Recurrent pulmonary embolism. A prospective study

DOI 10.1378/chest.95.5.976 1989;95;976-979Chest

M Monreal, J Ruiz, R Salvador, J Morera and A Arias study.Recurrent pulmonary embolism. A prospective

http://chestjournal.chestpubs.org/content/95/5/976

can be found online on the World Wide Web at: The online version of this article, along with updated information and services

) ISSN:0012-3692http://chestjournal.chestpubs.org/site/misc/reprints.xhtml(without the prior written permission of the copyright holder.reserved. No part of this article or PDF may be reproduced or distributedChest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights

ofbeen published monthly since 1935. Copyright1989by the American College is the official journal of the American College of Chest Physicians. It hasChest

© 1989 American College of Chest Physicians by guest on July 10, 2011chestjournal.chestpubs.orgDownloaded from


http://chestjournal.chestpubs.org/site/misc/reprints.xhtml

http://chestjournal.chestpubs.org/

4From the Hospital de Badalona Cermans Trias i Pujol, UniversidadAutonoma, Barcelona, Spain.

tProfessor of Medicine.lChief, Pneumology Service.Manuscript received May 31; revision accepted September 15.

976 Recurrent Pulmonary Embolism (Montrea! et a!)

Recurrent Pulmonary Embolism*A Prospective StudyManuel Monreal, M.D.;t Juan Ruiz, M.D.;t Rafael Salvador, M.D.;

Jose Morera, M.D.;� and Antonio Arias, M.D.

We have prospectively studied a series of 121 consecutive

patients with venous thromboembolism (38 with pulmonary

embolism, 83 with venous thrombosis of the lower extrem-ities) searching for recurrences of pulmonary embolismdespite adequate heparin therapy. A baseline ventilation-

perfusion lung scan was Obtained initially in every patient,

whether the original diagnosis was pulmonary embolism orvenous thrombosis. Repeat chest roentgenograms and lungscans were obtained routinely at eight days of heparin

treatment. The primary trial endpoints were a finding of aclinically apparent recurrent pulmonary embolism, orlaboratory evidence of subclinical pulmonary embolism.Eight items of clinical and laboratory information were

recorded at admission and then correlated with the lung

T he main objective in treating venous thromboem-

bolism with heparmn is to reduce the risk of further

pulmonary emboli. Despite adequate anticoagulant

treatment, recurrences occur.’� Can anticoagulation

failure be detected before these recurrences occur?

Usual laboratory methods to monitor anticoagulation

failure do not always reliably predict recurrences,9”#{176}

and then the problems arise from a lack of conclusive

evidence about which patients are at a higher risk of

recurrences. Several authors have suggested that the

finding ofa free-floating thrombus on venography may

represent a substantial risk ofembolism.” Others have

reported a higher anticoagulation failure rate in cancer

patients with venous �2 There have been a

number of prospective studies aimed at detecting

factors which lead to recurrence despite heparin

therapy, but most of them do not include a second

lung scan routinely performed to compare against the

baseline scan. Additionally, there have not been pro-

spective studies looking specifically at the role of

venography in predicting these recurrences.

We have prospectively studied a consecutive series

of patients with venous thromboembolism (both pul-

monary embolism and venous thrombosis), searching

for patients who develop recurrences of pulmonary

embolism despite heparin therapy. Furthermore, 5ev-

scan results. Recurrences were seen in seven of 38 patientswith an original diagnosis of pulmonary embolism, and in

five of 83 patients admitted because of venous thrombosis(p = 0.034). Recurrences were also more frequent in patients

with a free-floating thrombus on venography (p 0.014).

The risk of new defects in patients with venous thrombosisand without free-floating thrombus was 3.05 percent, ye-

nous thrombosis with free-floating thrombus, 13.33 percent;

patients with pulmonary embolism without free-floating

thrombus, 11.42 percent; and with free-floating thrombus,

38.67 percent. Venography seems thus mandatory in pa-tients with pulmonary embolism, as it recognizes a sub-

group of patients at a high risk of recurrences.

(Chest 1989; 95:976-79)

eral points ofclinical and laboratory information were

measured, and the risk of recurrences was assessed.

