1

A PILOT STUDY ON THE RANDOMIZATION OF INFERIOR VENA CAVA FILTER PLACEMENT FOR VENOUS THROMBOEMBOLISM PROPHYLAXIS IN HIGH-RISK

TRAUMA PATIENTS

By

ANITA RAJASEKHAR

A THESIS PRESENTED TO THE GRADUATE SCHOOL

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2012

2

© 2012 Anita Rajasekhar

3

To Raya, Anjali, and Tyson

4

ACKNOWLEDGMENTS

This work supported in part by the NIH/NCATS Clinical and Translational

Science Award to the University of Florida UL1TR000064. This study was supported by

a Cook medical education grant in the form of 50 donated inferior vena cava filters. We

thank Huazhi Liu for performing the statistical analysis for this study. We thank Tera

Thigpin and Jennifer Lanz, the research coordinators for this study. We thank the

University of Florida, Department of Surgery, Division of Acute Care Surgery faculty for

recruiting patients for this study.

5

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES ............................................................................................................ 6

LIST OF FIGURES .......................................................................................................... 7

ABSTRACT ..................................................................................................................... 8

CHAPTER

1 INTRODUCTION .................................................................................................... 10

Venous Thromboembolism in Trauma Patients ...................................................... 10 Pilot Feasibility Study .............................................................................................. 14

Study Objectives ..................................................................................................... 15

2 METHODS .............................................................................................................. 16

Patient Population ................................................................................................... 16 Study Design .......................................................................................................... 17 Screening and Follow-up ........................................................................................ 18

Study Outcomes ..................................................................................................... 19 Statistical Analysis .................................................................................................. 19

3 RESULTS ............................................................................................................... 21

Primary Outcomes .................................................................................................. 21 Secondary Outcomes ............................................................................................. 22

4 DISCUSSION ......................................................................................................... 26

LIST OF REFERENCES ............................................................................................... 32

BIOGRAPHICAL SKETCH ............................................................................................ 36

6

LIST OF TABLES

Table page 3-1 Sociodemographic and clinical variables according to treatment group ............. 23

3-2 Primary outcomes of feasibility ........................................................................... 23

3-3 Secondary clinical outcomes .............................................................................. 24

3-4 Pharmacologic prophylaxis ................................................................................. 24

3-5 Follow-up compliance ......................................................................................... 24

7

LIST OF FIGURES

Figure page 3-1 CONSORT flow diagram .................................................................................... 25

8

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science

A PILOT STUDY ON THE RANDOMIZATION OF INFERIOR VENA CAVA FILTER

PLACEMENT FOR VENOUS THROMBOEMBOLISM PROPHYLAXIS IN HIGH-RISK TRAUMA PATIENTS

By

Anita Rajasekhar

December 2012 Chair: Marian Limacher Major: Medical Sciences – Clinical and Translational Science

Prophylactic inferior vena cava filters (pIVCFs) for the prevention of pulmonary

embolism (PE) in high-risk trauma patients (HRTPs) are frequently inserted despite the

lack of supportive data demonstrating efficacy. This report details the two-year interim

analysis of the Filters in Trauma (FIT) Pilot Study, a single institution, prospective

randomized pilot feasibility study in a Level 1 trauma center. HRTPs were identified for

pIVCF placement by the Eastern Association for the Surgery of Trauma (EAST)

guidelines. From November 2008 to November 2010, 38 HRTPs were enrolled and

randomized to either pIVCF versus none. All patients received pharmacologic venous

thromboembolism (VTE) prophylaxis when safe. Primary outcomes included a priori

defined feasibility objectives and secondary outcomes were incidence of PE, deep vein

thrombosis (DVT), and death.

Thirty of 38 enrolled patients were eligible for analysis. The baseline socio-

demographic characteristics were balanced between both groups. Primary feasibility

results included: time from admission to enrollment (mean 45.7 ± 21.5hrs), time from

enrollment to randomization (mean 4.4 ± 8.5hrs), time from randomization to IVCF

9

placement (mean 17.2 ± 9.1hrs), adherence to weekly compression ultrasound within

first month (IVCF group = 46.7%, non-IVCF group = 66.7%), and one month follow-up

(IVCF group=80%, non-IVCF group=100%). At 6-month follow-up one PE occurred in

the non-filter group. DVT was not diagnosed in either group. One non-VTE related

death occurred in the filter group. Barriers to enrollment included inability to obtain

informed consent due to patient refusal and no next of kin identified. Our pilot study

demonstrates for the first time that a randomized controlled trial evaluating the efficacy

of pIVCFs in trauma patients is feasible.

10

CHAPTER 1 INTRODUCTION

Venous Thromboembolism in Trauma Patients

Venous thromboembolism (VTE), including deep vein thrombosis (DVT) and

pulmonary embolism (PE), is an important health care problem resulting in significant

morbidity, mortality, and health care expenditures. Clinicians attribute VTE to the

abnormalities described as Virchow’s triad: venous stasis, intimal injury, and a

hypercoagulable state. All three conditions commonly occur in trauma patients.

