841600

research-article2019
HPI0010.1177/1120700019841600HIP InternationalBala et al.

Original Research Article

HIP HIP
International

HIP International

Venous thromboprophylaxis after total

1­–8
© The Author(s) 2019
Article reuse guidelines:
hip arthroplasty: aspirin, warfarin, sagepub.com/journals-permissions
https://doi.org/10.1177/1120700019841600
DOI: 10.1177/1120700019841600

enoxaparin, or factor Xa inhibitors? journals.sagepub.com/home/hpi

Abiram Bala, Marlon J Murasko, David R Burk,

James I Huddleston 3rd, Stuart B Goodman,
William J Maloney and Derek F Amanatullah

Abstract
Introduction: Debate over the ideal agent for venous thromboembolism (VTE) prophylaxis after total hip arthroplasty
(THA) has led to changes in prescribing trends of commonly used agents. We investigate variation in utilisation and the
differences in VTE incidence and bleeding risk in primary THA after administration of aspirin, warfarin, enoxaparin, or
factor Xa inhibitors.
Methods: 8829 patients were age/sex matched from a large database of primary THAs performed between 2007 and
2016. Utilisation was calculated using compound annual growth rate. Incidence of postoperative deep venous thrombosis
(DVT), pulmonary embolism (PE), bleeding-related complications, postoperative anaemia, and transfusion were identified
at 2 weeks, 30 days, 6 weeks, and 90 days.
Results: Aspirin use increased by 33%, enoxaparin by 7%, and factor Xa inhibitors by 31%. Warfarin use decreased by
1%. Factor Xa inhibitors (1.7%) and aspirin (1.7%) had the lowest incidence of DVT followed by enoxaparin (2.6%), and
warfarin (3.7%) at 90 days. Factor Xa inhibitors (12%) and aspirin (12%) had the lowest incidence of blood transfusion
followed by warfarin (15%) and enoxaparin (17%) at 90 days. There was no difference in incidence of blood transfusion
or bleeding-related complications nor any detectable difference in symptomatic PE incidence.
Conclusions: The utilisation of aspirin and factor Xa inhibitors increased over time. Aspirin and factor Xa inhibitors
provided improved DVT prophylaxis with lower rates of postoperative anaemia compared to enoxaparin and warfarin.

Keywords
Anticoagulants, total hip arthroplasty, total hip replacement, thromboprophylaxis, VTE in total joint arthroplasty