MATERIAL AND METHODS

Since January 1985, 145 consecutive patients in our hospital were

diagnosed as having acute venous thromboembolism. All had

objective tests to confirm the diagnosis: venous thrombosis was

confirmed with ascending x-ray venography in all patients; pulmo-

nary embolism with a high-probability ventilation and perfusion

lung scan and venographic documentation of venous thrombosis on

lower extremities. Only those patients having an adequate course of

heparmn therapy were included, and therefore, ten ofthese patients

were not included (clinically massive pulmonary embolism, four;

previous abnormalities on blood coagulation tests, two; intracerebralhemorrhage, one; digestive bleeding, one; hematoma, one; throm-

bocytopenia, one. Additionally, another 14 patients were excludedbecause of adverse effects of heparmn therapy: digestive bleeding,

seven; hematuria, four; thrombocytopenia, two; retroperitoneal

hematoma, one.A total of 121 patients entered into the study; they were 64 men

and 57 women, aged 21 to 86 years, mean 64. A total of 38 patientshad clinically apparent pulmonary embolism, and 83 patients had

deep venous thrombosis on lower limbs. Each patient received eightdays ofintravenous heparmn therapy. Then oral warfarin therapy wasbegun. A baseline lung scan was obtained initially in every patient,

whether the original diagnosis was pulmonary embolism or venous

thrombosis. Repeat chest roentgenograms and lung scans were

obtained routinely at eight days after heparmn was started, and atany other time that signs or symptoms indicated.

The primary trial endpoint was a finding of confirmed, clinically

apparent recurrent pulmonary embolism during the first eight days

of heparmn therapy. In addition, laboratory evidence of subclinical

pulmonary embolism was categorized as symptomless new perfu-

sion lung scan defects seen at the repeat investigation in segmentspreviously well perfused (according to Callus et al, methods’s).

All lung scans were obtained with a standard gamma camera.

Pulmonary perfusion imaging was performed using #{176}�“Tc-macroag-



CHEST/95/5/MAV.1989 977

gregated albumin injected intravenously with the patient supine.

Imaging included six standard projections (anterior, posterior, right

and left laterals, and right and left posterior obliques). Ventilation

imaging was performed with ‘�‘Xe. All images were obtained in the

posterior projection. The ventilation-perfusion studies were inter-

preted in accordance with the criteria of Biello et al,’� by two staff

physicians who had no knowledge of either the clinical history or

venogram findings. They examined the perfusion studies to deter-

mine whether there were single or multiple perfusion defects, and

they noted the sizes of the largest defects to anatomic divisions ofthe lungs. Then they compared the major perfusion defects with

the ventilation defects, and the lung scan obtained eight days after

heparin started with the baseline scan. When interpretation of the

two readers differed, the readers reviewed the case together and

came to a consensus opinion.Biplanar ascending conventional venograms allowed the classifi-

cation of iliofemoral thrombi as occlusive or free-floating when

associated with a 5 cm or greater proximal segment of thrombus

outlined by radiopaque contrast material.” Furthermore, the degree

of closeness of the aforesaid thrombus was determined by the

distance in centimeters from the thrombus proximal end to the

vena cava bifurcation into the iliac veins.

Eight points ofclinical and laboratory information were recorded

at admission and then compared with the lung scan results: age,

sex, cause of the thrombosis (spontaneous or postoperative), asso-

ciated malignancy, closeness of the thrombus to the vena cava, free-

floating or occlusive thrombus, heparin dosage, and clinical diag-

nosis at admission (pulmonary embolism or venous thrombosis). A

poor laboratory response to heparmn (measured by activated partial

thromboplastin time) was not included in the laboratory information

because of technical problems. Therefore, we include the total

heparin dosage. A comparison was made between patients with

recurrences of pulmonary embolism on the second scan and those

without new perfusion defects. The chi-square test was used to

evaluate the statistical significance of the relationship between twocategorical variables. The SPSSX program” was used for such

purpose.

Data were then analyzed by means of a logistic regression model

in order to identify factors which predicted risk of recurrences of

pulmonary embolism. We applied a stepwise selection procedure

to choose those variables that showed a significant statistical

association with the risk ofrecurrences. The chi-squared likelihoodratio test was used to evaluate whether or not a given factor

improved the prediction of the model. The BMDP LR program”was used to perform stepwise selection of variables. In order to

estimate more exact coefficients we used the Walke,’.Duncan

method.

RESULTS

Recurrences were seen in seven patients with an

original diagnosis of pulmonary embolism (seven out

of38 patients, 18.4 percent), in six ofwhom recurrent

pulmonary embolism was established by new, high-

probability changes on the lung scan. The other

patient, a 72-year-old woman, developed sudden dysp-

nea, chest pain and hypotension five days after admis-

sion. She died and findings of pulmonary embolism

were seen at autopsy. In three patients, recurrence of

pulmonary embolism was asymptomatic, discovered

on the routine lung scan obtained on the eighth day of

treatment.