Hospitalized patients recovering from major trauma are among those at highest risk of

developing VTE. In fact, PE is the third most frequent cause of in-hospital death in

trauma patients who survive more than 24 hours after injury.1 The reported incidence of

VTE in trauma patients varies between studies, presumably as a result of heterogeneity

in the risk profile such as the nature of their injuries, the method of diagnosis, and

whether routine surveillance was performed. If VTE prophylaxis is not used, the

incidence of DVT may exceed 50%, with half of patients demonstrating no symptoms.2

PE may occur in as many as 25%, with 2% of patients symptomatic. Among trauma

patients with PE, mortality can be as high as 20 to 50%.2-3 Furthermore, post-thrombotic

syndrome, a chronic condition associated with chronic leg pain, intractable edema, and

leg ulcers, is a burdensome and potentially devastating complication of DVT. Also,

patients with a first episode of VTE are at increased risk for recurrent VTE. Therefore,

VTE results in significant morbidity and mortality, increasing length of hospital stay and

thus cost of care.

To date, Geerts et al. have performed the most comprehensive study on the

incidence of VTE in trauma patients.2 In this study, serial impedance plethysmography

11

and lower-extremity contrast venography were performed to detect DVT in a cohort of

716 high-risk patients admitted to a regional trauma unit. Patients did not receive

mechanical or pharmacologic VTE prophylaxis. Lower extremity DVT was found in 58%

of patients. DVT was proximal in 18% of patients. Fewer than 2% of patients diagnosed

with DVT had symptoms. Two percent of the study population developed PE, 20% of

which were fatal despite clinical and venography surveillance. In a subgroup analysis,

DVT was diagnosed in 50% of patients with major injuries to the face, chest, and

abdomen, 62% with spinal injuries, 54% with head injuries, and 69% with lower

extremity injuries.2

Effective prophylaxis is essential as the clinical diagnosis of VTE is often difficult

and treatment is frequently delayed or inadequate. Several studies have attempted to

identify risk factors that place the trauma patient at high-risk for VTE. The largest study

to date was reported by Knudson et al.3 Using the American College of Surgeons

National Trauma Data Bank, 1,602 patients with VTE were evaluated. Nine risk factors

were found to be significantly associated with VTE: age, spinal cord injury (SCI), head

injury, days on ventilator, venous injury, shock on admission, major surgical procedure,

and pelvic and lower extremity fracture.3 Geerts reported five independent risk factors

for DVT on multivariate analysis including age, blood transfusion, surgery, spinal cord

injury, and fracture of femur or tibia.2

While VTE risk can be decreased with pharmacologic prophylaxis2, a subset of

trauma patients do not qualify for anticoagulation as a result of their injury and are

therefore at great risk for VTE. To address this issue, mechanical methods to prevent

VTE in trauma patients have been explored. An inferior vena cava filter (IVCF) is an

12

endovascular device placed to prevent PE. Accepted indications for insertion of IVCFs

are: 1) known VTE when anticoagulation is contraindicated, 2) recurrent pulmonary

embolism (PE) despite adequate anticoagulation, 3) known complications to

anticoagulant therapy.1 Despite the paucity of reliable data on efficacy and safety,

indications for IVCFs have liberalized. The PREPIC (Prévention du Risque d'Embolie

Pulmonaire par Interruption Cave) study enrolled 400 non-trauma patients with proximal

DVT with or without PE who were randomly assigned to receive an IVCF or no filter. All

patients received therapeutic anticoagulation for known VTE. Early findings revealed

that IVCFs reduced the occurrence of symptomatic and asymptomatic PE (IVCF group

1.1% vs. no filter group 4.8%; p = 0.03). However, after eight years this initial benefit

was counterbalanced by a significant increase in DVT (IVCF group 36% vs. no filter

group 28%; p = .042) with no effect on long-term mortality.4,5 These data indicate that

filters, when used in conjunction with anticoagulation, offer a short-term reduction in the

total number of PEs at the cost of a long-term increase in recurrent DVT with no

reduction in mortality. Unfortunately, because 94% of patients received anticoagulation

for at least 3 months, these data offer no insight into the outcome of the typical patient

who has had a vena caval filter due to a contraindication to anticoagulant therapy.

Since the FDA approval of the Gunther Tulip™ Vena Cava Filter for retrieval in

2003, six new retrievable filters have been approved for use in the United States and

filter use has increased dramatically.6 In 2003, the FDA approved changes to three

existing permanent filters to allow percutaneous retrieval, but did not specify retrieval

indications.7 The rationale for using retrievable IVCFs is to offer mechanical protection

against PE during the limited high-risk period when anticoagulation may be

13

contraindicated. The largest proportional increase in the use of retrievable IVCFs in the

United States has been in patients at risk for PE but who have neither PE nor DVT (i.e.

prophylactic IVCFs).2 Overall, IVCF insertion has increased significantly during the past

10 years in trauma patients.8,9 Over the last decade IVCF placement increased by

111% in the Medicare population.10 Data from the National Hospital Discharge Survey

indicate that 49,000 hospitalized patients received an IVCF in 1999.6 Based on the most

recent data from this registry, 92,000 patients received IVCFs in 2006, representing an

almost 200% increase.6 In 2012, vena cava filter sales are expected to exceed 250,000

units of which 80% (200,000) are expected to be retrievable filters.11 The considerable

variation in use of IVCFs cannot be explained by patient or center characteristics.12

Acute complications of filter placement include improper positioning, tilting, and

insertion site thrombosis. Long term complications are DVT, inferior vena cava (IVC)

obstruction, filter migration, strut fracture, filter erosion through the IVC, and proximal

PE. Timely removal may prevent most of these complications, which provides the

rationale for the increasing popularity for retrievable filters. More than half of IVCFs are

placed for temporary prophylaxis and therefore should be removed.13-16 However,

retrospective studies report that the rate of removal of these retrievable filters is only

approximately 20%.17-20 The overall retrieval rate is a product of the clinical and

procedural rate.21 If attempted, ~ 90% of IVCF retrievals are successful.21 Procedural

factors associated with retrieval failure include prolonged dwelling time, advanced age,

filter head position, and filter design.22-24 The true time frame for successful retrieval of

IVCFs remains undefined, but proceduralists have reported retrieval as long as 494

days after insertion.25 Clinical factors that influence whether or not IVCF retrieval is

14

attempted include comorbidities, concurrent anticoagulation, primary indication for

placement, and documented plans for removal at the time of insertion or when the

patient is discharged.19,26 On August 8, 2010 the FDA issued a safety alert encouraging

timely removal of IVCFs in order to avoid long-term complications.27 Therefore,

improving the rate of IVCF retrieval is an important objective and strategies to increase

retrieval rates are needed.