Date received: 27 June 2018; accepted: 28 February 2019

Introduction balance between symptomatic VTE prevention and poten-

Prophylaxis for venous thromboembolism (VTE), which
includes deep venous thrombosis (DVT) and pulmonary
embolism (PE), remains an important aspect of routine
postoperative care after total hip arthroplasty (THA).
Increased utilisation of chemical prophylaxis and/or
mechanical compression has reduced the risk of sympto-
matic VTE in the perioperative period.1,2
Both the American Academy of Orthopaedic Surgeons
and the American College of Chest Physicians recommend
the use of a chemical VTE prophylaxis agent during the
perioperative period. However, despite significant analysis
of the current literature, there is no consensus regarding
the preferred therapy especially when considering the
tial bleeding complications.1,3–7 Availability of novel anti-
coagulants have unfortunately led to further confusion.8,9
Each VTE prophylaxis agent after primary THA has
been well studied in isolation or in comparison to another
agent, though the comparisons often omit one of the
commonly used agents.10–17 Given that the incidence of
Department of Orthopaedic Surgery, Stanford Hospital and Clinics,
Redwood City, CA, USA
Corresponding author:
Derek F Amanatullah, Department of Orthopaedic Surgery, Stanford
Hospital and Clinics, 450 Broadway Street, Redwood City, CA 94063-
6342, USA.
Email: dfa@stanford.edu
2 HIP International 00(0)
symptomatic VTE is relatively low, it is difficult to acquire
sufficient statistical power to discern differences between
agents. As such, the multitude of meta-analyses comparing
symptomatic VTE outcomes reinforces the inadequacy of
the retrospective and prospective studies in the area of
thromboprophylaxis.14,18–21 Many studies have further lim-
itations: low patient volumes, variety in outcome meas-
ures, differences in protocol, and inherent variability in
clinical practice. The ongoing Comparative Effectiveness
of Pulmonary Embolism Prevention after Hip and Knee
Replacement (PEPPER) trial might provide the highest-
quality answer once data are available, though it does not
include enoxaparin in its comparison.22 To our knowledge,
no study has directly compared aspirin, enoxaparin, warfa-
rin, and factor Xa inhibitors for thromboprophylaxis after
THA.
The purpose of this study was to answer the following
3 questions: are there differences in postoperative sympto-
matic VTE incidence between aspirin, enoxaparin, warfa-
rin, and factor Xa inhibitors; what are the differences in
bleeding risks between these agents; how has the utilisa-
tion of each agent changed over the 2007 to 2015 study
period given nationwide changes in guidelines?
Materials and methods
Our study used completely de-identified data and was
therefore exempt from institutional review board approval.
We retrospectively queried a combined Medicare and pri-
vate-payer Humana database with records of over 16 mil-
lion patients from 2007 to Quarter 1 of 2016 using
PearlDiver Technologies (Colorado Springs, CO, USA).
This consists specifically of administrative claims data
from patients enrolled in individual commercial insurance
with Humana and enrolled with dual coverage in Humana
and Medicare Advantage plans. Patients from all regions
of the United States are represented, though there was a
comparatively higher proportion of coverage in the south-
eastern United States. We used International Classification
of Diseases, 9th revision (ICD-9) procedural code 81.51
and current procedural terminology (CPT) code 27130 to
identify all primary THAs. We only selected THAs per-
formed between 2007 and Quarter 4 of 2015, which guar-
anteed 90 days of time for follow-up after any THA
identified. We only selected patients who had a single pri-
mary THA during the study period to allow for accurate
tracking of prescriptions and complications.
PearlDiver enables users to identify prescription bill-
ings for specified medications and allows for the tracking
of this billing event before or after a procedure. We first
identified all possible anticoagulant and antiplatelet agents
searchable in the database: aspirin, enoxaparin, warfarin,
apixaban, rivaroxaban, fondaparinux, heparin, dabigatran,
clopidogrel, dipyridamole, prasugrel, ticagrelor, ticlodi-
pine, and cilostazol. We excluded patients with prior
symptomatic VTE. Additionally, we excluded any patient
who had a prescription filled for 1 of the aforementioned
medications within 1 year prior to their THA; however, we
were not able to collect prior year prescription data from
2006 for patients who had primary THAs performed in
2007. We grouped all factor Xa inhibitors into a single
cohort (i.e. apixaban, rivaroxaban, and fondaparinux), and
identified patients who had either aspirin, enoxaparin,
warfarin, or a factor Xa inhibitor prescribed during admis-
sion or within 14 days of discharge after THA. Finally, we
matched these 4 study groups by age and sex, a process
that selects a random sample from an initial larger cohort
to fit the desired demographic proportions based on the
smallest sized cohort.
Incidences for symptomatic DVT and PE were queried
with the respective ICD-9 diagnosis codes at 2 weeks,
30 days, 6 weeks, and 90 days. We identified bleeding risk
incidence by searching for incidence of postoperative
anaemia, transfusion rate, and bleeding-related complica-
tions. Anaemia was defined as specific diagnosis of acute
post-haemorrhagic anaemia or acute postoperative anae-
mia. Transfusion was defined as all autologous and alloge-
neic blood transfusions. Bleeding-related complications
were collectively defined as: haemorrhage, haematoma,
haemarthrosis, or bleeding requiring surgical interven-
tion.23 We utilised the “first instance” command in the lan-
guage to search for these complications. This allowed us to
identify any VTE or bleeding code that appeared in a
patient’s record exclusively for the first time, so that pre-
existing symptomatic VTE or bleeding diagnoses did not
confound the postoperative outcomes.
Comorbidities were compared using the Charlson
Comorbidity Index (CCI) and a standardised list of
Elixhauser comorbidities.24,25 The ICD-9 and CPT codes
used in this study are found in the appendix.
A total of 649 patients who had aspirin thromboprophy-
laxis alone were age and sex matched to 3377 patients with
enoxaparin, 3245 patients with warfarin, and 1558 patients
with factor Xa inhibitors, for a total study size of 8829
patients (Table 1). There were no differences in the propor-
tion of age and sex between the groups due to the matching
system. The median CCI, which serves as a proxy for
10-year survival, was identical for aspirin (1, range 0–13),
enoxaparin (1, range 0–17), warfarin (1, range 0–19), and
factor Xa inhibitors (1, range 0–18), indicating similar
overall health status. The Elixhauser comorbidity profile
for each group (Table 2) was also evaluated. Despite the
equivalence in CCI, there were differences in comorbidi-
ties among the aspirin, enoxaparin, warfarin, and factor Xa
inhibitors.
The PearlDiver database maintains Health Insurance
Portability and Accountability Act (HIPAA) compliance.
Since small numbers patients could theoretically be identi-
fied using the PearlDiver database, we wanted to ensure
patient confidentiality, hence any query returning a group
Bala et al. 3

Table 1.  Matched cohort demographics.

Aspirin Enoxaparin Warfarin Factor Xa inhibitors p-value

Age 40–44 (yrs) 14 (2%) 73 (2%) 70 (2%) 34 (2%) 1.00
Age 45–49 47 (7%) 244 (7%) 235 (7%) 112 (7%) 1.00
Age 50–54 77 (12%) 401 (12%) 385 (12%) 185 (12%) 1.00
Age 55–59 133 (20%) 692 (20%) 665 (20%) 319 (20%) 1.00
Age 60–64 146 (22%) 759 (22%) 730 (22%) 351 (22%) 1.00
Age 65–69 107 (16%) 557 (16%) 535 (16%) 257 (16%) 1.00
Age 70–74 70 (11%) 364 (11%) 350 (11%) 168 (11%) 1.00
Age 75–79 42 (6%) 219 (6%) 210 (6%) 101 (6%) 1.00
Age 80–84 13 (2%) 68 (2%) 65 (2%) 31 (2%) 1.00
Female 307 (47%) 1598 (47%) 1535 (47%) 736 (47%) 1.00
Male 342 (53%) 1779 (53%) 1710 (53%) 822 (53%) 1.00
Total cohort 649 3377 3245 1558  

Table 2.  Comorbidities of study cohorts.

