CLINICAL RESEARCH STUDY

Inferior Vena Cava Filters in Elderly Patients with

Stable Acute Pulmonary Embolism
Paul D. Stein, MD, Fadi Matta, MD, Mary J. Hughes, DO
Department of Osteopathic Medical Specialties, College of Osteopathic Medicine, Michigan State University, East Lansing.

ABSTRACT

BACKGROUND: Patients aged >60 years with pulmonary embolism who were stable and did not require
thrombolytic therapy were shown to have a somewhat lower in-hospital all-cause mortality with vena cava
filters. In this investigation we further assess mortality with filters in stable elderly patients.
METHODS: In-hospital all-cause mortality according to use of inferior vena cava filters was assessed from
the National (Nationwide) Inpatient Sample, 2003-2012, in: 1) All patients with pulmonary embolism; 2)
All with pulmonary embolism who had none of the comorbid conditions listed in the Charlson Comorbidity
Index; 3) Patients with a primary (first-listed) diagnosis of pulmonary embolism, and 4) Patients with a
primary diagnosis of pulmonary embolism and none of the comorbid conditions listed in the Charlson
Comorbidity Index.
RESULTS: From 2003-2012, 2,621,575 stable patients with pulmonary embolism were hospitalized in the
US. Patients aged >80 years showed lower mortality with vena cava filters (all pulmonary embolism, 6.1%
vs 10.5%; all pulmonary embolism with no comorbid conditions, 3.3% vs 6.3%; primary pulmonary
embolism, 4.1% vs 5.7%; primary pulmonary embolism with no comorbid conditions, 2.1% vs 3.7%; all
P <.0001). In the all-patient category, patients aged 71-80 years showed somewhat lower mortality with
filters, 6.3% vs 7.4% (P <.0001), and those without comorbid conditions, 2.5% vs 2.8% (P ¼ .04). Those
aged 71-80 years with primary pulmonary embolism, irrespective of comorbid conditions, did not show
lower mortality with filters.
CONCLUSION: At present, in the absence of a randomized controlled trial, it seems prudent to consider a
vena cava filter in very elderly (aged >80 years) stable patients with acute pulmonary embolism.
Ó 2016 Elsevier Inc. All rights reserved.  The American Journal of Medicine (2016) -, ---

KEYWORDS: Elderly; Mortality; Pulmonary embolism; Vena cava filters

Among patients with pulmonary embolism who were stable
(not in shock or on ventilatory support) and did not receive
thrombolytic therapy, in-hospital all-cause mortality was
marginally lower in those who received an inferior vena
cava filter compared with those who did not, 21,420 of
297,700 (7.2%), compared with 135,240 of 1,712,800
(7.9%).1 This observation was based on administrative data
from the Nationwide Inpatient Sample.1 Subsequently, a
Funding: None.
Conflict of Interest: None.
Authorship: All authors had access to the data and participated in
preparation of the manuscript.
Requests for reprints should be addressed to Paul D. Stein, MD,
Department of Osteopathic Medical Specialties, College of Osteopathic
Medicine, Michigan State University, 909 Fee Road, East Lansing,
MI 48824.
E-mail address: steinp@msu.edu
randomized controlled trial of inferior vena cava filters in
acute pulmonary embolism, Prévention du Risque d’Em-
bolie Pulmonaire par Interruption Cave2 (PREPIC2),
showed no reduction of mortality with filters in stable
patients with pulmonary embolism.2 However, with only
200 patients included in the treatment arm and 199 patients
in the control arm, it was not possible to stratify according to
age or any other category.
An investigation of vena cava filters in elderly patients
(65 years of age) based on a national cohort study of
Medicare beneficiaries, showed no lower all-cause mortality
at 30 days with inferior vena cava filters.3 This prompted us
to assess our published in-hospital data in older patients
(>60 years) with pulmonary embolism who were stable and
did not require thrombolytic therapy.4 A somewhat lower
in-hospital all-cause mortality was shown in such patients
with vena cava filters 1999-2008, 8.0% with filters,

0002-9343/$ -see front matter Ó 2016 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.amjmed.2016.09.033
2 The American Journal of Medicine, Vol -, No -, - 2016