Pilot Feasibility Study

Despite conflicting data on their efficacy, pIVCFs have become part of the

treatment algorithm at many trauma centers for prevention of VTE in trauma patients at

highest risk. However, the use of pIVCF in any patient population remains controversial.

No randomized controlled trials have addressed the efficacy of pIVCFs in trauma

patients. Practice management guidelines have been published for trauma patients. In

2002, the Eastern Association for the Surgery of Trauma (EAST) cited Class III

evidence (retrospective data, expert opinion, or case report) in support of the use of

pIVCFs in high-risk trauma patients (HRTPs)28. However, the 2012 American College of

Chest Physicians (ACCP) guidelines recommended against the use of pIVCFs as

primary prophylaxis in trauma patients based on nearly the same body of evidence

(GRADE 2C-weak recommendation based on low strength of evidence).29 This

dichotomy in clinical recommendations may reflect differences in practice patterns

across trauma centers in North America. Therefore we designed this pilot study to

determine the feasibility of performing a prospective randomized controlled trial of

pIVCF use among high-risk trauma patients.

The goal of a pilot feasibility study is to identify issues related to study

conceptualization, design, sample size and patient selection, data collection

15

procedures, and approaches for data analysis.30 This type of clinical investigation is

essential for maximizing the success of a future prospective trial. Pilot studies with

explicit feasibility objectives and a priori success criteria are important foundation steps

to prepare for a large trial, ensuring a rigorous trial design that is implemented efficiently

and safely.31

The Filters in Trauma (FIT) Pilot Study was designed to optimally prepare for a

larger multicenter randomized controlled trial evaluating the efficacy and safety of

pIVCFs in trauma patients. Herein the outcomes of this pilot feasibility study are

represented. In addition, the logistics of the trial design and implementation barriers that

were identified over the first two years that the study was open for enrollment are

delineated.

Study Objectives

The primary outcome measures of this pilot study were the following feasibility

objectives: 1) timely enrollment of high-risk trauma patients, 2) acceptable time to

randomization, 3) timely receipt of allocated treatment, 4) adherence to weekly lower-

extremity surveillance compression ultrasounds (CUS). Secondary objectives were to

obtain preliminary data on VTE event rates at hospital discharge, 1 month, and 6

months post-discharge. Additionally, barriers to recruitment and randomization, use of

pharmacologic prophylaxis, prescription of sequential compression devices (SCDs),

compliance with follow-up, and removal rates of filters were examined.

16

CHAPTER 2 METHODS

Patient Population

This was a single institution prospective randomized pilot study to evaluate the

feasibility of investigating the use of pIVCFs in HRTPs at the University of Florida in

Gainesville, Florida. The study opened for enrollment in November 2008. All patients

were recruited from Shands Hospital at the University of Florida, a Level 1 trauma

center serving 13 counties in North Central Florida with a population base of 1.1 million

citizens. Prior to this study the standard of practice at the University of Florida for VTE

primary prophylaxis included placement of pIVCF in any trauma patient deemed high-

risk for VTE. In addition, immediate pharmacologic prophylaxis was used when deemed

safe by the primary service, usually within 24 hours.

The inclusion criteria for this study were based on the EAST guidelines for high-

risk trauma patients.28 Patients considered high-risk for VTE and therefore included in

the study were those over the age 18 admitted to the trauma service within 96 hours

before the enrollment, and with at least one of the following injury patterns: 1) spinal

cord injury with paralysis; 2) multiple complex pelvic fractures; 3) bilateral lower-

extremity (LE) long bone fractures, excluding fibula; 4) pelvic fracture plus one or more

LE long bone fracture(s) excluding fibula; 5) expected non-weight bearing bilaterally > 7

days; 6) body mass index (BMI) > 35kg/m2. All patients were required to have expected

hospitalization for > 1 week to ensure at least two CUS of the LE were performed prior

to hospital discharge. In addition, to be eligible, patients must have had a negative

baseline lower-extremity CUS. Patients were excluded if they had indications for

therapeutic anticoagulation, had a previous IVCF or contraindications for IVCF

17

placement, or were deemed to be ineligible for pharmacologic prophylaxis throughout

their entire hospital stay. Additional exclusion criteria included inability to obtain

informed consent from the patient or proxy; pregnant, institutionalized, or terminally ill

patients; or anticipated survival < 24hours.

The University of Florida Institutional Review Board (IRB) approved this study.

An independent data safety monitoring board (DSMB) comprising a hematologist, a

cardiologist, and a vascular surgeon at our institution was convened every four months

to investigate potential adverse events and safety issues. Pre-specified stopping rules

included a 10% absolute difference of fatal PEs or deaths between the two treatment

groups. This absolute difference was chosen based on heterogeneous population

studies in which a total event rate of fatal PEs or death were reported to be less than

10%.