Aspirin Enoxaparin Warfarin Factor Xa inhibitors p-value

Congestive heart failure 21 (3%) 152 (5%) 153 (5%) 73 (5%) 0.41
Valvular disease 64 (10%) 333 (10%) 253 (8%) 157 (10%) 0.01
Pulmonary circulation disorders 0 (<1%) 65 (2%) 56 (2%) 35 (2%) N/A
Peripheral vascular disease 57 (9%) 383 (11%) 290 (9%) 185 (12%) <0.01
Hypertension, uncomplicated 334 (51%) 1880 (56%) 1738 (54%) 896 (58%) 0.01
Hypertension, complicated 58 (9%) 358 (11%) 282 (9%) 144 (9%) 0.06
Hypertension, both 336 (52%) 1901 (57%) 1748 (54%) 905 (58%) <0.01
Paralysis 0 (<1%) 30 (1%) 30 (1%) 16 (1%) N/A
Other neurologic disorders 34 (5%) 263 (8%) 210 (6%) 120 (8%) 0.03
Chronic pulmonary disease 95 (15%) 682 (20%) 619 (19%) 326 (21%) <0.01
Diabetes, uncomplicated 107 (16%) 616 (18%) 558 (17%) 285 (18%) 0.52
Diabetes, complicated 47 (7%) 202 (6%) 141 (4%) 107 (7%) <0.01
Hypothyroidism 101 (16%) 506 (15%) 448 (14%) 256 (16%) 0.10
Renal failure 38 (6%) 213 (6%) 143 (4%) 114 (7%) <0.01
Liver disease 17 (3%) 150 (4%) 128 (4%) 69 (4%) 0.16
Chronic peptic ulcer disease 0 (<1%) 0 (<1%) 12 (<1%) 0 (<1%) N/A
HIV/AIDS 0 (<1%) 18 (1%) 0 (<1%) 0 (<1%) N/A
Lymphoma 0 (<1%) 34 (1%) 19 (1%) 0 (<1%) N/A
Metastatic cancer 0 (<1%) 52 (2%) 24 (1%) 19 (1%) N/A
Solid tumour without metastasis 42 (6%) 238 (7%) 210 (6%) 111 (7%) 0.74
RA/CVD 56 (9%) 404 (12%) 365 (11%) 214 (14%) <0.01
Coagulation deficiency 26 (4%) 90 (3%) 96 (3%) 50 (3%) 0.28
Obesity 106 (16%) 549 (16%) 496 (15%) 362 (23%) <0.01
Weight loss 19 (3%) 129 (4%) 117 (4%) 65 (4%) 0.53
Fluid and electrolyte disorders 58 (9%) 357 (11%) 359 (11%) 180 (12%) 0.30
Blood loss anaemia 0 (<1%) 47 (1%) 38 (1%) 20 (1%) N/A
Deficiency anaemias 93 (14%) 562 (17%) 474 (15%) 257 (17%) 0.08
Alcohol abuse 17 (3%) 108 (3%) 94 (3%) 39 (3%) 0.55
Drug abuse 21 (3%) 168 (5%) 96 (3%) 85 (5%) <0.01
Psychoses 43 (7%) 296 (9%) 249 (8%) 141 (9%) 0.10
Depression 76 (12%) 503 (15%) 440 (14%) 256 (16%) <0.01

HIV/AIDS, human immunodeficiency virus or acquired immune deficiency syndrome; RA/CVD, rheumatoid arthritis or collagen vascular disease.

of patients between 1 and 10 returned a “null” value. When We examined the utilisation of each agent across the
this occurred in our study, we calculated the incidence of 2007 to 2015 study period and compared yearly change by
the complication to be less than or equal to the maximum calculating the compound annual growth rate (CAGR).
possible value (10/cohort size). 4-way chi-square test was used for categorical data using
4 HIP International 00(0)