compared with 10.2% without filters.4 This led us to further which patients were excluded are shown in Table 1. We
assess whether there might be lower in-hospital mortality were unable to determine if the filters were temporary or
with vena cava filters in stable elderly patients, particularly permanent.
the very elderly, which is the purpose of this investigation. Comorbid conditions listed in the Charlson Comorbidity
Index6 as well as the ICD-9-CM codes used to identify
these comorbid conditions are shown in Table 2.
METHODS
We analyzed administrative data
CLINICAL SIGNIFICANCE Statistical Methods
from the National (Nationwide)
Differences in mortality rates (case
Inpatient Sample (NIS), Health-  The vast majority of patients with acute fatality rates) were assessed by
care Cost and Utilization Project, pulmonary embolism are stable. Fisher’s 2-tailed exact test using
Agency for Healthcare Research
and Quality, 2003-2012. Each5  In-hospital all-cause mortality of stable GraphPad Software (San Diego,
patients with acute pulmonary embolism CA). Relative risk and 95% con-
year of the NIS provides infor-
fidence intervals were calculated
mation on approximately 8 million decreased during 2003-2012 in patients
using calculator for confidence
inpatient stays from about 1000 with and without vena cava filters.
intervals of relative risk (www.
hospitals. The NIS is designed to
 Very elderly (aged >80 years) patients sign.ac.uk/methodology/risk.xls).
approximate a 20% sample of US
non-Federal, short-term, general, with stable acute pulmonary embolism Linear regression analyses were
5 showed lower in-hospital all-cause mor- performed using SPSS Version
and other specialty hospitals.
Beginning with data from 2012, tality with vena cava filters. 22 for Windows (SPSS Inc,
Chicago, IL).
the NIS was redesigned to improve
national estimates. To highlight the
design change, beginning with 2012 data, the database was
renamed from the “Nationwide Inpatient Sample” to the RESULTS
“National Inpatient Sample.” The NIS is now a sample of
discharge records from all Healthcare Cost and Utilization All Stable Patients with Pulmonary Embolism
Project-participating hospitals, rather than a sample of hos- From 2003-2012, 2,765,640 patients were discharged from
pitals from which all discharges were retained.5 short-stay hospitals in the US with pulmonary embolism.
We determined the in-hospital all-cause mortality Among these, 2,621,575 (94.8%) were stable and did not
according to age among stable patients with pulmonary receive thrombolytic therapy or a pulmonary embolectomy.
embolism, defined as those not in shock or on ventilatory Women were 54.0% (P <.0001). Mortality was 5.8% in
support. We analyzed 4 categories of stable patients: 1) All women and 6.1% in men (P <.0001). Most patients (75.0%)
patients with pulmonary embolism, 2) All patients with were White. Mortality among the races ranged from 5.8% to
pulmonary embolism who had none of the comorbid 8.7%. The majority of all stable patients with pulmonary
conditions listed in the Charlson Comorbidity Index,6 3) embolism, 59.5%, were aged 61 years or older, and 17.9%
Patients with a primary (first-listed) diagnosis of pulmonary were age >80 years.
embolism, and 4) Patients with a primary diagnosis of Among stable patients of all ages, in-hospital mortality in
pulmonary embolism and none of the comorbid conditions those who received a vena cava filter was 5.5%, compared
listed in the Charlson Comorbidity Index. with 6.0% who did not receive a vena cava filter (P <.0001)
Included patients were adults (aged 18 years) of both (Table 3). Among patients aged 61-70 years, mortality in
sexes and all races hospitalized in short-stay hospitals from those who received a vena cava filter was marginally
all regions of the US. We assume that patients with a lower than in those who did not, 5.5% compared with
first-listed diagnosis were admitted to the hospital because 5.8% (P <.001) (Figure 1, Table 3). In patients aged
of pulmonary embolism, and we define this as primary 71-80 years, mortality in those who received a filter was
pulmonary embolism. 6.3%, compared with 7.4% in those who did not
Excluded patients were those in shock or on ventilatory (P <.0001). In patients aged >80 years, in-hospital all-
support, who we define as unstable. Unstable patients cause mortality was 6.1% with a vena cava filter, compared
previously were shown to have a lower in-hospital mortality with 10.5% in those who did not receive a filter (P <.0001).
rate with vena cava filters.1 Patients administered throm- Among patients aged 60 years or younger, in-hospital
bolytic therapy or who underwent pulmonary embolectomy mortality was not lower with vena cava filters (Figure 1).
were also excluded. Such patients also were shown to have a Among patients of all ages, in-hospital mortality
lower in-hospital all-cause mortality with vena cava decreased from 2003-2012 in those who received a vena
filters.1,7 Patients younger than age 18 years were excluded. cava filter and in those who did not (Figure 2). In those
The International Classification of Diseases, Ninth aged >80 years, relative risk with a vena cava filter,
Revision, Clinical Modification (ICD-9-CM) codes for 2003-2012, ranged from 0.4 to 0.7, even though mortality
pulmonary embolism, vena cava filter, and conditions for in both groups decreased (Table 4).
Stein et al Vena Cava Filters in the Elderly 3