Study Design

Physicians on the admitting trauma service were the first points of contact to

introduce the study to eligible patients. After immediate stabilization and initial surgical

procedures, the principal investigator or study coordinator, who were not affiliated with

the direct care of the patient, obtained written informed consent from the patient or

legally authorized representative. Once enrolled in the study, patients underwent

computer-generated simple randomization to receive pIVCF placement or no pIVCF.

For patients allocated to pIVCF the vascular surgeon or trauma surgeon associated with

the study was notified and performed placement of the IVCF. All patients in the pIVCF

arm received a market-approved CookTM CelectTM retrievable IVCF donated by Cook

Medical (Bloomington, Indiana, USA). Filters were inserted at the bedside using

18

intravascular ultrasound guidance via the femoral route and advanced to an infrarenal

position.

When considered safe by the primary service, all patients received

pharmacologic VTE prophylaxis in the form of either enoxaparin (Sanofi-Aventis, Paris,

France) 30mg subcutaneous (SQ) twice daily, unfractionated heparin (UFH) 5000 units

SQ three times daily (for spinal cord injury), or fondaparinux (GlaxoSmithKline, London,

UK) 2.5mg SQ daily (if clinically indicated, e.g., suspicion of heparin induced

thrombocytopenia) at the discretion of the primary clinical service. Given the nature of

device placement, the patient, treating physician and nurses, ultrasound technologists,

and research personnel were not masked to the treatment allocation. Patients in both

treatment groups received all other standards of care for their injuries. Sequential

compression devices were employed when the site of injury did not preclude them.

Screening and Follow-up

Standard of clinical practice included weekly lower extremity surveillance

compression ultrasound (CUS) to monitor for asymptomatic DVT.32,33 CUS of the lower

extremities were also performed for evaluation of clinically suspected DVT For

symptoms or signs suggestive of PE, a spiral computed tomography of the chest using

a standard PE protocol was performed. On CUS, an inability to compress the vein or

actual visualization of a thrombus was characterized as a DVT. On computed

tomography of the chest, a filling defect noted in a main, segmental, or subsegmental

pulmonary artery was defined as a PE. Patients who developed a venous

thromboembolism (PE or DVT) were treated with therapeutic anticoagulation, if not

contraindicated, based on practitioner’s choice.

19

All patients were followed prospectively until death or hospital discharge and then

as an outpatient in 1 month (in follow-up clinic or via phone call if no trauma clinic

appointment was needed per standard of practice) and at 6 months (via phone call)

post-discharge. Following discharge IVCF removal was attempted as per the following

current vascular surgery practices. Patients were scheduled for a vascular surgery

clinic appointment 3 months after filter placement. A letter was sent from the vascular

surgery office to remind patients of this appointment. Upon return to vascular clinic

patients underwent duplex ultrasound for IVCF retrieval planning and clinical evaluation

to determine ambulatory status and ongoing VTE risks. The decision to offer removal

depended on whether the VTE risk had returned to baseline. At 6 months post-

discharge patients were asked about complications of the filter, diagnosis of any new

VTE, continued pharmacologic prophylaxis, and whether the filter had been removed.

Study Outcomes

The primary outcomes were the feasibility objectives described above. A priori

we specified the following criteria to define success of this pilot feasibility study: 1)

timely enrollment within 96 hours of admission, 2) randomization within 24 hours of

enrollment, 3) treatment allocation within 48 hours of randomization, and 4) adherence

with weekly surveillance CUS during hospitalized. Secondary outcome measures

included: 1) event rates of PE, DVT, and death; 2) investigation of barriers to

recruitment and randomization; 3) pharmacologic prophylaxis initiation; 4) compliance

with follow-up; 5) filter retrieval rates.

Statistical Analysis

The goal of this feasibility study was to inform the planning of a large multicenter

randomized controlled clinical trial (RCCT). By design, the FIT Pilot Study was not

20

powered to determine the relative benefit of pIVCF versus no pIVCF on the

development of PE. A sample size of convenience was chosen based on hospital

record review, which indicated that approximately 200 patients per year received a

pIVCF prior to initiation of the FIT Pilot Study. A sample size of 100 patients was

targeted over a 12-month enrollment phase. It was felt that this sample size would be

sufficient to address issues related to the consent process, to refine the protocol as

needed, identify barriers to conduction of the study, and perform an interim analysis.

Statistical analysis was performed with SAS (version 9.2; SAS Inc., Cary, NC).

Descriptive statistical analysis was performed to compare characteristics of patients in

the two treatment groups (chi square test or Fischer’s exact test for bivariate categorical

data, Student’s t test for continuous variables with normal distribution and Mann-

Whitney test for nonparametric data). Logistic regression was performed to compare

binary outcomes of presence or absence of thrombotic events. As patients were

randomized to treatment groups, we did not perform multivariate regression.

21

CHAPTER 3 RESULTS

Between November 2008 and November 2010, 38 patients were enrolled and 30

of 38 were eligible for analysis (Figure 3-1). Differences between socio-demographic

and clinical characteristics in the IVCF group and non-IVCF group were not significantly

different (Table 3-1). The DSMB met every 4 months and all DSMB reports were

conveyed to the IRB. There were no acute complications related to the device

placement and no serious adverse events that required stopping the study by the

oversight of the DSMB.