Table 3.  Utilisation and growth rate of each still had the lowest incidence of symptomatic DVT, fol-
thromboprophylaxis agent. lowed by enoxaparin (2.4%), and warfarin (3.4%). The
Year Aspirin Enoxaparin Warfarin Factor Xa symptomatic DVT incidence for aspirin was detectable at
inhibitors 90 days. By 90 days, aspirin (1.7%) and factor Xa inhibi-
tors (1.7%) had equally low incidences of symptomatic
2007 23 (4%) 265 (8%) 306 (9%) 32 (2%) DVT followed enoxaparin (2.6%), and warfarin (3.7%).
2008 47 (7%) 358 (11%) 361 (11%) 42 (3%) Due to the HIPAA limitations of the database, we could
2009 34 (5%) 345 (10%) 397 (12%) 46 (3%)
not accurately find the incidence of symptomatic PE for
2010 53 (8%) 376 (11%) 403 (12%) 47 (3%)
aspirin or factor Xa inhibitors. At 90 days, the sympto-
2011 25 (4%) 439 (13%) 406 (13%) 84 (5%)
matic PE incidence was lowest for enoxaparin (0.5%), fol-
2012 19 (3%) 317 (9%) 389 (12%) 285 (18%)
lowed by warfarin (0.7%).
2013 35 (5%) 380 (11%) 367 (11%) 317 (20%)
2014 121 (19%) 431 (13%) 342 (11%) 336 (22%)
There was a difference in the incidence of postoperative
2015 292 (45%) 466 (14%) 274 (8%) 369 (24%) anaemia among the agents at all time intervals (p < 0.01,
CAGR 33% 7% −1% 31% Table 5). The incidence of postoperative anaemia remained
relatively constant from 2 weeks until 90 days postopera-
CAGR, compound annual growth rate. tively for aspirin (23–24%), factor Xa inhibitors (26%),
warfarin (24–25%), and enoxaparin (25–26%). There was
the SNP tools statistical add-on package for Microsoft a difference in incidence of blood transfusion at all time
Excel (Redmond, WA, USA). Significance was set at an points (p < 0.01, Table 5) for aspirin (12%), factor Xa
alpha of < 0.05. A post hoc power analysis revealed that inhibitors (12%), warfarin (15%), and enoxaparin (17%).
we were powered ⩾ 80% to detect a 1.8% increase in the Due to the HIPAA limitations of the database, we could
incidence of symptomatic DVT, a 0.8% increase in the not accurately find the incidence of bleeding-related com-
incidence of symptomatic PE, 1.8% increase in the inci- plications for aspirin until 90 days (Table 5). There was no
dence of bleeding-related complications, a 4.7% increase difference in the incidence of bleeding-related complica-
in the incidence of anaemia, and a 5.8% increase in the tions at 90 days (p = 0.94) for aspirin (2%), factor Xa
incidence of transfusion between each anticoagulation inhibitors (1%), enoxaparin (1%), or warfarin (2%).
choice at each study interval.
Discussion
Results Utilisation trends of commonly used thromboprophylactic
Dramatic shifts in the utilisation of each VTE prophylactic agents reflect the lack of consensus on which agent most
agent occurred in 2012 and 2014 after the widespread adop- effectively balances the risk of symptomatic VTE versus
tion of factor Xa inhibitors, and the gradual adoption of bleeding. Aspirin and factor Xa inhibitors had the highest
aspirin as an acceptable form of VTE prophylaxis (Table 3). growth in utilisation over our study period when compared
Aspirin utilisation was stagnant until an increase in utilisa- to enoxaparin and warfarin. Aspirin or factor Xa inhibitors
tion in 2014, with a CAGR of 33% over our study period. provided improved VTE prophylaxis compared to enoxa-
Enoxaparin utilisation steadily increased until 2011. Then, it parin and warfarin with a lower risk of postoperative
dramatically decreased with the introduction of factor Xa anaemia.
inhibitors in 2012, followed by a steady increase in utilisa- Trends in utilisation for the 4 agents studied have not
tion until 2015, resulting in a CAGR of 7% over our study been examined in depth. One retrospective study found
period. Factor Xa inhibitors, which rapidly took the place of that aspirin was largely replaced as an agent of pharma-
enoxaparin and warfarin utilisation from 2012 to 2015, had cological thromboprophylaxis in the early 2000s by
a CAGR of 31% (Table 3). Warfarin utilisation peaked in enoxaparin and then warfarin after 2005, likely in
2011 and has been steadily declining, resulting in a decrease response to guideline change.26 Studies of recent utilisa-
in the CAGR of 1% over the study period. tion trends are even more limited. Another single-insti-
Due to the HIPAA limitations of the database, we could tution study found that the rates of prescribing aspirin as
not accurately find the incidence of symptomatic DVT for the VTE prophylaxis agent after total knee arthroplasty
aspirin until 90 days. There was a difference in the inci- increased after the convergence of the American
dence of symptomatic DVT among the agents at 90 days (p Academy of Orthopaedic Surgeons and the American
< 0.01, Table 4). At 2 weeks, factor Xa inhibitors (1.0%) College of Chest Physicians guidelines, while all other
had the lowest incidence of symptomatic DVT, followed agents including factor Xa inhibitors decreased.27 We
by enoxaparin (1.7%), and warfarin (2.4%). By 30 days, were unable to find a specific correlate for THA. It is
factor Xa inhibitors (1.4%) again had the lowest incidence likely that the parallel increase found in our study also
of symptomatic DVT, followed by enoxaparin (2.2%), and followed the consensus of these same guidelines. The
warfarin (3.1%). By 6 weeks, factor Xa inhibitors (1.5%) increase in factor Xa inhibitor use likely followed the
Bala et al. 5

Table 4.  Venous thromboembolic events.

Deep venous thrombosis 2 weeks 30 days 6 weeks 90 days

Aspirin <10 (<2%) <10 (<2%) <10 (<2%) 11 (2%)
Enoxaparin 57 (2%) 76 (2%) 82 (2%) 89 (3%)
Warfarin 78 (2%) 102 (3%) 109 (3%) 121 (4%)
Factor Xa Inhibitors 15 (1%) 22 (1%) 24 (2%) 27 (2%)
p-value N/A N/A N/A <0.01
Pulmonary embolism 2 weeks 30 days 6 weeks 90 days
Aspirin 0 (<1%) <10 (<2%) <10 (<2%) <10 (<2%)
Enoxaparin <10 (<1%) 12 (0.4%) 13 (0.4%) 16 (0.4%)
Warfarin 16 (0.5%) 17 (0.5%) 20 (0.6%) 24 (0.7%)
Factor Xa Inhibitors <10 (<1%) <10 <10 (<1%) <10 (<1%)
p-value N/A N/A N/A N/A

Table 5.  Risk of bleeding.