Table 1 International Classification of Diseases, Ninth Stable Patients with a Primary Diagnosis of
Revision, Clinical Modification (ICD-9-CM) Codes for Pulmonary Pulmonary Embolism
Embolism, Vena Cava Filter, and Conditions for Which Patients From 2003-2012, 1,516,753 stable patients with a primary
Were Excluded diagnosis of pulmonary embolism were hospitalized.
Condition ICD-9-CM Code Women constituted 54.7% (P <.0001). Mortality was the
same in women and men. Whites constituted 75.8%. Mor-
Pulmonary embolism 415.1
tality ranged from 2.8% to 3.4% among the races. Mortality
Vena cava filter 38.7*
Thrombolytic therapy 99.1* decreased from 2003-2012, both in those with vena cava
Embolectomy 38.05* filters and those without (Table 6). Mortality was lower in
Shock 785.5 those >aged 80 years who received filters, 4.1%
Ventilator dependence V46.1 compared with 5.7% (P <.0001), but in patients aged 80
*Procedure code.
years or younger, mortality was not lower (Table 6,
Figure 4). In those aged >80 years, relative risk with a
vena cava filter ranged from 0.6 to 0.9 every year except
2004, at which time relative risk was 1 (Table 7).
All Stable Patients with Pulmonary Embolism
and No Comorbid Conditions
There were 908,142 patients hospitalized during 2003-2012 Stable Patients with a Primary Diagnosis of
with stable pulmonary embolism and none of the comorbid Pulmonary Embolism and No Comorbid
conditions listed in the Charlson Comorbidity Index Conditions
(Table 5). Mortality decreased from 2003-2012 both in From 2003-2012, 615,805 stable patients with a primary
those with vena cava filters and those who did not receive diagnosis of pulmonary embolism and none of the comorbid
filters (Table 5). Mortality was lower in those aged >80 conditions listed in the Charlson Comorbidity Index were
years who received filters, 3.3% compared with 6.3% hospitalized. Mortality decreased from 2003-2012 in both
(P <.0001), and it was marginally lower in those aged those with vena cava filters and those who did not receive
61-80 years (Table 5, Figure 3). In those aged >80 filters (Table 8). Mortality was lower in those aged >80
years, relative risk with a vena cava filter in 2003-2012 years who received filters, 2.1% compared with 3.7%
ranged from 0.2 to 0.9 (Table 4). (P <.0001), but mortality was not lower in patients aged
71-80 years (Table 8, Figure 5). In patients 70 years of
age or younger, mortality comparing filters to no filters
Table 2 International Classification of Diseases, 9th Revision, varied. In those aged >80 years, relative risk with vena
Clinical Modification (ICD-9-CM) Comorbid Conditions Included cava filters, 2003-2012, ranged from 0.2 to 0.8 every year
in the Charlson Index6 except 2005, at which time relative risk was 1.0, but
mortality data were sparse in those who received a vena
Comorbid Condition ICD-9-CM Codes Used cava filter (Table 7).
Comorbid conditions Included in the Charlson Index
Acute myocardial infarction 410
Heart failure 428 DISCUSSION
Peripheral vascular disease 440.2, 443.9 Stable patients in the 4 categories investigated (all patients
Cerebrovascular disease 430-438 with pulmonary embolism, all patients who had no comor-
Dementia 290 bid conditions, patients with a primary diagnosis of pul-
Chronic obstructive pulmonary 490-496 monary embolism, and patients with a primary diagnosis of
disease pulmonary embolism and no comorbid conditions) showed
Rheumatologic disease 710.0, 710.1, 710.4,
lower in-hospital all-cause mortality with a vena cava filter,
714.0, 714.1,
providing they were very elderly (aged >80 years). The
714.2, 714.8
Ulcer disease 531-534 reduced mortality with filters in the very elderly, with 2
Acute or chronic liver disease 570, 571 exceptions, was shown each year. All stable patients with
Diabetes mellitus 250.0-250.3 pulmonary embolism, aged 71-80 years, with and without
Hemiplegia and hemiparesis 342.0-342.9 comorbid conditions, showed a somewhat lower mortality
Paraplegia 344.1 with vena cava filters. Patients with a primary diagnosis of
Moderate or severe renal disease 580-586, 588 pulmonary embolism, aged 71-80 years, with and without
Diabetes with chronic complications 250.4-250.6 comorbid conditions, did not show lower mortality with
Any neoplasms, leukemia, lymphoma 140-195, 200-208 filters. Mortality decreased from 2003-2012 in each of the 4
Metastatic cancer 196-199 categories, both in those who received filters and those who
HIV and AIDS 042
did not.
AIDS ¼ acquired immune deficiency syndrome; HIV ¼ human The present data, 2003-2012, showed lower mortalities
immunodeficiency virus.
among stable patients (5.5% mortality with filters compared
4 The American Journal of Medicine, Vol -, No -, - 2016

Table 3 Mortality in All Patients with Stable Pulmonary Embolism According to Year, Age and Inferior Vena Cava Filters
IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
N n % (95% CI) n n % (95% CI) RR (95% CI)
Year
2003 26,779 2014 7.5 (7.2-7.8) 156,880 12,536 8.0 (7.9-8.1) 0.9 (0.9-1.0)
2004 29,424 2217 7.5 (7.2-7.8) 166,220 12,777 7.7 (7.6-7.8) 1.0 (0.9-1.0)
2005 32,352 2207 6.8 (6.5-7.1) 182,510 12,833 7.0 (6.9-7.1) 1.0 (0.9-1.0)
2006 37,709 2319 6.1 (5.9-6.4) 201,990 13,304 6.6 (6.5-6.7) 0.9 (0.9-1.0)
2007 39,491 2313 5.9 (5.6-6.1) 220,553 14,382 6.5 (6.4-6.6) 0.9 (0.9-0.9)
2008 43,110 2428 5.6 (5.4-5.8) 249,313 15,451 6.2 (6.1-6.3) 0.9 (0.9-1.0)
2009 43,389 2150 5.0 (4.8-5.2) 255,416 14,117 5.5 (5.4-5.6) 0.9 (0.9-0.9)
2010 42,980 1798 4.2 (4.0-4.4) 267,595 13,642 5.1 (5.0-5.2) 0.8 (0.8-0.9)
2011 40,214 1535 3.8 (3.6-4.0) 282,276 13,938 4.9 (4.9-5.0) 0.8 (0.7-0.8)
2012 35,735 1350 3.8 (3.6-4.0) 267,640 11,915 4.5 (4.4-4.5) 0.9 (0.8-.0.9)
All years 371,183 20,331 5.5 (5.4-5.5) 2,250,393 134,895 6.0 (6.0-6.0) 0.9 (0.9-0.9)
Age, years
18-20 1016 18 1.8 (1.0-2.6) 13,391 135 1.0 (0.8-1.2) 1.7 (1.1-2.8)
21-30 6681 228 3.4 (3.0-3.8) 94,901 1578 1.7 (1.6-1.7) 2.0 (1.8-2.3)
31-40 15,460 444 2.9 (2.6-3.1) 166,014 3735 2.2 (2.2-2.3) 1.3 (1.2-1.4)
41-50 35,222 1395 4.0 (3.8-4.2) 283,760 8751 3.1 (3.0-3.1) 1.3 (1.2-1.3)
51-60 59,718 3106 5.2 (5.0-5.4) 384,562 18,764 4.9 (4.8-4.9) 1.1 (1.0-1.1)
61-70 78,295 4336 5.5 (5.4-5.7) 448,212 26,190 5.8 (5.8-5.9) 1.0 (0.9-1.0)
71-80 92,085 5761 6.3 (6.1-6.4) 473,046 34,994 7.4 (7.3-7.5) 0.9 (0.8-0.9)
>80 82,706 5043 6.1 (5.9-6.3) 386,507 40,748 10.5 (10.4-10.6) 0.8 (0.6-0.6)
All ages 371,183 20,331 5.5 (5.4-5.5) 2,250,393 134,895 6.0 (6.0-6.0) 0.9 (0.9-0.9)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.

