the following risk factors: myocardial infarction/congestive heart failure, age, varicose veins, immobility, obesity, hyperviscosity/hypercoagulable syndromes, estrogen therapy, history of venous thromboembolism, and malignancy. This VTE risk score has been validated in other surgical settings [12].

Prophylaxis for venous thromboembolism

We searched the database for all charges corresponding to provision of either mechanical or pharmacologic VTE prophylaxis administered at any point within the first 7 days after the index spinal surgery date. This time period was chosen because the majority of patients undergoing surgery were discharged within 7 days. Mechanical prophylaxis was considered to be present when billing charges for intermittent compression devices or antiembolism stockings of the lower extremities were identified. Pharmacologic prophylaxis was considered to be present if any of the following medications were dispensed: injectable heparin sodium at doses between 5000 and 7500 units, low molecular weight heparins (specifically enoxaparin and dalteparin, as there were no other types of low molecular weight heparin administered), and fondaparinux. We excluded 6067 patients who received either intravenous heparin or warfarin sodium within the first 7 days, because these agents are not used for routine VTE prophylaxis, and because it was not possible to determine whether their use was for pre-existing conditions requiring full-dose anticoagulation.

Thromboembolic and hemorrhagic outcomes

We searched for secondary discharge diagnosis codes consistent with deep vein thrombosis (DVT) or pulmonary embolism (PE) during the index hospitalization (ICD-9 codes: 415.1x, 451.1x, 451.2, 451.81, 451.9, 453.1, 453.2, 453.40, 453.41, 453.42, 453.8, 453.9, and 997.2 with a secondary diagnosis of DVT or PE). In addition, we determined whether patients were readmitted within 30 days with a primary discharge diagnosis code for VTE. Adverse surgical outcomes were identified by searching for specific ICD-9 diagnosis and procedure codes. Postoperative hemorrhages were identified by using codes for hemorrhage/hematoma complicating a procedure (998.1, 998.11, and 998.12) and extradural hemorrhage (852.4 and 432.0). We also searched for codes associated with other surgical complications, specifically wound disruption (998.3, 998.31, and 998.32), postoperative infection (998.5, 998.51, and 998.59), and other complications (998.13, 998.7, and 998.8), and for procedure codes associated with control of hemorrhage and reoperation (39.98, 39.99, 83.14, 83.19, 83.39, 83.44, 83.49, 86.04, 86.22, 86.28, 96.58, 96.59, 97.15, and 97.16). Finally, we searched for charges indicating transfusion of ≥ 2 units of packed red blood cells within the first 4 days after the index surgery.

Statistical analysis

Rates of pharmacologic prophylaxis, mechanical prophylaxis, both forms and no prophylaxis after surgery were determined for different categories of spinal fusion. Bivariable analysis with chi-squared tests for categorical variables, and t-tests and non-parametric tests for continuous variables, were performed to test the association between subject/hospital characteristics and receipt of prophylaxis. We then developed multivariable logistic regression models to test the independent association between clinical characteristics and receipt of VTE prophylaxis. Candidate variables significant at < 0.2 were considered for model inclusion. Generalized estimating equations, with the use of proc genmod (SAS Institute, Cary, NC, USA), were used to account for potential clustering effects at the hospital and individual provider level. We next developed multivariable logistic regression models using generalized estimating equations modeling the association of clinical characteristics and receipt of prophylaxis with the likelihood of developing diagnosed VTE within 30 days, with the significance level set at 0.05.

We recognized that patients did not receive VTE prophylaxis at random, which introduced the threat of treatment and allocation biases unaccounted for in our base multivariable models. To address these biases, we developed a propensity score representing the likelihood that each patient received VTE prophylaxis. The propensity score was derived with a separate multivariable logistic regression model including all available patient and hospital predictor covariates, where covariates for the propensity score were retained at a significance level of < 0.2 [13]. The final propensity score was then included as a separate covariate in the multivariable models modeling VTE outcomes. All analyses were performed with sas version 9.2 (SAS Institute).


We identified a total of 80 183 spinal fusions performed between 1 October 2003 and 30 September 2005 in 244 hospitals. The median number of spinal fusions performed by each hospital was 223 (interquartile range 57–466). The most common procedures were cervical fusions (39 292, 49.0%), followed by lumbar fusions (38 309, 47.8%), and then miscellaneous fusions (2582, 3.2%), which were primarily thoracic spinal fusions and fusions not otherwise specified (Table 1). The mean age of subjects was 53.3 years, and 45.9% were male. The median length of hospital stay was 3 days, and most patients were discharged home after the hospitalization (Table 1).