Anaemia 2 weeks 30 days 6 weeks 90 days

Aspirin 147 (23%) 149 (23%) 150 (23%) 156 (24%)
Enoxaparin 845 (25%) 856 (25%) 861 (26%) 880 (26%)
Warfarin 771 (24%) 784 (24%) 788 (24%) 801 (25%)
Factor Xa Inhibitors 406 (26%) 408 (26%) 410 (26%) 417 (26%)
p-value <0.01 <0.01 <0.01 0.01
Transfusion 2 weeks 30 days 6 weeks 90 days
Aspirin 74 (12%) 75 (12%) 76 (12%) 79 (12%)
Enoxaparin 564 (17%) 570 (17%) 572 (17%) 580 (17%)
Warfarin 468 (14%) 477 (15%) 478 (15%) 482 (15%)
Factor Xa Inhibitors 177 (11%) 180 (12%) 182 (12%) 186 (12%)
p-value <0.01 <0.01 <0.01 <0.01
Bleeding complication 2 weeks 30 days 6 weeks 90 days
Aspirin <10 (<2%) <10 (<2%) <10 (<2%) 11 (2%)
Enoxaparin 34 (1%) 44 (1%) 47 (1%) 50 (1%)
Warfarin 30 (1%) 42 (1%) 44 (1%) 49 (2%)
Factor Xa inhibitors 16 (1%) 18 (1%) 19 (1%) 21 (1%)
p-value N/A N/A N/A 0.94