In contrast, only five patients in whom the original

diagnosis was thrombosis had new defects (5 out of 83

patients, 6 percent). This rate of recurrence was

statistically lower than that in the group of patients

with pulmonary embolism (p = 0.034). Only one pa-

tient in this group had clinical symptoms of pulmonary

embolism, leaving the remaining four patients asymp-

tomatic.

A significant association was also found between the

finding on venogram of a free-floating thrombus and

recurrences ofpulmonary embolism. Seven out of 34

patients with such a finding developed recurrence,

while only five patients with an adherent thrombus

did (seven of 34, 20.58 percent vs five of 87, 5.74

percent; p = 0.0141). No other items correlate with

the development of recurrences of pulmonary embo-

lism (Table 1).

We made a multivariate analysis using the logistic

regression model. The stepwise selection procedure

selected as a first step the variable “free-floating

thrombus” as the most significant factor associated

with the risk of recurrences (x2 = 5.88, p = 0.02). In a

second step, the clinical diagnosis of pulmonary em-

bolism at admission was selected as the second varia-

ble most significantly associated with recurrence

(x2 = 4.70, p = 0.03). Table 2 shows the logistic regres-sion results on the two selected factors by the Walker-

Duncan method. The relative risk of recurrences for

free-floating thrombus is 4.9, and that for pulmonary

embolism diagnosis is 4. i (both measures adjusted for

the other regressor).

Finally, Table 3 shows the risk of recurrences for

each factor considered alone, and the risk for the

combination of both two variables. Patients admitted

because of venous thrombosis have a low risk of

developing perfusion defects on lung scan while

receiving heparmn therapy. They either have or do

not have a free-floating thrombus on venogram. Con-

trarily, venography is useful in patients admitted

because ofa pulmonary embolism, because the detec-

tion of a free-floating thrombus identifies the group of

Table 1-Statistical Sign�ficance ofthe CorrelationBetween Eight Variables and Recurrent Thromboembolism

Variables

No Recurrences

(n = 109)

Recurrences

(n = 12) p

Age, yr 63.80 66.25 0.579Sex:M 57 7

F 52 5 0.690

DVT cause: spontaneous 73 11

postoperative 36 1 0.078

Associated malignancy 20 4 0.216Closeness of thrombus,

cm 22 29 0.308

Free-floating thrombus 27 7 0.014Heparin dosage, IU/24 h 30520 32750 0.176

Diagnosis at admission:Venous thrombosis 78 5

Pulmonary embolism 31 7 0.034



Table 2-Logistic Regression: Risk ofRecurrent Pulmonary Embolism According to Presence ofFree-Flaating Thrombus andClinical Diagnosis at Admission

Variable Coefficient Standard Error Relative Risk* 95% Confidence Interval

Free-floating 1.58691 0.64873 4.9 1.4-17.4

thrombus

Clinical diagnosis 1.4113 0.64918 4.1 1.1-14.6

ofpulmonary embolism

�‘The risk of each category of patients was estimated based on the

following formula:lJ1+EXP(-[-3.4594+ 1.58691 [PE]+ 1.4113 (FF1’)])

where PE is a clinical diagnosis ofpulmonary embolism, and FFEthe presence on venography of a free-floating thrombus.

978 Recurrent Pulmonary Embolism (Montrea! et a!)

4Relative risk = EXP (logistic regression coefficient).

patients at a higher risk ofrecurrences despite heparmn

therapy.

DISCUSSION

Our data indicate that recurrences of pulmonary

embolism are relatively common in patients early in

the course of heparin therapy, and that patients at a

highest risk may be identified by venography.

Prevention of recurrences is a frequently stated aim

of anticoagulant therapy. The overall early mortality

rate in treated acute pulmonary embolism is 3 to 5

percent.3 In series which have used angiographic data,

the incidence of nonfatal recurrences in treated pa-

tients ranges between zero and 19 percent.’7 Some of

these recurrences could certainly be traced back to

inadequate therapy, but in others, the complication

occurred while laboratory parameters showed ade-

quate anticoagulation. Thus, it becomes evident that,

while most patients respond favorably to heparmn

therapy, a small number of them did not. The nonre-

sponders may require emergency alternative thera-

peutic actions because they definitely have an in-

creased risk of thromboembolic complications and

death if treated with heparmn alone. Many authorities

on the subject now admit that anticoagulation alone is

adequate treatment for nearly all patients, and that

venous interruption, preferably using a filter device,

should be reserved for patients with either definite

recurrence of pulmonary embolism or a high risk of

emboli combined with a contraindication to anticoag-

Table 3-Risk ofRecurrences ofPulmonary EmbolismAccording to Presence or Not ofa Free-Floating Thrombus