Table 1.   Characteristics of subjects who received prophylaxis for venous thromboembolism after spinal fusion surgery as compared with those who did not receive prophylaxis
 All (n = 80 183)Received prophylaxis (pharmacologic and/or mechanical) (n = 48 605)Did not receive prophylaxis (n = 31 578)P-value
  1. Data are given as N (%) or mean ± standard deviation, unless otherwise stated.

Age (years)53.3 ± 13.853.8 (13.9)52.6 (13.4)< 0.001
Male36 775 (45.9)21 975 (45.2)14 800 (46.9)< 0.001
 White60 067 (74.9)36 645 (75.4)23 422 (74.2)< 0.001
 Black6424 (8.0)4017 (8.3)2407 (7.6)
 Hispanic1771 (2.2)1014 (2.1)757 (2.4)
 Other11 921 (14.9)6929 (14.3)4992 (15.8)
Primary payer
 Medicare21 837 (27.2)13 950 (28.7)7887 (25.0)< 0.001
 Medicaid3569 (4.5)2120 (4.4)1449 (4.6)
 Managed care32 264 (40.2)19 246 (39.6)13 018 (41.2)
 Indemnity18 802 (23.4)11 304 (23.3)7498 (23.7)
 Uninsured1136 (1.4)610 (1.3)526 (1.7)
 Capitated814 (1.0)312 (0.6)502 (1.6)
 Other1761 (2.2)1063 (2.2)698 (2.2)
Admitting source
 Emergency room3906 (4.9)2166 (4.5)1740 (5.5)< 0.001
 Outpatient75 560 (94.2)46 002 (94.7)29 558 (93.6)
 Transfer714 (0.9)436 (0.9)278 (0.9)
Standard metropolitan statistical area
 Rural5062 (6.3)3283 (6.8)1779 (5.6)< 0.001
 Urban75 121 (93.7)45 322 (93.3)29 799 (94.4)
 Midwest18 012 (22.5)10 696 (22.0)7316 (23.2)< 0.001
 Northeast8705 (10.9)47 86 (9.9)3919 (12.4)
 South40 266 (50.2)26 156 (53.8)14 110 (44.7)
 West13 200 (16.5)6967 (14.3)6233 (19.7)
Number of hospital beds
 0–99582 (0.7)299 (0.6)283 (0.9)< 0.001
 100–1994595 (5.7)3127 (6.4)1468 (4.7)
 200–2999108 (11.4)6475 (13.3)2633 (8.3)
 300–39915 838 (19.8)10 067 (20.7)5771 (18.3)
 400–49913 102 (16.3)8767 (18.0)4335 (13.7)
 ≥ 50036 958 (46.1)19 870 (40.9)17 088 (54.1)
Teaching hospital57 705 (72.0)35 605 (73.3)22 100 (70.0)< 0.001
Type of spinal fusion surgery
 Cervical39 292 (49.0)20 717 (42.6)18 575 (58.8)< 0.001
 Lumbar38 309 (47.8)26 084 (53.7)12 225 (38.7)
 Other2582 (3.2)1804 (3.7)778 (2.5)
Surgical approach
 Anterior41 897 (52.3)23 008 (47.3)18 889 (59.8)< 0.001
 Posterior3630 (45.2)24 237 (49.9)11 993 (38.0)
 Circumferential1221 (1.5)847 (1.7)374 (1.2)
 Other/unknown835 (1.0)513 (1.1)322 (1.0)
Levels of vertebrae fused
 155 (0.1)34 (0.1)21 (0.1)< 0.001
 2–364 480 (80.4)38 842 (79.9)25 638 (81.2)
 ≥ 410 624 (13.2)6913 (14.2)3711 (11.8)
 Not specified5024 (6.3)2816 (5.8)2208 (7.0)
Comorbid conditions
 Hypertension30 878 (38.5)19 463 (40.0)11 415 (36.2)< 0.001
 Chronic pulmonary disease10 235 (12.8)6358 (13.1)3877 (12.3)< 0.001
 Diabetes mellitus9915 (12.4)6143 (12.6)3772 (12.0)0.004
 Metastatic cancer319 (0.4)223 (0.5)96 (0.3)< 0.001
 Solid tumor without metastasis298 (0.4)197 (0.4)101 (0.3)0.05
 Diagnosed obesity5379 (6.7)3299 (6.8)2080 (6.6)0.27
 Depression7641 (9.5)4993 (10.3)2648 (8.4)< 0.001
 Coagulopathy689 (0.9)489 (1.0)200 (0.6)< 0.001
 Paralysis1357 (1.7)861 (1.8)496 (1.6)0.03
Caprini et al. [11] venous thromboembolism risk score
 ≤ 412 920 (16.1)7625 (15.7)5295 (16.8)< 0.001
 536 775 (45.9)21 626 (44.5)15 149 (48.0)
 615 738 (19.6)9728 (20.0)6010 (19.0)
 711 519 (14.4)7457 (15.3)4062 (12.9)
 82196 (2.7)1472 (3.0)724 (2.3)
 9817 (1.0)558 (1.2)259 (0.8)
 ≥ 10218 (0.3)139 (0.3)79 (0.3)
Diagnosis-related group severity score
 148 648 (60.7)27 983 (57.6)20 665 (65.4)< 0.001
 224 771 (30.9)15 875 (32.7)8896 (28.2)
 35949 (7.4)4159 (8.6)1790 (5.7)
 4815 (1.0)588 (1.2)227 (0.7)
Length of stay (median days) (interquartile range)3 (1–4)3 (1–5)2 (1–4)< 0.001
Any intensive care unit stay7710 (9.6)4878 (10.0)2832 (9.0)< 0.001
Discharge status
 To home70 006 (87.3)41 418 (85.2)28 588 (90.5)< 0.001
 Transfer to another acute facility235 (0.3)162 (0.3)73 (0.2)
 Skilled nursing facility/rehabilitation center9434 (11.8)6722 (13.8)2712 (8.6)
 Dead/hospice177 (0.2)116 (0.2)61 (0.2)
 Other331 (0.4)187 (0.4)144 (0.5)