results of the RECORD1, RECORD2, and ADVANCE3
trials. In these trials, factor Xa inhibitors were found to
be superior to the once highly utilised enoxaparin.15,16,28
We believe our results capture prescribing trends during
a time of changing clinical practice guidelines, provid-
ing a new contribution to the literature.
In comparison to randomised controlled trials or retro-
spective analyses, studies on large administrative data-
bases lack clinical observation, patient specific identifiers,
and the guarantee of accurately reported complications.
Furthermore, the evidence regarding the validity and accu-
racy on administrative claims data is mixed.29,30
We were unable to determine the dose and frequency of
the thromboprophylactic agents used, which are important
factors in determining their efficacy. Although we selected
patients to guarantee a minimum of 90 days of follow-up,
some patients may have been lost to follow-up due to
change in insurance or having passed away. Our inability
to identify individuals who experienced a fatal PE is one
potential limitation of the current study. Additionally, we
excluded patients who had anticoagulation or antiplatelet
agents prescribed the year before, but we may have still
captured some of these patients in 2007. We likely missed
a large proportion of patients using aspirin for two reasons:
(1) many who are prescribed aspirin may not need pre-
scription billing because aspirin can be purchased over the
counter; and (2) we initially excluded all patients pre-
scribed aspirin within 1 year before THA to prevent con-
founding our results. The authors recognise that a large
portion of the aspirin only cohort was lost with this
approach; however, it allowed us to strictly identify the
cohort of patients who had aspirin prescribed as the sole
thromboprophylactic agent after THA. In addition, the
authors also recognise that prescriptions are used as a
proxy for anticoagulant use, and we are unable to account
for adherence rates.31
The database used does not offer the patient-level data
that would have allowed us to perform a multivariate
6 HIP International 00(0)
regression to account for all potential differences in medi-
cal comorbidites. Despite this, we were able to match
each group by age and sex to find that the CCI for each
group was identical. As such, we believe that the overall
health status of each group is similar. We were also not
able to control for all risk factors specific to symptomatic
VTE that can influence the selection of a thromboprophy-
lactic agent. For example, high symptomatic VTE risk
patients may have been appropriately prescribed enoxapa-
rin, and low symptomatic VTE risk patients prescribed
aspirin. Since we excluded all patients with pre-existing
symptomatic VTE diagnoses, we may have missed the
highest risk patients. We are unable to capture in-hospital
VTE and its impact on anticoagulation.32 This certainly
warrants future study on this topic. We also are unable to
assess the impact of mechanical prophylaxis in the con-
text of these results as the protocols are unlikely to be
standardised across institutions.33 In addition, we were
unable to examine differences between various racial and
ethnic populations because of database limitations.
Another important consideration is that our transfusion
outcomes may be confounded by decreased blood transfu-
sion rates over time in THA and increased adaptation of
agents such as tranexamic acid, which occurred concur-
rently with the increased utilisation of aspirin and factor
Xa inhibitors.34 As we did not specifically look at TXA
utilisation, we are unable to study how that specifically
impacted our collective bleeding outcomes. Finally, it is
likely that certain patients may have had more frequent
surveillance in the postoperative period; for example,
patients on warfarin may require International Normalised
Ratio (INR) monitoring. This could invariably elevate the
reported symptomatic VTE and bleeding outcomes.
There is no direct correlate to our study in the literature
as other studies omit one or more of the studied agents. Our
study was underpowered to detect symptomatic PE, so
symptomatic VTE incidence for each agent relies mainly
on the observed incidence of symptomatic DVT. Aspirin
was found to have a low symptomatic VTE rate, with no
difference when compared to enoxaparin in several meta-
analyses, one retrospective review, and one randomised
controlled trial.11,14,18,21 These findings correspond to the
findings in our study. Comparing aspirin to warfarin, sev-
eral studies found aspirin to be as effective for VTE proph-
ylaxis across patients with varying levels of risk, while
others found aspirin to be inferior.11,35,36 In our study, aspi-
rin was found to have a stronger VTE prophylaxis profile,
specifically for symptomatic DVT, which may be due in
part to increased utilisation, in conjunction with early
mobilisation, and medication compliance. Data directly
comparing aspirin and factor Xa inhibitors is limited. A ret-
rospective study demonstrates no difference between aspi-
rin and other anticoagulants for VTE prophylaxis, but this
study group includes modern anticoagulants other than
only factor Xa inhibitors, so the results are not reflective of
a direct comparison.12 We found that factor Xa inhibitors
had a similar symptomatic VTE incidence when compared
to aspirin. Factor Xa inhibitors are more effective in pre-
venting symptomatic VTE compared to enoxaparin as
found in several recent meta-analyses of randomised con-
trolled trials.7,20 These findings correlate with ours at cor-
responding time points and may be attributable to both
mechanism of action as well as ease of administration and
compliance in comparison to enoxaparin.37,38 Data directly
comparing factor Xa inhibitors to warfarin is limited,
though a randomised clinical trial found no differences
between the groups.39 In our study, factor Xa inhibitors had
a lower symptomatic DVT incidence than warfarin. A
recent study examining patients prescribed warfarin for
thromboprophylaxis found difficulty in maintaining target
therapeutic INR range postoperatively, specifically that
