Inferior Vena Cava Filter Use and Risk of DVT/PE Post Arthroplasty
Daniel R Mesko DO1*, Nikkole Haines MD1, Alex Sadauskas BS2 and Scott M Sporer MD1
1Department of Orthopaedic Surgery, RUSH University Medical Center, USA
2Department of Orthopaedics, Saint Louis University School of Medicine, USA
Submission: June 15, 2018; Published: July 10, 2018
*Corresponding author: Daniel R Mesko, Department of Orthopaedic Surgery, RUSH University Medical Center, 1161 W Harrison St Chicago, IL 60612, Tel: (877) 632 6637; Email: email@example.com
How to cite this article: Mesko DR, Haines N, Sadauskas A, Sporer SM. Inferior Vena Cava Filter Use and Risk of DVT/PE Post Arthroplasty. Ortho &
Rheum Open Access J 2018; 12(3): 555836. DOI: 10.19080/OROAJ.2018.12.555836.
Background: Despite anticoagulation use, total joint arthroplasty has a 0.16-1.51% pulmonary embolism (PE) rate and a 0.1-0.2% fatal PE rate. The purpose of this study was to describe the post-operative thrombotic events and complications in patients at high-risk in two groups, those who had an IVC filter placed prior to total joint replacement surgery and those who did not.
Methods: Patients from a single surgeon’s practice were reviewed between January 1, 2012 to December 31, 2015 that were identified as high-risk for post-operative thrombotic events. Patient demographics, medical history, Charlson Comorbidity Index, filter-related complications, as well as post-operative deep vein thrombosis, pulmonary embolism, bleeding events and hospitalizations were collected retrospectively from the electronic medical records.
Results: Fifty-five patients were identified, thirty-six of whom had a filter placed. Two of these had filter placement and removal on separate occasions for separate arthroplasty surgeries. Of the IVC filters placed, two insertional complications were identified. No filter retrieval complications occurred in the filters that were removed. There were nineteen patients deemed high risk for postoperative thrombotic events that did not undergo filter placement. Of these, there were no cases of pulmonary embolism. One patient was found to have a chronic lower-extremity DVT on post-operative work-up.
Conclusion: In our series, there were two complications related to filter placement. The group of nineteen high risk patients without a filter placed showed no cases of symptomatic PE. Further multicenter study is required to make definitive recommendations on the use of IVC filters in select arthroplasty patients.
The use of regional anesthesia, expeditious surgical times, early ambulation, sequential compression devices and anticoagulants have all helped reduce the rate of symptomatic pulmonary embolism (PE) and deep vein thrombosis (DVT) . Despite these improvements, symptomatic PE, including fatal PE, still occurs after joint arthroplasty. PE’s following arthroplasty surgery without prophylaxis occur at a rate between 4-10% with a fatality rate of 0.2-5% [2,3]. Fatal PE’s following total joint arthroplasty (TJA) with more modern techniques and prophylaxis have been found to occur from 0.08-0.2% [2,3]. While thrombotic events remain relatively uncommon, a projected increase of both primary and revision hip and knee arthroplasty cases to over 4.4 million annually by 2030 represents a potential for over 8,800 PE-related deaths in the same time frame . Utilization of more aggressive anticoagulation has been looked at in depth to manage patients potentially at higher risk for such events; however, this has not demonstrated any reduction in the incidence of fatal PE after total joint arthroplasty .
In addition to not reducing fatal PE, potent anticoagulation increases the risks of wound complications, infections, and all-cause mortality compared to less aggressive anticoagulation regimens . Furthermore, many fatal PE’s occur prior to anticoagulation initiation postoperatively . A prophylactic IVC filter is a supplemental and/or alternative method for prevention of PE that has been investigated more in the bariatric, spine and trauma literature than the arthroplasty literature [8–18]. The use of prophylactic IVC filters in high-risk arthroplasty patients prior to surgery has not been widely utilized. According to current AAOS guidelines, there is a lack of compelling evidence for or against filter use in patients who have a contraindication to anticoagulation or known residual venous thromboembolic disease .
CHEST guidelines recommend against prophylactic IVC filters,
even in patients with contraindications to pharmacologic or
mechanical prophylaxis, due to low quality evidence supporting
their benefit and due to cases of complications during insertion
and retrieval . The purpose of this study was to look at the
use of prophylactic IVC filters in patients undergoing total joint
replacement surgery and to evaluate complication rates related
to filter placement/removal as well as post-thrombotic events in
these high-risk patients. Our hypothesis is that IVC filters, in high
risk patients, show equal or reduced rates of fatal PE compared
to a similar group of high risk patients from the same surgeon as
well as the current literature rates, with minimal complications
during filter placement / retrieval.