on Venography and Clinical Diagnosis at Admission5

Risk of Recurrences, %

Venous thrombosis without free- 3.05

floating thrombus

Venous thrombosis with free-floating 13.3

thrombus

Pulmonary embolism without free- 11.42

floating thrombus

Pulmonary embolism with free- 38.67

floating thrombus

Bleeding complications are seldom difficult to de-

tect, but the diagnosis of recurrent thromboembolism

may be more elusive. Many authors have demon-

strated that pulmonary embolism may be clinically

quite silent.�#{176}� The use of lung scanning for the

documentation of recurrence in this series, although a

more objective determination than has been used in

most previous reports, might be open to criticism

because of the lack of specificity of this test. It had

been suggested that new abnormalities may appear on

the lung scan in the absence of re�embo1ization.m

However, these artifactual changes have been well

documented in only a few cases. We used lung

scanning to confirm pulmonary embolism (and recur-

rences) but required a high-probability ventilation-

perfusion lung scan result. Such a result has a predic-

tive positive value of roughly 75 percent.tm Therefore,

when used properly, lung scanning is probably a

reliable method for detecting recurrent pulmonary

embolism.

Another possible criticism is the diagnosis of free-

floating thrombus. Since there is usually very little

evidence of the nature of the proximal end of thethrombus in the presence of an occluded vein, it is

necessary” to do retrograde contrast injections in

order to definitely rule out free-floating thrombi.

Therefore, we cannot be sure that the patients with

obstructed veins did not have a proximal free-floating

segment. Whatever the case, in our experience,

patients admitted because ofpulmonary embolism are

at a higher risk of recurrence (38 percent), if a free-

floating thrombus is shown by venography. Thus, we

recommend performing routine venography on all

patients presenting with pulmonary embolism. These

patients may require alternative therapeutic actions.

Once this information is available, a rational decision

regarding a choice between heparmn or interventional

therapy (vena cava filtering device), as claimed by

some authors,� would be made.

In contrast, the development of new perfusion

defects on lung scan in patients presenting with deep

venous thrombosis is lower, even in the presence of a

free-floating thrombus, If the logic is correct that a

free-floating tail represents a mechanical risk not well

controlled by anticoagulation, then it should not make

a difference whether the patient already has an em-



CHEST/95/5/MAY, 1989 979

bolism or not. We do not have an explanation for this

discrepancy. An increase in recurrence rate from 3 to

13 percent in patients with venous thrombosis and a

free-floating thrombus is a marked increase, but would

this increase in recurrence rate justify a more aggres-

sive therapeutic approach? Further studies are needed

to definitely confirm or rule out these findings. But at

present, our findings suggest that venography may

play a role in predicting those patients with venous

thromboembolism at highest risk to develop recur-

rences.

REFERENCES

1 Bynum U, Parkey RW, Wilson JE. Tests offibrin metabolism in

recurrent venous thromboembolism. Arch Intern Med 1977;

137:1385-89

2 Pollack EW. Early diagnosis of anticoagulant failure: can recur-

rent pulmonary embolism be prevented? mt Surg 1980; 65:219-22

3 Wilson JE, Bynum U, Parkey RW. Heparmn therapy in venous

thromboembolism. Am J Med 1981; 70:808-16

4 Menzolan JO, Sequeira JC, Doyle JE, Cantelmo NL, NowakM, Tracey K, et al. Therapeutic and clinical course ofdeep vein

thrombosis. Am J Surg 1983; 146:581-85

5 Hattersley PC, Mitsuoka C, King JH. Heparin therapy for

thromboembolic disorders: a prospective evaluation of 134 cases

monitored by the activated coagulation time. JAMA 1983;

250:1413-16

6 Doyle DJ, Turpie ACC, Hirsh J, Best C, Kinch D, Levine MN,

et al. Adjusted subcutaneous heparin or continuous intravenousheparin in patients with acute deep vein thrombosis: a random-

ized trial. Ann Intern Med 1987; 107:441-45

7 Holm HA, Finnanger B, Hartmann A, Laerum F, Lohren 0,Ruud TE, et al. Heparmn treatment of deep venous thrombosis

in 280 patients: symptoms related to dosage. Acta Med Scand

1984; 215:46-53

8 Rosiello BA, Chan CK, Tencza F, Matthay RA. Timing of oral

anticoagulant therapy in the treatment of angiographically

proven acute pulmonary embolism. Arch Intern Med 1987;