INR values outside of targeted range were often subthera-
peutic, providing a possible explanation for this discrep-
ancy.40 Our study aligns with much of the literature, while
providing a direct comparison of all 4 agents with the same
methodology.
Bleeding risk is often cited as a disadvantage to the use
of factor Xa inhibitors, which have a higher bleeding risk
when compared to enoxaparin.17,41,42 With our study que-
ries and definitions, we found no difference in bleeding-
related complications and a decreased transfusion rate
compared to traditional anticoagulants. A possible expla-
nation may lie in our grouping of factor Xa inhibitors.
Multiple meta-analyses of high-quality trials have demon-
strated an increased risk of bleeding when comparing fac-
tor Xa inhibitors to enoxaparin, contrasting our study,
however the dose or the specific factor Xa inhibitor used
invariably played a role.7,20,43 In these studies, rivaroxa-
ban had increased bleeding risk compared to enoxaparin,
while apixaban had no difference in bleeding risk or even
reduced risk. Low dose factor Xa inhibitors also had a
reduced bleeding risk. It is possible that by grouping all
factor Xa inhibitors, the lower dose, lower risk factor Xa
inhibitors may have masked the effects of the higher risk
factor Xa inhibitors on bleeding outcomes. Aspirin had
the lowest bleeding risk overall in our study, but parallel
relationships in current literature are limited. The litera-
ture on aspirin versus enoxaparin is mixed, with aspirin
having lower or equivalent rates of bleeding complica-
tions.21,44 The literature also remains mixed on bleeding
outcomes for aspirin compared to warfarin, though our
study demonstrated lower postoperative anaemia and
transfusion rates with aspirin.21,35
Choice of thromboprophylaxis agent after primary
THA remains an important issue, and an ideal agent has
yet to be identified. We found that either aspirin or a factor
Xa inhibitor is an effective thromboprophylactic agent
with a favourable complication profile. Enoxaparin,
despite higher rates of bleeding complications, may still
benefit patients who are appropriately selected. It is
Bala et al. 7
important that research on the safety and efficacy of these
agents continues as prescribing patterns and utilisation
rates continue to change. We believe that our study will
serve as a springboard for future studies, including appro-
priate patient selection for specific VTE prophylaxis
agents, and comparison of low-bleeding risk Xa inhibitors
such as apixaban and aspirin.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
article.
Funding
The author(s) received no financial support for the research,
authorship and/or publication of this article.
References
1. Falck-Ytter Y, Francis CW, Johanson NA, et al. Prevention
of VTE in orthopedic surgery patients: antithrombotic ther-
apy and prevention of thrombosis, 9th ed: American College
of Chest Physicians Evidence-Based Clinical Practice
Guidelines. Chest 2012; 141(Suppl.): e278S–e325S.
2. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of
venous thromboembolism: American College of Chest
Physicians Evidence-Based Clinical Practice Guidelines
(8th Edition). Chest 2008; 133(Suppl.): 381S–453S.
3. Anderson FA Jr, Huang W, Friedman RJ, et al. Prevention
of venous thromboembolism after hip or knee arthroplasty:
findings from a 2008 survey of US orthopedic surgeons. J
Arthroplasty 2012; 27: 659–666.
4. Farfan M, Bautista M, Bonilla G, et al. Worldwide
adherence to ACCP guidelines for thromboprophylaxis
after major orthopedic surgery: a systematic review of
the literature and meta-analysis. Thromb Res 2016; 141:
163–170.
5. Knesek D, Peterson TC and Markel DC. Thromboembolic
prophylaxis in total joint arthroplasty. Thrombosis 2012;
2012: 837896.
6. Lewis CG, Inneh IA, Schutzer SF, et al. Evaluation of the
first-generation AAOS clinical guidelines on the prophy-
laxis of venous thromboembolic events in patients undergo-
ing total joint arthroplasty: experience with 3289 patients
from a single institution. J Bone Joint Surg Am 2014; 96:
1327–1732.
7. Venker BT, Ganti BR, Lin H, et al. Safety and efficacy of
new anticoagulants for the prevention of venous thrombo-
embolism after hip and knee arthroplasty: a meta-analysis. J
Arthroplasty 2016; 32: 645–652.
8. Lieberman JR and Pensak MJ. Prevention of venous throm-
boembolic disease after total hip and knee arthroplasty. J
Bone Joint Surg Am 2013; 95: 1801–1811.
9. Pierce TP, Elmallah RK, Jauregui JJ, et al. What‘s new in
venous thromboembolic prophylaxis following total knee
and total hip arthroplasty? An Update. Surg Technol Int
2015; 26: 234–237.
10. Anderson DR, Dunbar MJ, Bohm ER, et al. Aspirin
versus low-molecular-weight heparin for extended venous
thromboembolism prophylaxis after total hip arthroplasty:
a randomized trial. Ann Intern Med 2013; 158: 800–806.
11. Bayley E, Brown S, Bhamber NS, et al. Fatal pulmonary
embolism following elective total hip arthroplasty: a 12-year
study. Bone Joint J 2016; 98-B: 585–588.
12. Chu JN, Maselli J, Auerbach AD, et al. The risk of venous
thromboembolism with aspirin compared to anticoagulants
after hip and knee arthroplasty. Thromb Res 2017; 155: 65–71.
13. Eriksson BI, Borris LC, Friedman RJ, et al. Rivaroxaban
versus enoxaparin for thromboprophylaxis after hip arthro-
plasty. N Engl J Med 2008; 358: 2765–2775.
14. Feng W, Wu K, Liu Z, et al. Oral direct factor Xa inhibitor
versus enoxaparin for thromboprophylaxis after hip or knee
arthroplasty: systemic review, traditional meta-analysis,
dose-response meta-analysis and network meta-analysis.
Thromb Res 2015; 136: 1133–1144.
15. Kakkar AK, Brenner B, Dahl OE, et al. Extended duration
rivaroxaban versus short-term enoxaparin for the prevention
of venous thromboembolism after total hip arthroplasty: a
double-blind, randomised controlled trial. Lancet 2008;
372: 31–39.
16. Lassen MR, Gallus A, Raskob GE, et al. Apixaban versus
enoxaparin for thromboprophylaxis after hip replacement. N
Engl J Med 2010; 363: 2487–2498.
17. Ning GZ, Kan SL, Chen LX, et al. Rivaroxaban for
thromboprophylaxis after total hip or knee arthroplasty: a
meta-analysis with trial sequential analysis of randomized