After institutional review board study approval was
obtained, our digital database was used to identify patients
undergoing primary total joint arthroplasty between January 1,
2012 to December 31, 2015 who were identified by the clearing
medical physician as high-risk for thrombotic event. Of the 55
patients identified, an IVC filter was placed prior to undergoing
total joint arthroplasty in 36 of these patients. For inclusion,
we required chart follow-up for a minimum of 6 weeks postoperatively
or until the patient was deceased. Thirty-six patients
with thirty-eight IVC filter placements were identified (10 total
hips, 21 total knees, 1 simultaneous bilateral total knees and 6
staged bilateral knees). Retrospectively collected data included
patient demographics, medical history, CCI, surgical procedure
details, post-operative course, filter details and related events.
The high-risk cohort who received an IVC filter included
twenty females and eighteen males with a mean age of 70.8
+/- 9.3 years. The average Charlson Comorbidity Index (CCI)
of this patient group was 4.3+/- 2.6. Indications for IVC filter
placement were prior PE (16 patients), both PE & DVT (11
patients), history of DVT with additional risk factors (7 patients),
prior recurrent DVT (3 patients) and a contraindication to
anticoagulation (1 patient). Filter placement took an average
time of 8.9 +/-5 minutes. Patient selection for filter utilization
and type was determined by the medical team, hematologist
and interventional radiologist. In total, 5 Celect filters (Cook,
Bloomington, Indiana), 28 Günther Tulip filters (Cook Medical,
Bloomington, Indiana), 4 VenaTech LP (permanent) filters (B.
Braun Medical Inc, Bethlehem, Pennsylvania), and a single
Denali filter (Bard Medical, Murray Hill, New Jersey) were used.
Average follow-up following total joint arthroplasty was 27.3 +/-
The nineteen high-risk patients who did not have an IVC
filter placed included 6 males and 13 females with a mean age
of 70.2 +/- 7.3 years. The average CCI of this group was 3.7+/-
1.5. Factors placing these patients at high-risk included a history
of PE (2 patients), history of PE and DVT (5 patients), a history
of DVT as well as other risk factors (8 patients), and a history
of DVT alone (4 patients). Descriptive statistics were calculated
with SPSS 24 (IBM, Armonk, NY). There were no statistical
differences between the demographics of the two groups
(Table 1). Due to the small numbers in the outcomes studies, no
comparative analyses could be performed.
*both with BMI >40; ** 1-Factor 2 gene mutation 1-factor V mutation, 1-APL antibody, 1- sarcoidosis, 2-afib & prostate CA, 1-TIA, 1-CVA;
***3-Morbid Obesity, 1-Factor V Leiden, 1-Atrial Fibrillation, 1-Metastatic Cancer & 1-Subacute Femoral DVT; ****1-Gastric Cancer with Bleeding
Two patients had filter-related complications. One patient
presented to the emergency department one day after filter
placement with bleeding from the right internal jugular entry
site. This was managed with application of surgical skin glue
and same-day discharge home. The second patient developed
retroperitoneal hemorrhage one day following filter placement.
This was managed with an eight-night admission and a diagnostic
angiogram, that was negative for an active extravasation. Due to
the hemorrhage, the patient’s scheduled total hip arthroplasty
(THA) was delayed and, in the interim, all anticoagulants were
held. The patient then developed bilateral DVT and PE resulting
in an additional 7-night hospitalization. Both patients eventually
underwent successful arthroplasty surgery without any further
symptomatic thromboembolic events.
Of the filters placed, four were intended as permanent
filters, two patients passed away from unrelated causes prior to
planned removal, two patients refused removal and one patient’s
filter was left in place long term due to a history of intracranial
hemorrhage and subsequent recommendations against systemic
anti-coagulation. Of the twenty-nine remaining filters, 27 (93%)
were successfully removed without complication at an average
of 120.7 +/- 95.8 days and 2 (7%) were lost to follow-up prior
to removal. The average IVC extraction procedure took 12 +/-
6.8 minutes (Table 2). One extraction procedure was aborted
initially as a 30-40% occluding thrombus was discovered within
the filter; this filter was removed three months later.
Four patients (11%) who received an IVC filter died during
the follow-up period. No deaths were related to the filter or
thromboembolic events. Three deaths were cancer-related,
and one was secondary to a hypertensive event. In the group
of 19 high risk patients who did not have a filter placed, all
were managed with warfarin and enoxaparin. Two patients
were worked up for DVT post-op, one of which was found to
have a chronic thrombus and the second was found to have
post-operative swelling not attributed to a thrombotic event.