VTE prophylaxis, either mechanical or pharmacologic, was administered in the hospital setting to 60.6% of subjects within the first 7 days after surgery. Mechanical prophylaxis alone was the most common practice (47.6%), and 13.0% of subjects received pharmacologic prophylaxis (either alone or in combination with mechanical prophylaxis). A total of 39.4% of subjects received neither mechanical nor pharmacologic prophylaxis. The most common drug used in patients receiving pharmacologic prophylaxis was heparin (66%), followed by low molecular weight heparins (37%). Fondaparinux was rarely used, being found for only 0.1% of all patients receiving pharmacologic prophylaxis. Patients undergoing lumbar fusion were more likely than those undergoing cervical fusion to receive pharmacologic prophylaxis: 20.2% as compared with 4.8%, P < 0.01.

There was wide variation in the proportion of subjects receiving prophylaxis after spinal surgery across the 244 hospitals. On average, hospitals provided VTE prophylaxis to 61.7% (standard deviation 32.4%) of their patients undergoing spinal fusion, and 26.2% of hospitals administered prophylaxis to ≥ 90% of their patients. However, 33.2% of hospitals administered prophylaxis to fewer than 50% of their patients postoperatively (Fig. 1).

Figure 1.

 Proportion of 244 US hospitals providing venous thromboembolism prophylaxis to patients after spinal fusion surgery.

Factors associated with receipt of thromboembolism prophylaxis

Table 1 presents the bivariate associations between clinical variables and receipt of prophylaxis. Owing to the large size of the dataset, many associations were statistically significant. All factors that reached a significance level of < 0.2 on bivariate analysis were entered into a multivariable model, adjusted for clustering at the hospital and surgeon level. There were no missing data for any of the final included variables. The factors that were significantly associated with VTE prophylaxis in the multivariable model were undergoing lumbar or other fusions as compared with cervical fusions, insurance type, depression, higher DRG severity of illness score, higher Caprini VTE risk score, and an intensive care unit stay (Table 2).

Table 2.   Factors associated in the multivariable model with postoperative receipt of venous thromboembolism prophylaxis among 80 183 subjects who underwent spinal fusion surgery
 Adjusted odds ratio (95% CI)*
  1. CI, confidence interval. *The multivariable model included all of the variables listed in the table, and was adjusted for clustering at the hospital and surgeon levels.