147:1469-73

9 Salzman EW, Deykin D, Shapiro RM, Rosenberg R. Manage-

ment ofheparin therapy: controlled prospective trial. N Engl JMed 1975; 1046-50

10 Wheeler AP, Jacquiss RDB, Newman JH. Physician practices in

the treatment of pulmonary embolism and deep venous throm-bosis. Arch Intern Med 1988; 148:1321-25

11 Scott Norris C, Creenfield U, Herrman JB. Free-floating

iliofemoral thrombus: a risk ofpulmonary embolism. Arch Surg1985; 12f�806-08

12 Oster MW. Thrombophlebitis and cancer: a review. Angiology

1976; 27:587-67

13 Callus A, Jackaman J, Tillett J, Mills W, Wycherley A. Safety

and efficacy of warfarin started early after submassive venousthrombosis or pulmonary embolism. Lancet 1986; 11:1293-96

14 Biello DR. Mattar AC, McKnight RC, Siegel BA. Ventilation-

perfusion studies in suspected pulmonary embolism. AJR 1979;

133:1033-3715 Norusis MJ. SPSSX user’s guide. New York: McGraw Hill Book

Company, 1983

16 Engelman L, Frane JW, Jennrich RI. Biomedical computer

programs P�series. Berkeley: University ofCalifornia Press, 1982

17 Egermayer P. Value of anticoagulants in the treatment of

pulmonary embolism: a discussion paper. J Roy Soc Med 1981;

74:675-81

18 Creenfield U. Current indications for and results of Creenfield

filter placement. J Vasc Surg 1984; 1:502-04

19 Dietzman RH, NicoloffDM, Varco RL. Treatment of pulmonaryemboli: vena caval interruption or anticoagulation? Geriatrics

1975; 64-6920 Kisther RL, Ball JJ, Nordyke RA, Freeman CC. Incidence of

pulmonary embolism in the course of thrombophlebitis of the

lower extremities. Am J Surg 1972; 124:169-76

21 Moser KM. LeMoineJR. Is embolic risk conditioned by location

ofdeep venous thrombosis? Ann Intern Med 1981; 94:439-44

22 Williams J\� Eikman EA, Greenberg S. Asymptomatic pulmo-

nary embolism: a common event in high risk patients. Ann Surg1982; 193:323-27

23 Monreal M, Moreno V, Martorell A, Lafoz E, Foz M, Ruiz J,et al. Predicting pulmonary embolism in postoperative patients

with deep venous thrombosis of lower limbs. Ann Vasc Surg

1987; 1:421-25

24 Monreal M, Rey-Joly C, Ruiz J, Salvador J, Lafoz E, Viver E.Asymptomatic pulmonary embolism in patients with deep vein

thrombosis: Is it useful to take a lung scan to rule out this

condition? J Cardiovasc Surg (in press)25 Moser KM, Longe AM, Ashburn WL, Cuisan M. Suprious

scintiphotographic recurrence ofpulmonary emboli. Am J Med1973; 55:434-38

26 McNeill BJ. Ventilation-perfusion studies in the diagnosis ofpulmonary embolism: concise communication. J Nuclr Med

1980; 21:319-23

27 Savelyer VS, Yablokov EC, Kirienko AJ. Surgical prevention ofpulmonary embolism. World J Surg 1980; 4:709-15



DOI 10.1378/chest.95.5.976 1989;95; 976-979Chest

M Monreal, J Ruiz, R Salvador, J Morera and A AriasRecurrent pulmonary embolism. A prospective study.

July 10, 2011This information is current as of

http://chestjournal.chestpubs.org/content/95/5/976Updated Information and services can be found at:

Updated Information & Services

http://chestjournal.chestpubs.org/content/95/5/976#related-urlsThis article has been cited by 9 HighWire-hosted articles:

Cited Bys

http://www.chestpubs.org/site/misc/reprints.xhtmlonline at: Information about reproducing this article in parts (figures, tables) or in its entirety can be foundPermissions & Licensing

http://www.chestpubs.org/site/misc/reprints.xhtmlInformation about ordering reprints can be found online:

Reprints

the right of the online article.Receive free e-mail alerts when new articles cite this article. To sign up, select the "Services" link to

Citation Alerts

slide format. See any online figure for directions. articles can be downloaded for teaching purposes in PowerPointCHESTFigures that appear in Images in PowerPoint format



http://chestjournal.chestpubs.org/content/95/5/976#related-urls

http://www.chestpubs.org/site/misc/reprints.xhtml

http://www.chestpubs.org/site/misc/reprints.xhtml


Documents

Recurrent pulmonary embolism. A prospective study