Primary Outcomes

Results of the primary feasibility objectives (Table 3-2) included: time from

admission to enrollment (mean 45.7 ± 21.5hrs), time from enrollment to randomization

(mean 4.4 ± 8.5hrs), time from randomization to placement of IVCF (mean 17.2 ±

9.1hrs), and adherence to weekly CUS within first month (IVCF group = 46.7%, non-

IVCF group = 66.7%, p=0.28). Examining eligible patients that were ultimately not

enrolled in the study elucidated barriers to recruitment. One hundred four patients were

screened for participation in the study. Fifty-three patients who were considered eligible

for the study were not enrolled for the following reasons (Figure 3-1): patient or proxy

refusal of consent (n=23), inability to obtain informed consent and no next of kin

identified (n=12), research staff or translator unavailable to obtain informed consent

(n=6), delayed notification by primary service of eligible patient (n=2), patient received

pIVCF prior to notification of eligibility (n=5), site unable to place IVCF (n=4), patient

unavailable for consent (n=1). There were no barriers to randomization once informed

consent was obtained.

22

Secondary Outcomes

At 6-month follow-up, one PE was diagnosed in the non-IVCF group, one death

unrelated to VTE in the IVCF group, and no DVT in either group (Table 3-3). The

average time to pharmacologic prophylaxis was less than 24 hours in each cohort. A

trend towards extended pharmacologic prophylaxis in the non-IVCF patients was noted

at hospital discharge. At 1 month more patients in the non-IVCF group (93.3% in non-

IVCF and 46.7% in IVCF, p=0.01) were still on pharmacologic prophylaxis (Table 3-4).

Physician ordering of SCDs was documented in 80% of patients receiving IVCF and

93.3% of patients in the non-IVCF group (p=0.29).

In the IVCF group, we were able to reach 80% of patients for 1-month follow-up

compared with 100% in the non-IVCF group (p=0.07). At 6 months post-discharge,

follow-up dropped off in each group (40% in IVCF and 60% in non-IVCF, p=0.28) (Table

3-5). Of the 15 patients who received an IVCF, 13 patients were eligible for 3-month

follow-up with the vascular clinic and 10 patients for 6-month follow-up. No patients had

their filter removed at 3 months. At 6 months 2 of the 10 patients (20%) had their filter

removed. Of the remaining patients who did not have their filter removed, 20% were lost

to follow-up and 20% were deemed to be at continued risk for VTE.

23

Table 3-1. Sociodemographic and clinical variables according to treatment group

Variable IVCF (n = 15)

Non-IVCF (n = 15)

Total p-value

Male 11 10 0.7 Female 4 5 0.7 Age (years) 39.3± 16.2 54.2 ± 25.2 46.5 ± 22.0 0.12 GCS 14.3 ± 1.0 13.5 ± 3.7 13.9 ± 2.7 0.37 ISS 24.2 ± 15.4 24.9 ± 12.1 24.6 ± 13.5 0.84 Insurance status

Yes 9 11 0.45 No 6 4 0.45

Mechanism of injury Penetrating 0 1 0.32 Blunt 15 14 0.32

Qualifying injury pattern BLE fx 2 7 0.05 Complex pelvic fx 5 2 0.2 SCI BMI>35

2 3

4 4

0.37 0.67

NWB ≥7d 1 0 0.32 BLE fx+ SCI 0 1 0.32 Pelvic+ LE fx 2 1 0.32 NWB> 7d + BMI>35 1 0 0.32

GCS, Glasgow coma scale; ISS, injury severity score; BLE, bilateral lower extremity; fx, fracture; SCI, spinal cord injury; NWB, non-weight bearing; BMI, body mass index, LE, lower extremity; d, days Table 3-2. Primary outcomes of feasibility

Outcome IVCF (n = 15)

Non-IVCF (n = 15)

Total p-value

Time from admission to enrollment (h)

46.9 ±19.9 44.4 ± 23.6 45.7 ± 21.5 0.65

Time from enrollment to randomization (h)

2.8 ± 6.6 6.4 ± 10.4 4.4 ± 8.5 0.06

Time from randomization to IVCF placement (h)

17.2 ± 9.1 NA 17.2 ± 9.1

Adherence to weekly CUS within 1st month by service (%)

Trauma (n=14) 66.7 100 0.16 Neurosurgery (n=7) 50 60 0.82 Orthopedics (n=8) 0 50 0.13 Vascular (n=1) NA 0 Total (n=30) 46.7 66.7 0.28

h, hours; CUS, compression ultrasound

24

Table 3-3. Secondary clinical outcomes

Outcome IVCF (n = 15)

Non-IVCF (n = 15)

DVT Inpatient 0 0 At 1-mo follow-up 0 0 At 6-mo follow-up 0 0

PE Inpatient 0 0 At 1-mo follow-up 0 0 At 6-mo follow-up 0 1

Death VTE related 0 0 Non-VTE related 1 0

DVT, deep vein thrombosis; PE, pulmonary embolism; VTE, venous thromboembolism Table 3-4. Pharmacologic prophylaxis