with 6.0% without filters) than shown in 1999-2008 (7.2%
mortality with filters, compared with 7.9% without filters).4
This reflects the decreasing mortality from pulmonary em-
bolism in stable patients. We previously showed that mor-
tality decreased from 1999-2008 in stable low-risk patients,
irrespective of the use of vena cava filters.8 The present
investigation shows that this trend has continued through
2012. A decreasing 30-day mortality with and without vena
cava filters was shown by others in 1999-2010.3 The lower
mortality with vena cava filters that we observed reflects the
effectiveness of vena cava filters in the elderly (aged >60
years). Elderly patients constituted over half of those in the
present investigation.
Among all stable patients aged >60 years in both the
present data, 2003-2012, and in 1999-2008,4 in-hospital
mortality was lower in those who received a vena cava

12 9

All Stable PE All Stable PE

No IVCF
10 8

8 7
Mortality (%)
Mortality (%)

6 6
IVCF
No IVCF
4 5

2 4

IVCF
- 3
18-20 21-30 31-40 41-50 51-60 61-70 71-80 > 80 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Age (Years) Year

Figure 1 In-hospital all-cause mortality according to age Figure 2 In-hospital all-cause mortality according to year
and use of an inferior vena cava filter (IVCF) in all stable and use of inferior vena cava filter (IVCF) in all stable
patients with pulmonary embolism (PE). With IVCF, patients with pulmonary embolism (PE). With IVCF,
r ¼ 0.962, P <.0001. With no IVCF, r ¼ 0.968, P <.0001. r ¼ 0.991, P <.0001. With no IVCF, r ¼ 0.993, P <.0001.
Stein et al Vena Cava Filters in the Elderly 5

Table 4 Mortality in Patients Older Than Aged 80 Years

IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
n n % (95% CI) n n % (95% CI) RR (95% CI)
All stable pulmonary embolism
Year
2003 5387 534 9.9 (9.1-10.7) 26,180 3615 13.8 (13.4-14.2) 0.7 (0.7-0.8)
2004 5851 561 9.6 (8.8-10.3) 29,165 3948 13.5 (13.1-13.9) 0.7 (0.7-0.8)
2005 7134 486 6.8 (6.2-7.4) 33,060 4004 12.1 (11.8-12.4) 0.6 (0.5-0.6)
2006 8496 586 6.9 (6.4-7.4) 36,389 4139 11.4 (11.0-11.7) 0.6 (0.6-0.7)
2007 8871 608 6.9 (6.3-7.4) 38,606 4240 11.0 (10.7-11.3) 0.6 (0.6-0.7)
2008 10,147 596 5.9 (5.4-6.3) 43,999 4743 10.8 (10.5-11.1) 0.6 (0.5-0.6)
2009 9794 525 5.4 (4.9-5.8) 43,249 4401 10.2 (9.9-10.4) 0.6 (0.5-0.6)
2010 9536 477 5.0 (4.6-5.4) 43,278 3931 9.1 (8.8-9.3) 0.6 (0.5-0.6)
2011 9500 345 3.6 (3.3-4.0) 48,476 4252 8.8 (8.5-9.3) 0.4 (0.4-0.5)
2012 7990 325 4.1 (3.6-4.5) 44,105 3475 7.9 (7.6-8.1) 0.5 (0.5-0.6)
All years 82,706 5043 6.1 (5.9-6.3) 386,507 40,748 10.5 (10.4-10.6) 0.6 (0.6-0.6)
Stable pulmonary embolism and no comorbid conditions
Year
2003 1502 99 6.6 (5.3-7.8) 8471 796 9.4 (8.8-10.0) 0.7 (0.6-0.9)
2004 1712 71 4.1 (3.2-5.1) 9025 737 8.2 (7.6-8.7) 0.5 (0.4-0.7)
2005 1908 77 4.0 (3.2-4.9) 10,155 777 7.7 (7.1-8.1) 0.5 (0.4-0.7)
2006 2079 72 3.5 (2.7-4.2) 10,827 750 6.9 (6.4-7.4) 0.5 (0.4-0.7)
2007 1967 96 4.9 (3.9-5.8) 9595 653 6.8 (6.3-7.3) 0.7 (0.6-0.9)
2008 2359 94 4.0 (3.2-4.8) 10,977 656 6.0 (5.5-6.4) 0.7 (0.5-0.8)
2009 2032 30 1.5 (1.0-2.0) 10,849 682 6.3 (5.8-6.7) 0.2 (0.2-0.4)
2010 2246 78 3.5 (2.7-4.2) 10,207 407 4.0 (3.6-4.4) 0.9 (0.7-1.1)
2011 2126 16 0.8 (0.4-1.1) 11,804 582 4.9 (4.5-5.3) 0.2 (0.1-0.3)
2012 1580 15 0.9 (0.5-1.4) 10,150 425 4.2 (3.8-4.6) 0.2 (0.1-0.4)
All years 19,511 648 3.3 (3.1-3.6) 102,060 6465 6.3 (6.2-6.5) 0.5 (0.5-0.6)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.