No documented bleeding events related to patients’ systemic
anticoagulation were discovered. No patients in this cohort died
during the follow-up period.
Prevention of thromboembolic events in patients at highrisk
for DVT/PE or with a contraindication to anticoagulation
presents a significant challenge. Readmission rates among
arthroplasty patients, secondary to a symptomatic DVT/PE,
were found by White et al.  to be 2.4 times higher in patients
with a prior history of thromboembolic event (Table 3). Other
characteristics such as age greater than 70 years old, female
gender, elevated BMI, total knee arthroplasties, increased CCI,
presence of DVT, anemia, COPD and even depression have been
identified as risk factors for pulmonary embolism [21,22].
The question remains that despite some risk inherent to filter
placement, are there specific patients that may benefit from IVC
filter placement prior to total joint arthroplasty?
The AAOS guidelines regarding filter use for high-risk
patients or those who are not able to be anticoagulated is listed as
inconclusive . Even when high risk patients are considered,
CHEST guidelines do not recommend prophylactic IVC filters, as
PE can still occur, and filters carry a 2-6% rate of “substantial
harm” in their review . Filter placement and removal is
not without potential for risk. Issues such as carotid artery
puncture, filter migration/tilt, filter breakage with embolism
to the heart and lung, need for thoracotomy, post thrombotic
syndrome, skin changes and inability to remove the filter due
to wall adhesion have all been reported . Filter placement
also does not eliminate the possibility of PE, as shown by one of
the patients in our study, as any formation and embolization of
thrombus proximal to the filter site or in the upper extremities
will not be prevented.
Few studies have been published looking specifically at IVC
filter placement prophylactically in high risk total joint patients.
Vaughn et al., published on 37 patients who had permanent filters
placed preoperatively to undergoing total joint arthroplasty.
They demonstrated only one case of asymptomatic PE and found
no filter related complications at an average follow up of 18.8
months . Emerson et al.  also showed no PE diagnoses in
a group of 33 patients; however, filter complications, including
two vagus nerve injuries, a hemothorax and a placement error
above the renal vein, occurred. In another cohort study, Austin
et al., looked at filters placed in total joint arthroplasty patients.
However, only 18 of these were preoperatively placed, while the
remaining 109 filters were placed in reaction to confirmed VTE
related events at a mean of 2.7 days post-operatively.
Across the entire cohort, recurrent PE was seen in two
patients; it was not clear if these were in the prophylactic
IVC filter placement patient group or the reactive placement
group. Filter related complications included femoral artery
pseudoaneurysm, filter malposition, recurrent or extending DVT
in five patients, and 15 patients had a persistent painful lower
extremity (post thrombotic syndrome) . As permanent IVC
filters have been associated with an increased risk for DVT,
the interest in removable filters has grown . Strauss et al.
 reported on retrievable IVC filters in 58 patients, including
12 arthroplasty patients. Sixty-four percent of the filters were
retrieved, all without insertional or retrieval complications.
Dhand et al.  have published the largest study to date with
109 retrievable filters placed prior to TJA. Their study group had
two symptomatic PEs despite filter placement, neither of which
were fatal. All but one of the filters were successfully removed.
One major filter related complicated included a fracture of the
filter leg, which embolized and necessitated a thoracotomy for
Our study demonstrated results similar to prior studies
in our filter patients. Overall, two patients (5.3%) had an
adverse event (both were DVT) during the follow-up period.
No patients sustained a PE post arthroplasty. Two patients
(5.3%) had a complication related to filter placement, and four
patients (11%) died of unrelated causes during the follow-up
period. One patient had a major complication, a retroperitoneal
bleed and subsequent DVT and PE, and second patient had
a minor complication. The limitations of our study include
the retrospective nature of the study. It is possible a patient
presented outside of our system and therefore an adverse event
or VTE avoided detection. Complications and adverse events
were identified via chart review, therefore non-symptomatic
PE/DVT may have been over-looked. Our study is also limited
due to the small study numbers over the time investigated.
Though small, it is reflective of the more limited volume of highrisk
patients and judicious use of preoperative IVC filters in
anticipated total joint arthroplasty patients.
In our high-risk cohort, IVC filters placed prophylactically
for total hip and knee arthroplasty showed a relatively low
complication rate from filter placement and retrieval, without any
cases of postoperative PE’s. The control group also demonstrated
zero post-operative PE’s. As the technology and technique of
IVC filter and their placement was evolved, prophylactic use
in populations at high risk for thromboembolic events or with
a contraindication to anticoagulation after joint arthroplasty
should be further investigated through a multicenter study to
best identify who may be best served by this technology.