Pharmacologic prophylaxis IVCF Non-IVCF p-value

Average time to initiation (h) 20.0 ± 13.2 23.5 ± 20.3 0.79 Pharmacologic prophylaxis at enrollment (%) 93.3 86.7 0.55 Pharmacologic prophylaxis at hospital discharge (%) 46.7 93.3 0.01 Pharmacologic prophylaxis at 1-month follow-up (%) 6.7 6.7 1.0

h, hours Table 3-5. Follow-up compliance

Follow-up IVCF (%) Non-IVCF (%) p-value

1 month 80.0 100.0 0.07 6 month 40.0 60.0 0.28

25

Figure 3-1. CONSORT flow diagram

26

CHAPTER 4 DISCUSSION

The results of this pilot study demonstrate that it is feasible to perform a

randomized study comparing pIVCF versus no IVCF for VTE prevention in high-risk

trauma patients. We designed the FIT Pilot Study to address four specific feasibility

objectives and several secondary outcomes. Our a priori defined feasibility objectives

were met. Patients were enrolled within 96 hours of admission to the trauma service,

randomized within 24 hours of enrollment, and received an IVCF if allocated to one

within 24 hours of randomization. Although adherence to weekly surveillance CUS was

suboptimal, patients who were not transferred off the trauma service throughout their

hospitalization did have better compliance with CUS. Conversely, patients transferred to

other surgical services during their hospitalization often did not continue to receive

weekly CUS. Lack of adherence to protocol may be explained by unfamiliarity with the

FIT Study by the non-trauma services as well as the fact that weekly CUS were not part

of their service’s standard of practice even before the FIT study opened, as was the

case with the trauma service.32,33

Fifty-three eligible patients were not enrolled. Barriers to enrolling these patients

included mainly patient refusal to participate or no next of kin identified for informed

consent, a common dilemma in trauma studies.34 Other barriers to recruitment were

delayed notification by the primary service of an eligible patient and inability to obtain

timely informed consent due to logistics (research staff/translator unavailable, patient

unavailable due to multiple operating room visits, site unable to place IVCF). As the

study progressed over 2 years, trauma attendings, residents, and physician extenders

became more familiar with the FIT study, which led to timely notification of the research

27

staff of eligibility status by the primary trauma service. To enhance this collaboration, we

advertised the FIT study with posters on the trauma units, as well as laminated pocket

cards detailing inclusion/exclusion criteria and research staff contact information. We

also regularly attended trauma service meetings to reintroduce the FIT study to new

residents rotating on the service. We faced several logistic barriers to enrolling patients.

For example, we had difficulty obtaining baseline CUS or informed consent within 96

hours of admission in patients with severe injuries who returned to the operating room

multiple times within the first few days of hospitalization. Also, since we did not initially

have a dedicated study coordinator, all eligible patients who were admitted to the

trauma service over a weekend were not enrolled.

Exclusion and inclusion criteria may have contributed to our low recruitment

rates. We excluded patients if they were ineligible for pharmacologic prophylaxis during

their entire hospitalization. The clinical question that was to be addressed was whether

IVCFs add any protection against venous thromboembolism beyond pharmacologic

prophylaxis delivered when deemed safe. In designing a multicenter randomized

controlled trial patients who are ineligible for pharmacologic prophylaxis should be

considered for inclusion. Finally, modifications to the inclusion criteria to reflect only the

highest-risk patients have contributed to the decreased recruitment rates. Revisions of

the inclusion criteria such as spinal cord injuries with paralysis, bilaterally long bone LE

fractures not including fibula fractures, non-weight bearing bilaterally, multiple complex

pelvic fractures may have narrowed the eligible patient population for the study. Prior to

the opening, the FIT study the criteria for pIVCF placement in HRTPs at the University

of Florida was determined by the attending physician and therefore highly variable and

28

likely more inclusive. These less strict inclusion criteria prior to the opening of the FIT

Study likely contributed to our over-estimation of sample size calculation based on prior

experiences.

No obstacles to randomization were encountered once informed consent was

signed. Randomization occurred within 24 hours of enrollment. Similarly, we found no

barriers to timely placement of the IVCF. Placement of the IVCF occurred within 24

hours of randomization to this group.

Several changes to the informed consent process were made over the course of

2 years. Initially, when the study opened in November 2008, the treating trauma

surgeon introduced the study and obtained informed consent. However, given the direct

patient care that the primary surgeon delivered to the patient and the variation in the

informed consent process between each trauma surgeon we shifted the duty of

obtaining informed consent to the principal investigator. Unfortunately, due to logistical

issues such as manpower and time constraints, this led to the inability to consent all

eligible patients in a timely manner. Approximately 1 year into the study we secured a

dedicated research coordinator for the FIT study who then performed all informed

consents. This key step of obtaining informed consent by one person led to a 65%

increase in enrollment rates as well as improved patient/physician compliance with the

study protocol. In addition, we instituted an informed consent documentation form to

help streamline the informed consent process. This form served as a reminder checklist

to document: 1) eligibility requirements, 2) the opportunity given to the patient and

family to discuss the study and ask questions, 3) copies of informed consent placed in

medical records and given to the patient, and 4) primary and secondary contact

29

information of patient and next-of-kin. In addition, the informed consent documentation

form served as a standardized script to explain the study and answer questions most

commonly asked by the patient and family.

Follow-up at 1-month post-discharge was significantly better than follow-up after

6 months. If a patient was not scheduled for a clinic appointment as per standard of

practice for their injuries the research coordinator conducted follow-up over the phone.

The sole reason for lack of follow-up with patients was inability to contact patients after

discharge because of inaccurate contact information in the hospital records. We

therefore introduced a contact information sheet as part of the informed consent

process. This included the patient’s and legal representative’s contact phone number

and address. At 3 months no patients had their IVCF removed. At 6 months only 2 of

the 10 patients eligible for removal had successful retrieval. Of those that did not have

retrieval, only 20% were deemed to be at continued risk for VTE justifying non-retrieval

of IVCF. Of the remainder who did not have the IVCF removed, 20% were lost to follow-

up with the vascular surgery service. This is consistent with prior reports in the literature

of retrieval rates of 20%.35 Reasons for lack of follow-up to remove filters and methods

to improve retrieval rates needs to be addressed prospectively given the recent

attention to complications of IVCFs reported in the literature.24, 36-37 On August 8, 2010,

the FDA issued a safety alert recommending that clinicians responsible for the ongoing

care of patients with retrievable IVCFs consider removing the filter as soon as protection

from PE is no longer needed.27 Therefore, improved retrieval rates of IVCFs is

mandatory and methods for increasing retrieval rates should be explored.