Table 5 Mortality in All Patients with Stable Pulmonary Embolism and No Comorbid Conditions According to Year, Age and Inferior Vena
Cava Filters
IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
n n % (95% CI) n n % (95% CI) RR (95% CI)
Year
2003 8008 291 3.6 (3.2-4.0) 64,533 2244 3.5 (3.3-3.6) 1.0 (0.9-1.2)
2004 8724 243 2.8 (2.4-3.1) 67,144 1976 2.9 (2.8-3.1) 0.9 (0.8-1.1)
2005 9631 249 2.6 (2.3-2.9) 72,178 2109 2.9 (2.8-3.0) 0.9 (0.8-1.0)
2006 10,525 206 2.0 (1.7-2.2) 77,937 1899 2.4 (2.3-2.5) 0.8 (0.7-0.9)
2007 9751 220 2.3 (2.0-2.5) 75,097 1578 2.1 (2.0-2.2) 1.1 (0.9-1.2)
2008 10,923 261 2.4 (2.1-2.7) 86,939 1751 2.0 (1.9-2.1) 1.2 (1.0-1.3)
2009 10,465 157 1.5 (1.3-1.7) 89,174 1552 1.7 (1.7-1.8) 0.9 (0.7-1.0)
2010 10,659 164 1.5 (1.3-1.8) 92,844 1295 1.4 (1.3-1.5) 1.1 (0.9-1.3)
2011 9584 94 1.0 (0.8-1.2) 95,511 1450 1.5 (1.4-1.6) 0.6 (0.5-0.8)
2012 8415 105 1.2 (1.0-1.5) 90,100 1090 1.2 (1.1-1.3) 1.0 (0.8-1.3)
All years 96,685 1990 2.1 (2.0-2.1) 811,457 16,944 2.1 (2.1-2.1) 1.0 (0.9-1.0)
Age, years
18-20 705 0 e 9303 53 0.6 (0.4-0.7) e
21-30 4083 57 1.4 (1.0-1.8) 64,604 485 0.8 (0.7-0.8) 1.8 (1.4-2.4)
31-40 7434 114 1.5 (1.3-1.8) 99,481 851 0.9 (0.8-0.9) 1.8 (1.5-2.2)
41-50 13,490 170 1.3 (1.1-1.4) 139,523 1358 1.0 (0.9-1.0) 1.3 (1.1-1.5)
51-60 15,930 284 1.8 (1.6-2.0) 140,022 2015 1.4 (1.4-1.5) 1.2 (1.1-1.4)
61-70 16,650 236 1.4 (1.2-1.6) 130,659 2164 1.7 (1.6-1.7) 0.9 (0.8-1.0)
71-80 18,882 481 2.5 (2.3-2.8) 125,805 3553 2.8 (2.7-2.9) 0.9 (0.8-1.0)
>80 19,511 648 3.3 (3.1-3.6) 102,060 6465 6.3 (6.2-6.5) 0.5 (0.5-0.6)
All ages 96,685 1,990 2.1 (2.0-2.1) 811,457 16,944 2.1 (2.1-2.1) 1.0 (0.9-1.0)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.
6 The American Journal of Medicine, Vol -, No -, - 2016

8
cava filters in pulmonary embolism is based on the results of
All Stable PE No Comorbid CondiƟons the randomized controlled PREPIC2 investigation.2 Chest
7
No IVCF recommended against the use of inferior vena cava in
6 patients with pulmonary embolism.9 The results of the
5
present investigation suggest that vena cava filters reduce in-
hospital mortality in very elderly stable patients (aged >80
Mortality (%)

4
3 with acute pulmonary embolism. The data are consistent
2
1 prospective randomized controlled trial will be performed
- with vena cava filters in stable patients with pulmonary
18-20 21-30 31-40 41-50 51-60
Age (Years)
Figure 3 In-hospital all-cause mortality according to age
and use of inferior vena cava filter (IVCF) in all stable
patients with pulmonary embolism (PE) and no comorbid
conditions. With IVCF, r ¼ 0.874, P ¼ .005. With no IVCF,
r ¼ 0.821, P ¼ .01.
filter. Both of these analyses differed from Medicare data in
patients aged 65 years that showed no lower mortality at 30
days.3 Thirty-day mortality is higher than in-hospital mor-
tality.3 Whether this explains the difference between our
observations and those based on Medicare data is uncertain.
The 2016 Chest guideline9 and expert panel report for
venous thromboembolic disease in regard to the use of vena
IVCF
years). The very elderly constitute 17.9% of stable patients
over years 2003-2012, and are based on observations in
469,000 patients older than aged 80 years. It is remote that a
61-70 71-80 > 80
embolism older than aged 80 years. The PREPIC2 study
required 6 years for 18 centers to recruit 399 patients.2 To
recruit 399 patients aged >80 years at the same rate of
recruitment, 33 years would be required for 18 centers, or 6
years would be required for 100 centers.
We speculate that the lower mortality in elderly patients
who receive a vena cava filter may relate to the fragile
condition of elderly patients. Perhaps the prevention of even
a small recurrent pulmonary embolism in elderly patients
may be life saving.
Among stable patients, subgroups in addition to elderly
patients appear to have lower in-hospital all-cause mortality
with vena cava filters. Among stable patients with pulmonary
embolism and chronic obstructive pulmonary disease, mor-
tality was 8.7% with vena cava filters, compared with 11.0%
without filters in those older than aged 50 years.10 The greatest
reduction of mortality with vena cava filters in patients with