Surgery type
 Cervical fusionReferent
 Lumbar fusion1.76 (1.58–1.96)
 Other fusion1.59 (1.41–1.79)
Insurance status
 Uninsured0.88 (0.80–0.97)
 Indemnity1.05 (0.98–1.11)
 Capitated1.03 (0.84–1.26)
 Managed care1.02 (0.97–1.07)
 Medicaid1.00 (0.94–1.06)
 Other1.03 (0.94–1.13)
Comorbid conditions
 Depression1.04 (1.00–1.09)
Caprini et al. [11] venous thromboembolism risk score1.03 (1.01–1.05)
Diagnosis-related group severity score
 21.08 (1.04–1.11)
 31.22 (1.10–1.35)
 41.51 (1.21–1.89)
Any intensive care unit stay1.15 (1.06–1.25)

There were 5414 patients who received both pharmacologic and mechanical prophylaxis during their hospitalization. On multivariable analysis, patients more likely to receive combined prophylaxis strategies had risk factors for thromboembolism, including undergoing lumbar fusion (adjusted odds ratio [OR] 4.7, 95% confidence interval [CI] 3.8–5.7), as compared with cervical fusion), surgery involving ≥ 4 vertebral levels (OR 1.2, 95% CI 1.1–1.3, as compared with two or three levels), an intensive care unit stay (OR 1.6, 95% CI 1.4–1.7), higher severity of illness (OR 3.2, 95% CI 1.4–1.7, for a DRG score of 4 as compared with 1), and higher VTE risk score (OR 1.1, 95% CI 1.1–1.2).

Outcome events

There were 222 DVTs and 149 PEs diagnosed in 359 patients in our study, for an overall cohort rate of 0.45%. Of patients with VTE, 56% were diagnosed during the index hospitalization, and an additional 44% were diagnosed after discharge but within 30 days of the index surgery. The rate of postoperative hemorrhage was 1.1%, and only 1.2% of patients had ≥ 2 units of transfused blood in the first 4 days after surgery (Table 3).

Table 3.   Unadjusted rates of diagnosed thromboembolism, surgical complications, blood transfusions and 30-day readmission after spinal fusion surgery, stratified by type of thromboembolism prophylaxis
 Type of venous thromboembolism prophylaxisAll patients, N (%)
Pharmacologic alone, N (%)Mechanical alone, N (%)Both pharmacologic and mechanical, N (%)No prophylaxis, N (%)
  1. DVT, deep vein thrombosis; PE, pulmonary embolism.

Cervical fusionN = 1208N = 18 827N = 682N = 18 575N = 39 292
 DVT/PE23 (1.9)46 (0.2)15 (2.2)36 (0.2)120 (0.3)
 Surgical site bleeding15 (1.2)135 (0.7)9 (1.3)138 (0.7)297 (0.8)
 Other surgical complications18 (1.5)98 (0.5)6 (0.9)104 (0.6)226 (0.6)
 Transfusions11 (0.9)44 (0.2)10 (1.5)38 (0.2)103 (0.3)
 Readmission in 30 days78 (6.4)570 (3.0)43 (6.3)559 (3.0)1250 (3.2)
Lumbar fusionN = 3384N = 18 333N = 4367N = 12 225N = 38 309
 DVT/PE28 (0.8)69 (0.4)45 (1.0)48 (0.4)190 (0.5)
 Surgical site bleeding58 (1.7)238 (1.3)77 (1.8)153 (1.3)526 (1.4)
 Other surgical complications71 (2.1)304 (1.7)86 (2.0)210 (1.7)671 (1.8)
 Transfusions49 (1.4)398 (2.2)98 (2.2)238 (2.0)783 (2.0)
 Readmission in 30 days195 (5.8)897 (4.9)247 (5.7)626 (5.1)1965 (5.3)
Other spinal fusionN = 444N = 995N = 365N = 778N = 2582
 DVT/PE20 (4.5)7 (0.7)8 (2.2)14 (1.8)49 (1.9)
 Surgical site bleeding12 (2.7)26 (2.6)17 (4.7)18 (2.3)73 (2.8)
 Other surgical complications10 (2.3)25 (2.5011 (3.0)23 (3.0)69 (2.7)
 Transfusions14 (3.2)32 (3.2)18 (4.9)35 (4.5)99 (3.8)
 Readmission in 30 days47 (10.6)96 (9.7)41 (11.2)67 (8.6)251 (9.7)

Factors associated with higher thromboembolism rates on unadjusted analysis included surgery on multiple vertebral levels (0.88% for operations involving ≥ 4 levels vs. 0.38% with < 4 levels, P < 0.001), and lumbar and other spinal fusions (0.5% and 1.9%, respectively, as compared with 0.3% after cervical fusions, P < 0.001). Higher unadjusted thromboembolism rates were also observed in subjects who received pharmacologic or pharmacologic plus mechanical prophylaxis than in subjects who did not receive prophylaxis (Table 3). Despite adjusting for the propensity to receive VTE prophylaxis, we continued to observe in multivariable models a higher risk of postoperative VTE among patients who received prophylaxis (Table 4).