30

Despite recent reports of filter complications24, 36-37, placement of pIVCFs may be

appropriate when anticoagulation may be contraindicated in the short-term. When

removed in a timely fashion, retrievable filters may avoid the long-term complications

associated with permanent IVCFs. However, the efficacy of pIVCFs beyond appropriate

pharmacologic prophylaxis remains to be elucidated.

We conducted this pilot study feasibility to determine the feasibility of conducting

a future large multicenter RCCT evaluating the efficacy of pIVCFs in high-risk trauma.

Our results will help inform the design of a future large trial. We conclude that by

addressing the multifaceted issues in recruitment, protocol adherence, and follow-up we

faced during this pilot study a multicenter RCCT is feasible. Several key considerations

emerge from the results of this study that should be addressed in the protocol of a

future large trial. First, since the quality endpoints for our feasibility objectives were met

(e.g., enrollment within 96 hours of admission), shorter acceptable time frames for

quality endpoints can likely be implemented to prevent exclusion of patients that

develop VTE during the first few days of hospitalization. During the first 2 years of this

study our institution did not have an electronic medical record system. For recruitment

we relied on manual screening of the daily trauma census as well as timely notification

of a potential candidate by the primary trauma service. A future trial can maximize

recruitment rates by utilizing a computerized alert system that automatically and

accurately identifies potential study subjects, thereby improving efficiency of enrollment.

This electronic tool is especially important in time sensitive trials in trauma patients. In

addition, this computerized method would effectively reduce the screening burden while

minimizing error associated with recruitment by research staff.

31

In summary, this comprehensive pilot study served a critical function in

identifying conceptual and methodological issues and was an essential foundation step

in preparing for a larger trial investigating the role of IVCFs in trauma patients.

32

LIST OF REFERENCES

1. Rogers FB. Venous thromboembolism in trauma patients: a review. Surgery. 2001;130(1):1-12.

2. Geerts WH, Code, KI, Jay RM, Chen E, Szalai JP. A prospective study of venous thromboembolism after major trauma. NEJM 1994; 331(24): 1601-1606.

3. Knudson MM, Ikossi DG, Khaw L, Morabito D, Speetzen LS. Thromboembolism after trauma: an analysis of 1602 episodes from the American College of Surgeons National Trauma Data Bank. Annals of Surgery 2004; 240(3): 490-498.

4. Decousus H, Leizorovicz A, Parent F, Page Y, et al. A clinical trial of vena cava filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. NEJM 1998;338:409-415.

5. The PREPIC Study Group. Eight-year follow-up of patients with permanent vena cava filters in the prevention of pulmonary embolism: the PREPIC (Prevention du Risque d'Embolie Pulmonaire par Interruption Cave) Randomized Study. Circulation 2005;112(3):416-22.

6. Stein PD, Matta F, Hull RD. Increasing use of vena cava filters for prevention of pulmonary embolism. Am J Med 2011;124:655– 61.

7. Kitchens CS AB, Kessler CM, Alving BM. Consultative hemostasis and thrombosis. Philadelphia, PA: Saunders Elsevier, 2007, pp. 542-545.

8. Shackford SR, Cook A, Rogers FB, Littenberg B, Osler T. The increasing use of vena cava filters in adult trauma victims: data from the American College of Surgeons National Trauma Data Bank. J Trauma. 2007;63(4):764-9.

9. Aryafar H, Kinney TB. Optional inferior vena cava filters in the trauma patient. Semin Intervent Radiol 2010;27(1):68–80

10. Duszak R, Parker L, Levin DC, Rao VM. Placement and removal of inferior vena cava filters: national trends in the Medicare population. J Am Coll Radiol 2011;8:483-489.

11. Smouse B, Johar A. 2010. Is market growth of vena cava filters justified? Endovasc Today. 74-77.

12. Dossett LA, Adams RC, Cotton BA. Unwarranted national variation in the use of prophylactic inferior vena cava filters after trauma: an analysis of the National Trauma Databank. J Trauma 2011;70(5):1066-70

33

13. Kaufman JA, Rundback JH, Khee ST, Geerts W, et al. Development of a research agenda for inferior venca cava filters: proceedings from a multidisciplinary research consensus panel. J Vasc Interv radiol 2009;20:697-707.

14. Helling TS, Kaswan S, Miller SL, Tretter JF. Practice patterns in the use of retrievable inferior vena cava filters in a trauma population: a single center experience. J Trauma 2009: 67: 1293-1296.

15. Tschoe M, Kim HS, Brotman DJ, Streiff MB. Retrievable vena cava filters: a clinical review. J Hosp Med 2009; 4:441-448.

16. Athanasoulis CA, Kaufman JA, Halpern EF, Waltman AC, Geller SC, Fan CM. Inferior vena cava filters: review of a 26-year single-center clinical experience. Radiology 2000; 216:54-66.

17. Michaelov T, Blickstein D, Levy D, Berliner S, Ellis MH. Removal of retrievable vena cava filters in routine practice: a multicenter study. Eur J Intern Med. 2011 Dec;22(6):e87-9.