Table 6 Mortality in Patients with Stable Primary Pulmonary Embolism According to Year, Age and Inferior Vena Cava Filters
IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
n n % (95% CI) n n % (95% CI) RR (95% CI)
Year
2003 15,425 792 5.1 (4.8-5.5) 100,557 3933 3.9 (3.8-4.0) 1.3 (1.2-1.4)
2004 16,921 916 5.4 (5.1-5.7) 105,040 3887 3.7 (3.6-3.8) 1.4 (1.3-1.5)
2005 17,719 786 4.4 (41-4.7) 114,393 3628 3.2 (3.1-3.3) 1.4 (1.3-1.5)
2006 18,812 734 3.9 (3.6-4.2) 127,220 3920 3.1 (3.0-3.2) 1.3 (1.2-1.4)
2007 19,100 643 3.4 (3.2-3.7) 131,833 3641 2.8 (2.7-2.8) 1.2 (1.1-1.3)
2008 20,685 713 3.4 (3.2-3.7) 140,467 3838 2.7 (2.6-2.8) 1.3 (1.2-1.4)
2009 20,562 670 3.3 (3.0-3.5) 142,633 3471 2.4 (2.4-2.5) 1.3 (1.2-1.4)
2010 20,348 563 2.8 (2.5-3.0) 152,435 3492 2.3 (2.2-2.4) 1.2 (1.1-1.3)
2011 19,355 511 2.6 (2.4-2.9) 161,068 3488 2.2 (2.1-2.2) 1.2 (1.1-1.3)
2012 17,370 435 2.5 (2.3-2.7) 154,810 3215 2.1 (2.0-2.1) 1.2 (1.1-1.3)
All years 186,297 6763 3.6 (3.5-3.7) 1,330,456 36,513 2.7 (2.7-2.8) 1.3 (1.3-1.3)
Age, years
18-20 387 10 2.6 (1.0-4.1) 8152 22 0.3 (0.2-0.4) 9.4 (4.5-19.6)
21-30 2890 73 2.5 (2.0-3.1) 61,297 396 0.6 (0.6-0.7) 3.8 (3.0-4.9)
31-40 7848 147 1.9 (1.6-2.2) 109,867 766 0.7 (0.6-0.7) 2.7 (2.2-3.2)
41-50 18,021 348 1.9 (1.7-2.1) 182,584 2185 1.2 (1.1-1.2) 1.6 (1.4-1.8)
51-60 29,212 974 3.3 (3.1-3.5) 229,461 4800 2.1 (2.0-2.1) 1.6 (1.5-1.7)
61-70 38,394 1351 3.5 (3.3-3.7) 256,808 6607 2.6 (2.5-2.6) 1.4 (1.3-1.4)
71-80 46,021 2061 4.5 (4.3-4.7) 267,377 9570 3.6 (3.5-3.6) 1.2 (1.2-1.3)
>80 43,524 1799 4.1 (3.9-4.3) 214,910 12,167 5.7 (5.6-5.8) 0.7 (0.7-0.8)
All ages 186,297 6763 3.6 (3.5-3.7) 1,330,456 36,513 2.7 (2.7-2.8) 1.3 (1.3-1.3)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.
Stein et al Vena Cava Filters in the Elderly 7

6
solid malignant tumors, but there was variability according
Primary Stable PE No IVCF to the type of tumor and age of the patient.11 Among pa-
5 tients with pulmonary embolism and solid malignant tu-
mors, mortality among those aged >30 years with filters
4 was 10.2%, compared with 14.9% without filters.11 The
IVCF greatest reduction of mortality with filters was in patients
Mortality (%)

3
with pulmonary embolism and carcinoma of the prostate,
4.3% mortality with filters compared with 10.5% without
2
filters. Among patients with carcinoma of the cervix, mor-
1
tality with filters was 5.0%, compared with 11.7% without
filters. Those with pulmonary embolism and carcinoma of
- the rectum, rectosigmoid junction, and anus showed 6.2%
18-20 21-30 31-40 41-50 51-60 61-70 71-80 > 80 mortality with filters and 12.8% without filters.11
Age (Years) A few stable patients, 1.4%, received thrombolytic
therapy.1 Such patients who received an inferior vena cava
Figure 4 In-hospital all-cause mortality according to age filter had a lower case in-hospital mortality rate than those
and use of inferior vena cava filter (IVCF) in stable patients who did not, 6.4% compared with 15%.1 Stable patients
with a primary diagnosis of pulmonary embolism (PE).
who underwent pulmonary embolectomy also showed lower
With IVCF, r ¼ 0.801, P ¼ .02. With no IVCF, r ¼ 0.941,
in-hospital mortality with vena cava filters than without,
P ¼ .001.
14% compared with 36%.7
A strength of this investigation is the large number of
pulmonary embolism and chronic obstructive pulmonary dis- adults of all ages, both sexes, all races, and from all regions
ease was shown in patients aged >80 years, 9.1% mortality of the US. Results in those aged >80 years were consistent
with filters, compared with 14.4% without filters.10 in each category of patients with pulmonary embolism.
In-hospital all-cause mortality also was lower with vena A weakness is that results are based on administrative
cava filters in stable patients with pulmonary embolism and data, which are subject to inaccuracies of coding. We