Table 4.   Factors associated in multivariable models with postoperative and 30-day diagnosis of venous thromboembolism among 80 183 subjects who underwent spinal fusion surgery; results from multivariable models with and without adjustment for the propensity of receiving venous thromboembolism prophylaxis
  1. CI, confidence interval; OR, odds ratio.

 Multivariable model results, n = 80 183
OR (95% CI)
Multivariable model results adjusted for propensity score, n = 80 183
OR (95% CI)
Type of prophylaxis
 Pharmacologic2.3 (1.6–3.3)2.3 (1.6–3.3)
 Mechanical1.0 (0.8–1.3)1.0 (0.8–1.3)
 Both2.3 (1.6–3.2)2.2 (1.6–3.1)
Diagnosis-related group severity of illness score
 22.4 (1.7–3.4)2.3 (1.6–3.4)
 38.2 (5.8–11.6)7.9 (5.4–11.6)
 423.6 (14.4–38.7)22.3 (12.8–38.8)
Any intensive care unit stay1.3 (1.0–1.7)1.3 (1.0–1.7)
Congestive heart failure0.4 (0.2–0.7)0.5 (0.2–0.7)
Coagulopathy1.7 (1.1–2.7)1.7 (1.1–2.6)
Hospital bed size
 400–4991.1 (0.7–1.5)1.1 (0.7–1.5)
 > 5001.4 (1.1–1.9)1.5 (1.1–2.00)
Caprini et al. [11] venous thromboembolism risk score1.3 (1.2–1.4)1.3 (1.2–1.4)

We identified 896 (1.1%) surgical site hemorrhages in the cohort. On bivariate analysis, receipt of pharmacologic prophylaxis was associated with greater unadjusted odds of surgical site bleeding than receiving no prophylaxis: OR 1.7 (95% CI 1.4–2.2) with pharmacologic therapy alone, and OR 2.0 (95% CI 1.6–2.5) for pharmacologic plus mechanical prophylaxis. However, when we adjusted for other clinical factors (including age, sex, type of surgery, number of fusion levels, and DRG severity score), the association was not statistically significant: adjusted OR 1.1 (95% CI 0.8–1.5) for pharmacologic therapy alone, and OR 1.2 (95% CI 0.9–1.5) for pharmacologic plus mechanical therapy.

The primary analyses excluded 6067 patients who received intravenous heparin or warfarin after their surgery, because we were unable to determine whether these patients had pre-existing VTE. However, because excluding these patients may also have led to the exclusion of patients treated for incident thromboembolism during the hospitalization, we conducted an additional analysis including individuals who received intravenous heparin or warfarin during their hospitalization. When these patients were included, the total number of patients with diagnosis codes for VTE increased to 674, for an overall cohort VTE rate of 0.78%. The number of surgical site hemorrhages increased to 1057, or 1.2% of the cohort, and the number of patients with ≥ 2 units of transfused blood increased to 1142 (1.3% of the cohort).


In this large sample of patients undergoing spinal fusion surgery, 39.4% received neither pharmacologic nor mechanical forms of VTE prophylaxis in the postoperative period. Significant practice variation was observed across hospitals in terms of the proportion of spinal surgery patients receiving prophylaxis. Overall rates of diagnosed VTE after spinal fusions were low, even among patients who did not receive VTE prophylaxis, as were rates of surgical site bleeding.

Prophylaxis was more commonly administered to patients with a higher baseline risk for VTE. Patients undergoing lumbar or thoracic fusion surgery were more likely to receive prophylaxis than patients undergoing cervical spine fusion surgery. In addition, physicians were more likely to administer prophylaxis to patients with a higher estimated VTE risk score and to patients who had greater cumulative severity of illness. These findings probably reflect selective administration of prophylaxis to patients who were considered more likely to develop VTE postoperatively. However, these patterns were not consistent across all participating sites, as there was wide variability in the proportion of patients receiving postoperative prophylaxis across individual medical centers.