18. Gaspard SF, Gaspard DJ. Retrievable inferior vena cava filters are rarely removed. Am Surg 2009; 75:426–8.

19. Mission JF, Kerlan RK Jr, Tan JH, Fang MC. Rates and predictors of plans for inferior vena cava filter retrieval in hospitalized patients. J Gen Intern Med 2010;25:321–5

20. Gasparis AP, Spentzouris AP, Meisner RJ. Improving retrieval rates of temporary inferior vena cava filters. J Vasc Surg 2011 Dec;54(6 Suppl):34S-8S.e1.

21. Thoung Van Ha. Retirevable filters: maximizing retrieval rates. Semin Roentgenol. 2011 Apr;46(2):154-8.

22. Marquess JS, Burke CT, Beecham AH, Dixon RG, Stavas JM, Sag AA, Koch GG, Mauro MA. Factors associated with failed retrieval of the Gunther Tulip inferior vena cava filter. J Vasc Interv Radiol 2008;19(9):1321-7.

23. Hermsen JL, Ibele AR, Faucher LD, Nale JK, Schurr MJ, Kudsk KA. Retrievable inferior vena cava filters in high-risk trauma and surgical patients: factors influencing successful removal. World J Surg 2008;32(7):1444-9.

24. Nicholson W, Nicholson WJ, Tolerico P, Taylor B, Solomon S, et al. Prevalence of fracture and fragment embolization of Bard retrievable vena cava filters and clinical implications including cardiac perforation and tamponade. Arch Intern Med 2010;170(20):1827-1831.

34

25. Smouse HB, Rosenthal D, Thuong VH, Knox MF, Dixon RG, Voorhees WD 3rd, McCann-Brown JA. Long-term retrieval success rate profile for the Gunther Tulip vena cava filter. J Vasc Interventional Radiology 2009; 20(7):871-7.

26. Irwin E, Byrnes M, Schultz S, Chipman J, Beal A, Ahrendt M, Beilman G, Croston JK. A systematic method for follow-up improves removal rates for retrievable inferior vena cava filters in a trauma patient population. J Trauma. 2010;69(4):866-9.

27. Inferior Vena Cava (IVC) Filters: Initial Communication: Risk of Adverse Events with Long Term Use. 2010. (Accessed November 7th, 2010, at http://www.fda.gov/Safety/MedWatch/SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm221707.htm.)

28. Rogers FB, Cipolle MD, Velmahos G, Rozycki G, Luchette FA. Practice management guidelines for the prevention of venous thromboembolism in trauma patients: the EAST practice management guidelines work group. J Trauma 2002;53:142-164.

29. Gould MK, Garcia DA, Wren SM, Karanicolas PJ, Arcelus JI, Heit JA, Samama CM; American College of Chest Physicians. Prevention of VTE in nonorthopedic surgical patients: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. CHEST. 2012;141(2_suppl):e227S-e277S.

30. Jairath N, Hogerny M, Parsons C. The role of the pilot study: a case illustration from cardiac nursing research. Appl Nurs Res. 2000;13(2):92-96.

31. Cook DJ, Rocker G, Meade M, Guyatt G, et al.; PROTECT Investigators; Canadian Critical Care Trials Group. Prophylaxis of Thromboembolism in Critical Care (PROTECT) Trial: a pilot study. J Crit Care. 2005;20(4):364-372.

32. Piotrowski JJ, Alexander JJ, Brandt CP, McHenry CR, Yuhas JP, Jacobs D. Is deep vein thrombosis surveillance warranted in high-risk trauma patients? Am J Surg. 1996; 172(2):210-3.

33. Adams RC, Hamrick M, Berenguer C, Senkowski C, Ochsner MG. Four years of an aggressive prophylaxis and screening protocol for venous thromboembolism in a large trauma population. J Trauma. 2008;65(2):300-6.

34. Holcomb JB, Weiskopf R, Champion H, Gould SA. Challenges to effective research in acute trauma resuscitation: consent and endpoints. Shock. 2011;35(2):107-113.

35. Johnson ON 3rd, Gillespie DL, Aidinian G, White PW, Adams E, Fox CJ. The use of retrievable vena cava filters in severely injured military trauma patients. J Vasc Surg. 2009; 49:410-416.

35

36. Mahrer A, Zippel D, Garniek A, Golan G, Bensaid P, Simon D, Rimon U. Retrievable Vena Cava Filters in Major Trauma Patients: Prevalence of Thrombus Within the Filter. Cardiovasc Intervent Radiol. 2008;31:785-789.

37. Arabi M, Willatt JM, Shields JJ, Cho KJ, Cwikiel WB. Retrievablily of optional inferior vena cava filters with caudal migration and caval penetration: report of three cases. J Vasc Interv Radiol. 2010;21(6):923-6.

36

BIOGRAPHICAL SKETCH

Anita Rajasekhar, M.D., is an associate professor in the Department of Medicine

at the University of Florida College of Medicine. Dr. Rajasekhar received her M.D.

from the University of Florida in 2004. She then completed her internal medicine

residency in 2007 and hematology/oncology fellowship in 2010 at the University of

Florida. Dr. Rajasekhar is board certified in medical oncology and hematology. She

joined the University of Florida as an assistant professor in 2010. Dr. Rajasekhar

graduated with a Master of Science in Clinical and Translational Science from the

University of Florida in 2012. Dr. Rajasekhar’s primary academic interests include

diagnosis and treatment of benign hematologic conditions, particularly related to

thrombosis and hemostasis. Her academic focus is on the prevention and treatment

of venous thromboembolism. She is a member of the American Society of

Hematology (ASH).