Table 7 Mortality in Patients Older Than Aged 80 Years

IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
n n % (95% CI) n n % (95% CI) RR (95% CI)
Stable primary pulmonary embolism
Year
2003 3190 214 6.7 (5.8-7.5) 15,516 1174 7.6 (7.2-8.0) 0.9 (0.8-1.0)
2004 3576 271 7.6 (6.7-8.4) 17,481 1322 7.6 (7.2-7.9) 1.0 (0.9-1.1)
2005 4081 190 4.7 (4.0-5.3) 19,392 1260 6.5 (6.2-6.8) 0.7 (0.6-0.8)
2006 4323 178 4.1 (3.5-4.7) 21,496 1289 6.0 (5.7-6.3) 0.7 (0.6-0.8)
2007 4506 152 3.4 (2.8-3.9) 22,123 1167 5.3 (5.0-5.6) 0.7 (0.6-0.8)
2008 5217 212 4.1 (3.5-4.6) 23,464 1331 5.7 (5.4-6.0) 0.7 (0.6-0.8)
2009 4767 158 3.3 (2.8-3.8) 22,243 1228 5.5 (5.2-5.8) 0.6 (0.5-0.7)
2010 4868 203 4.2 (3.6-4.7) 23,499 1219 5.2 (4.9-5.5) 0.8 (0.7-0.9)
2011 4881 121 2.5 (2.0-2.9) 25,721 1142 4.4 (4.2-4.7) 0.6 (0.5-0.7)
2012 4115 100 2.4 (2.0-2.9) 23,975 1035 4.3 (4.1-4.6) 0.6 (0.5-0.7)
All years 43,524 1799 4.1 (3.9-4.3) 214,910 12,167 5.7 (5.6-5.8) 0.7 (0.7-0.8)
Stable primary pulmonary embolism and no comorbid conditions
Year
2003 970 48 4.9 (3.6-6.3) 5727 353 6.2 (5.5-6.8) 0.8 (0.6-1.1)
2004 1173 28 2.4 (1.5-3.3) 6086 275 4.5 (4.0-5.0) 0.5 (0.4-0.8)
2005 1124 43 3.8 (2.7-4.9) 6916 275 4.0 (3.5-4.4) 1.0 (0.7-1.3)
2006 1215 28 2.3 (1.5-3.1) 7388 343 4.6 (4.2-5.1) 0.5 (0.3-0.7)
2007 1184 9 0.8 (0.3-1.3) 6756 204 3.0 (2.6-3.4) 0.3 (0.1-0.5)
2008 1285 30 2.3 (1.5-3.2) 6896 201 2.9 (2.5-3.3) 0.8 (0.6-1.2)
2009 1117 11 1.0 (0.4-1.6) 6477 247 3.8 (3.3-4.3) 0.3 (0.1-0.5)
2010 1148 29 2.5 (1.6-3.4) 6603 203 3.1 (2.7-3.5) 0.8 (0.6-1.2)
2011 1152 6 0.5 (0.1-0.9) 7365 183 2.5 (2.1-2.8) 0.2 (0.1-0.5)
2012 905 10 1.1 (0.4-1.8) 6440 160 2.5 (2.1-2.9) 0.5 (0.2-0.9)
All years 11,273 242 2.1 (1.9-2.4) 66,654 2444 3.7 (3.5-3.8) 0.6 (0.5-0.7)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.
8 The American Journal of Medicine, Vol -, No -, - 2016

Table 8 Mortality in Patients with Stable Primary Pulmonary Embolism and No Comorbid Conditions According to Year, Age, and Inferior
Vena Cava Filters
IVCF IVCF Died IVCF Mortality No IVCF No IVCF Died No IVCF Mortality
n n % (95% CI) n n % (95% CI) RR (95% CI)
Year
2003 4851 134 2.8 (2.3-3.2) 47,008 922 2.0 (1.8-2.1) 1.4 (1.2-1.7)
2004 5332 88 1.7 (1.3-2.0) 48,367 689 1.4 (1.3-1.5) 1.2 (0.9-1.4)
2005 5545 117 2.1 (1.7-2.5) 52,280 750 1.4 (1.3-1.5) 1.5 (1.2-1.8)
2006 5622 54 1.0 (0.7-1.2) 57,053 794 1.4 (1.3-1.5) 0.7 (0.5-0.9)
2007 5033 33 0.7 (0.4-0.9) 54,345 493 0.9 (0.8-1.0) 0.7 (0.5-1.0)
2008 5487 75 1.4 (1.1-1.7) 58,373 524 0.9 (0.8-1.0) 1.5 (1.2-1.9)
2009 5365 59 1.1 (0.8-1.4) 59,082 528 0.9 (0.8-1.0) 1.2 (0.9-1.6)
2010 5149 55 1.1 (0.8-1.3) 61,998 487 0.8 (0.7-0.9) 1.4 (1.0-1.8)
2011 4912 25 0.5 (0.3-0.7) 64,648 479 0.7 (0.7-0.8) 0.7 (0.5-1.0)
2012 4335 35 0.8 (0.5-1.1) 61,020 355 0.6 (0.5-0.6) 1.4 (1.0-2.0)
All years 51,631 675 1.3 (1.2-1.4) 564,174 6021 1.1 (1.0-1.1) 1.2 (1.1-1.3)
Age, years
18-20 257 0 e 6179 11 0.2 (0.1-0.3) e
21-30 1720 13 0.8 (0.3-1.2) 45,075 143 0.3 (0.3-0.4) 2.4 (1.3-4.2)
31-40 3839 26 0.7 (0.4-0.9) 72,237 256 0.4 (0.3-0.4) 1.9 (1.3-2.8)
41-50 7334 49 0.7 (0.4-0.9) 102,099 394 0.4 (0.3-0.4) 1.7 (1.3-2.3)
51-60 8204 98 1.2 (1.0-1.4) 98,680 700 0.7 (0.7-0.8) 1.7 (1.4-2.1)
61-70 8928 54 0.6 (0.4-0.8) 89,212 778 0.9 (0.8-0.9) 0.7 (0.5-0.9)
71-80 10,076 193 1.9 (1.6-2.2) 84,038 1295 1.5 (1.5-1.6) 1.2 (1.1-1.4)
>80 11,273 242 2.1 (1.9-2.4) 66,654 2444 3.7 (3.5-3.8) 0.6 (0.5-0.7)
All ages 51,631 675 1.3 (1.2-1.4) 564,174 6021 1.1 (1.0-1.1) 1.2 (1.1-1.3)
CI ¼ confidence interval; IVCF ¼ inferior vena cava filter; RR ¼ relative risk.