There is no consensus regarding the optimal choice of VTE prophylaxis after spinal fusion surgery. Prior small studies have reported generally low rates of clinically significant thromboembolism after spinal surgery in patients receiving mechanical prophylaxis, ranging from 0.0% to 4% [14–20]. Asymptomatic DVT rates may be higher, however, with one study using venography to screen patients undergoing spine surgery reporting a DVT rate of 26.5% after lumbar surgery [21]. A meta-analysis of 4383 patients undergoing elective spine surgery found a DVT rate of 1.09% and a PE rate of 0.06%; this analysis also found pharmacologic prophylaxis to be more effective in preventing VTE than mechanical or no prophylaxis, at the expense of a higher rate of epidural hematomas [22]. The 0.45% rate of diagnosed postoperative VTE in our study compares favorably with the 0.81% rate of symptomatic DVT in medical patients not receiving anticoagulant prophylaxis, and the 2–5% symptomatic VTE rate after elective hip arthroplasty [4,23].

The rate of any postoperative hemorrhage in our cohort was 1.1%, and 1.2% of patients received ≥ 2 units of transfused blood in the first 4 days after surgery. In comparison, 0.4% of 1954 spine surgery patients who received nadroparin for VTE prophylaxis developed major hemorrhage postoperatively [24]. We lacked detailed information on the location or management of hemorrhages in our cohort, and so were unable to determine the severity or consequences of such bleeds. However, concerns about hemorrhagic complications, particularly epidural hemorrhages, may contribute to cautious use of pharmacologic prophylaxis after spinal procedures.

Our analyses found, paradoxically, higher rates of VTE in subjects who received pharmacologic prophylaxis than in those who received neither pharmacologic nor mechanical prophylaxis. Randomized trials have clearly shown that pharmacologic prophylaxis reduces VTE rates in medical and surgical patients [4], and our findings should not be used to suggest a lack of efficacy with VTE prophylaxis. The higher VTE rates observed in patients who received pharmacologic prophylaxis in our study are probably attributable to confounding by indication, whereby patients at higher baseline risk for thrombosis were more likely to receive more effective forms of VTE prophylaxis. This supposition is supported by the observations that patients with higher severity of illness and higher estimated VTE risk scores were more likely to receive combined mechanical and pharmacologic prophylaxis. Although we attempted to control for baseline risk for VTE and minimize confounding by indication through the use of a propensity score analysis, residual confounding remains a major limitation of observational studies such as ours. Thus, the direction and effect size of the ORs from our multivariable analysis of the efficacy of VTE prophylaxis should be interpreted with caution, and our study should not be used to draw firm conclusions about the relative efficacy and safety of different approaches to VTE prophylaxis.

There are a number of limitations to this analysis. We used a window of 7 days to assess for VTE prophylaxis, although the median length of stay was shorter than this. Patients with uncomplicated procedures may have been discharged prior to consideration of VTE prophylaxis. The diagnoses of postoperative VTE and hemorrhage were dependent on ICD-9 codes. Although prior studies have shown that ICD-9 codes are sensitive for the presence of clinical VTE, the positive predictive value may be lower when the codes are in the secondary position [25]. In addition, reliance on administrative codes may lead to underestimation of the true prevalence of VTE, as asymptomatic clots were unlikely to be detected. We were not able to assess for severity of hemorrhage in our study, although rates of transfusion and reoperation were low. Patients who presented with VTE after discharge to an institution other than the index hospital or who were managed only in the outpatient or emergency department setting would not have been captured by our study, potentially leading to underestimation of the true rate of VTE. Our primary analysis excluded patients who received intravenous heparin or warfarin after their surgery, which may have also led us to underestimate the rate of events. Finally, as discussed above, our study was an observational study of actual clinical practice where VTE prophylaxis was not randomly assigned to patients. Residual confounding resulting from non-randomized allocation of prophylaxis prevents our study from being able to support one thromboprophylaxis strategy over another.

A large proportion of patients undergoing spinal fusion surgery receive neither pharmacologic nor mechanical means of thromboembolism prophylaxis in the USA. However, rates of diagnosed thromboembolism after spinal fusion surgery appear to be low, even among patients who received no VTE prophylaxis. These low rates of adverse outcomes support the lack of definitive recommendations regarding the optimal form of VTE prophylaxis after spinal fusion surgery.


This study was supported by National Institutes of Health grants K23 AG028978 (M. C. Fang) and K24 HL098372 (A. D. Auerbach). The sponsor had no role in the design and conduct of the study, the collection, management, analysis and interpretation of the data, or the preparation, review or approval of the manuscript. The corresponding author had full access to all of the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Disclosure of Conflict of Interests

J. D. Lurie has served as a consultant for the Foundation for Informed Medical Decision Making, Blue Cross-Blue Shield, Sanofi-Aventis, and Regeneron. The other authors state that they have no conflict of interest.