controlled trials. Sci Rep 2016; 6: 23726.
18. An VV, Phan K, Levy YD, et al. Aspirin as thromboprophy-
laxis in hip and knee arthroplasty: a systematic review and
meta-analysis. J Arthroplasty 2016; 31: 2608–2616.
19. As-Sultany M, Pagkalos J, Yeganeh S, et al. Use of oral
direct factor Xa inhibiting anticoagulants in elective hip and
knee arthroplasty: a meta-analysis of efficacy and safety
profiles compared with those of low-molecular-weight hep-
arins. Curr Vasc Pharmacol 2013; 11: 366–375.
20. Boyd RA, DiCarlo L and Mandema JW. Direct oral anti-
coagulants vs.enoxaparin for prevention of venous throm-
boembolism following orthopedic surgery: a dose-response
meta-analysis. Clin Transl Sci 2017; 10: 260–270.
21. Wilson DG, Poole WE, Chauhan SK, et al. Systematic
review of aspirin for thromboprophylaxis in modern elec-
tive total hip and knee arthroplasty. Bone Joint J 2016; 98-
B: 1056–1061.
22. NIH U.S. National Library of Medicine. ClinicalTrials.
gov. Comparative effectiveness of pulmonary embolism
prevention after hip and knee replacement (PEPPER).
https://clinicaltrials.gov/ct2/show/NCT02810704 (2016,
accessed 6 June 2018).
23. Healy WL, Iorio R, Clair AJ, et al. Complications of total
hip arthroplasty: standardized list, definitions, and stratifi-
cation developed by the hip society. Clin Orthop Relat Res
2016; 474: 357–364.
24. Elixhauser A, Steiner C, Harris DR, et al. Comorbidity
measures for use with administrative data. Med Care 1998;
36: 8–27.
25. Gordon M, Stark A, Sköldenberg OG, et al. The influence of
comorbidity scores on re-operations following primary total
hip replacement: comparison and validation of three comor-
bidity measures. Bone Joint J 2013; 95-B: 1184–1191.
8 HIP International 00(0)
26. Rosenman M, Liu X, Phatak H, et al. pharmacologic proph-
ylaxis for venous thromboembolism among patients with
total joint replacement: an electronic medical records study.
Am J Ther 2016; 23: e336–e344.
27. Shah SS, Satin AM, Mullen JR, et al. Impact of recent
guideline changes on aspirin prescribing after knee arthro-
plasty. J Orthop Surg Res 2016; 11: 123.
28. Lassen MR, Ageno W, Borris LC, et al. Rivaroxaban versus
enoxaparin for thromboprophylaxis after total knee arthro-
plasty. N Engl J Med 2008; 358: 2776–2786.
29. Bozic KJ, Bashyal RK, Anthony SG, et al. Is administra-
tively coded comorbidity and complication data in total
joint arthroplasty valid? Clin Orthop Relat Res 2013; 471:
201–205.
30. Clair AJ, Inneh IA, Iorio R, et al. Can administrative data be
used to analyze complications following total joint arthro-
plasty? J Arthroplasty 2015; 30(Suppl.): 17–20.
31. Lewis S, Kink S, Rahl M, et al. Aspirin: are patients actually
taking it? A quality assessment study. Arthroplast Today
2018; 4: 475–478.
32. Hood BR, Cowen ME, Zheng HT, et al. Association of aspi-
rin with prevention of venous thromboembolism in patients
after total knee arthroplasty compared with other antico-
agulants. A noninferiority analysis. JAMA Surg 2019; 154:
65–72.
33. Shahi A, Bradbury TL, Guild GN 3rd, et al. What are the
incidence and risk factors of in-hospital mortality after
venous thromboembolism events in total hip and knee
arthroplasty patients? Arthroplast Today 2018; 4: 343–347.
34. Bedard NA, Pugely AJ, Lux NR, et al. Recent trends in
blood utilization after primary hip and knee arthroplasty. J
Arthroplasty 2017; 32: 724–727.
35. Huang RC, Parvizi J, Hozack WJ, et al. Aspirin is as effec-
tive as and safer than warfarin for patients at higher risk
of venous thromboembolism undergoing total joint arthro-
plasty. J Arthroplasty 2016; 31(Suppl.): 83–86.
36.
Intermountain Joint Replacement Center Writing
Committee. A prospective comparison of warfarin to aspirin
Appendix.  List of Codes.
Procedure or Complication
THA
DVT
PE
Anemia (postop)
Bleeding Complications
Transfusion Rate
THA, Total Hip Arthroplasty; DVT, Deep Venous Thrombosis; PE, Pulmonary Embolism.
for thromboprophylaxis in total hip and total knee arthro-
plasty. J Arthroplasty 2012; 27: 1–9.e2.
37. Carrothers AD, Rodriguez-Elizalde SR, Rogers BA, et al.
Patient-reported compliance with thromboprophylaxis
using an oral factor Xa inhibitor (rivaroxaban) following
total hip and total knee arthroplasty. J Arthroplasty 2014;
29: 1463–1467.
38. Wilke T, Moock J, Müller S, et al. Nonadherence in out-
patient thrombosis prophylaxis with low molecular weight
heparins after major orthopaedic surgery. Clin Orthop Relat
Res 2010; 468: 2437–2453.
39. Bern MM, Hazel D, Deeran E, et al. Low dose compared to
variable dose Warfarin and to Fondaparinux as prophylaxis for
thromboembolism after elective hip or knee replacement sur-
gery; a randomized, prospective study. Thromb J 2015; 13: 32.
40. Nam D, Sadhu A, Hirsh J, et al. The use of warfarin for DVT
prophylaxis following hip and knee arthroplasty: how often
are patients within their target INR range? J Arthroplasty
2015; 30: 315–319.
41. Adam SS, McDuffie JR, Lachiewicz PF, et al. Comparative
effectiveness of new oral anticoagulants and standard throm-
boprophylaxis in patients having total hip or knee replacement:
a systematic review. Ann Intern Med 2013; 159: 275–284.
42. Gomez-Outes A, Terleira-Fernandez AI, Suarez-Gea
ML, et al. Dabigatran, rivaroxaban, or apixaban versus
enoxaparin for thromboprophylaxis after total hip or knee
replacement: systematic review, meta-analysis, and indirect
treatment comparisons. BMJ 2012; 344: e3675.
43. Neumann I, Rada G, Claro J, et al. Oral direct factor xa
inhibitors versus low-molecular-weight heparin to prevent
venous thromboembolism in patients undergoing total hip
or knee replacement: a systematic review and meta-analy-
sis. Ann Intern Med 2012; 156: 710–719.
44. Jameson SS, Baker PN, Charman SC, et al. The effect of
aspirin and low-molecular-weight heparin on venous throm-
boembolism after knee replacement: a non-randomised
comparison using National Joint Registry Data. J Bone Joint
Surg Br 2012; 94: 914–918.
Codes
CPT: 27130, ICD-9: 81.51
451.11,451.19,451.2,451.81,451.9,453.40,453.41,453.42,453.8,453.9
415.1,415.11,415.13,415.19
285.1
998.1,998.11,998.12,998.13,719.10,719.16,719.16,39.98
CPT: 36430,ICD-9: 99.0, 99.00,99.02,99.03,99.04