assumed that any inaccuracies of coding would be similar
among patients with and without vena cava filters. In regard
to coding for pulmonary embolism, review and reabstraction
of a sample of Medicare hospitalizations showed that 92% of
codeable cases for pulmonary embolism were on the ab-
stract.12 The specificity of ICD-9-CM coding is high. It is
4 elderly, it would be necessary to speculate that the insertion
Primary Stable PE No Comorbid CondiƟons
No IVCF
3
Mortality (%)
remote that procedures coded in discharge abstracts were not
performed. Because reimbursement for procedures is linked
to the extent of coding, it is also likely that the sensitivity of
coding for the use of vena cava filters, thrombolytic therapy,
ventilator support, and embolectomy was high.
Another weakness is the possibility of “immortal time
bias” (survivor bias).13 Because lower in-hospital mortality in
stable patients was observed only in the elderly or very
of filters was delayed more in elderly patients than in younger
patients, or that the higher death rates among elderly patients
affected them more in the waiting period before filters were
inserted. We do not know, and cannot determine, the time
interval between hospital admission and filter insertion. It is

IVCF
2
clear that a prospective randomized controlled trial of elderly
and very elderly patients will not be done in the foreseeable
1 future. We are left with the choice of dismissing this and
comparable investigations because immortal time bias cannot
be excluded, or we can offer elderly patients the possibility of
- a lower mortality with vena cava filters, although we cannot
18-20 21-30 31-40 41-50 51-60 61-70 71-80 > 80
guarantee the absence of bias.
Age (Years)

Figure 5 In-hospital all-cause mortality according to age CONCLUSION

and use of inferior vena cava filter (IVCF) in stable patients
with a primary diagnosis of pulmonary embolism (PE) and
no comorbid conditions. With IVCF, r ¼ 0.849, P ¼ .01.
With no IVCF, r ¼ 0.809, P ¼ .02.
At present, it seems prudent to consider a vena cava filter in
very elderly (aged >80 years) stable patients with a primary
or secondary diagnosis of acute pulmonary embolism,
with or without comorbid conditions. Although future
Stein et al Vena Cava Filters in the Elderly
prospective study is warranted, it seems remote that a ran-
domized controlled trial of very elderly stable patients will
be performed in the foreseeable future.
References
1. Stein PD, Matta F, Keyes DC, et al. Impact of vena cava filters on
in-hospital case fatality rate from pulmonary embolism. Am J Med.
2012;125:478-484.
2. Mismetti P, Laporte S, Pellerin O, et al; PREPIC2 Study Group. Effect
of a retrievable inferior vena cava filter plus anticoagulation vs
anticoagulation alone on risk of recurrent pulmonary embolism: a
randomized clinical trial. JAMA. 2015;313:1627-1635.
3. Bikdeli B, Wang Y, Minges KE, et al. Vena caval filter utilization and
outcomes in pulmonary embolism: Medicare hospitalizations from
1999 to 2010. J Am Coll Cardiol. 2016;67:1027-1035.
4. Dalen JE, Stein PD. Is there a sub-group of patients with pulmonary
embolism who may benefit from an IVC filter? J Am Coll Cardiol.
2016;67:1036-1037.
5. Agency for Healthcare Research and Quality. HCUP Databases. Health-
care Cost and Utilization Project (HCUP). February 2016. Available at:
www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed July 23, 2016.
9
6. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of
classifying prognostic comorbidity in longitudinal studies: develop-
ment and validation. J Chronic Dis. 1987;40:373-383.
7. Stein PD, Matta F. Case fatality rate with pulmonary embolectomy for
acute pulmonary embolism. Am J Med. 2012;125:471-477.
8. Stein PD, Matta F, Alrifai A. Case fatality rate in pulmonary embolism
according to age and stability. Clin Appl Thromb Hemost. 2012;19:
668-672.
9. Kearon C, Akl EA, Ornelas J, et al. Antithrombotic therapy for VTE
disease: CHEST guideline and expert panel report. Chest. 2016;149:
315-352.
10. Stein PD, Matta F. Vena cava filters in hospitalized patients with
chronic obstructive pulmonary disease and pulmonary embolism.
Thromb Haemost. 2013;109:897-900.
11. Stein PD, Matta F, Sabra MJ. Case fatality rate with vena cava filters in
hospitalized stable patients with cancer and pulmonary embolism. Am J
Med. 2013;126:819-824.
12. Kniffin WD Jr, Baron JA, Barrett J, et al. The epidemiology of
diagnosed pulmonary embolism and deep venous thrombosis in the
elderly. Arch Intern Med. 1994;154:861-866.
13. Ho AM, Dion PW, Ng CS, Karmakar MK. Understanding immortal
time bias in observational cohort studies. Anaesthesia. 2013;68:
126-130.