Annals of Internal Medicine
Residual Thrombosis on Ultrasonography to Guide the Duration of
Anticoagulation in Patients With Deep Venous Thrombosis
A Randomized Trial
Paolo Prandoni, MD, PhD; Martin H. Prins, MD, PhD; Anthonie W.A. Lensing, MD, PhD; Angelo Ghirarduzzi, MD; Walter Ageno, MD;
Davide Imberti, MD; Gianluigi Scannapieco, MD; Giovanni B. Ambrosio, MD; Raffaele Pesavento, MD; Stefano Cuppini, MD;
Roberto Quintavalla, MD; and Giancarlo Agnelli, MD, for the AESOPUS Investigators*
Background: The optimal duration of oral anticoagulant therapy in
patients with deep venous thrombosis (DVT) of the lower extrem-
ities remains uncertain.
Objective: To assess whether tailoring the duration of anticoagu-
lation on the basis of the persistence of residual thrombi on ultra-
sonography reduces the rate of recurrent venous thromboembolism
(VTE) compared with the administration of conventional fixed-
duration treatment in adults with proximal DVT.
Design: Parallel, randomized trial from 1999 to 2006. Trained
physicians who assessed outcomes were blinded to patient assign-
ment status, but patients and providers were not.
Setting: 9 university or hospital centers in Italy.
Patients: 538 consecutive outpatients with a first episode of acute
proximal DVT at completion of an uneventful 3-month period of
anticoagulation.
Intervention: Patients were randomly assigned (stratified by center
and secondary vs. unprovoked DVT by using a computer-
generated list that was accessible only to a trial nurse) to fixed-
duration anticoagulation (no further anticoagulation for secondary
thrombosis and an extra 3 months for unprovoked thrombosis)
or flexible-duration, ultrasonography-guided anticoagulation (no
further anticoagulation in patients with recanalized veins and con-
tinued anticoagulation in all other patients for up to 9 months for
secondary DVT and up to 21 months for unprovoked thrombosis).
For the primary outcome assessment, 530 patients completed the
trial.
T he optimal duration of oral anticoagulant therapy after
an episode of deep venous thrombosis (DVT) of the
lower extremities is uncertain (1). Although patients with
DVT have a substantial risk for recurrent thromboembo-
lism after discontinuing anticoagulation (2, 3) and prolon-
gation of treatment has been demonstrated to be effective,
indefinite anticoagulation is not widely implemented be-
cause the risk for bleeding is perceived to outweigh the
benefit in most patients (4). Indefinite anticoagulation is
commonly reserved for patients with permanent risk factors
for thrombosis, whereas patients with unprovoked DVT gen-
erally receive 6 months of anticoagulation and those with
DVT associated with transient risk factors receive treatment
for 3 months (5, 6). Despite this adaptation of treatment,
recurrent thromboembolism after discontinuation of therapy
still occurs (7), and outcomes can probably be further im-
proved by using individual prognostic information.
Recent studies suggest that the presence of residual
thrombosis on ultrasonography when warfarin therapy is
Article
Measurements: The rate of confirmed recurrent VTE during 33
months of follow-up.
Results: Overall, 46 (17.2%) of 268 patients allocated to fixed-
duration anticoagulation and 32 (11.9%) of 270 patients allocated
to flexible-duration anticoagulation developed recurrent VTE (ad-
justed hazard ratio [HR], 0.64 [95% CI, 0.39 to 0.99]). For patients
with unprovoked DVT, the adjusted HR was 0.61 (CI, 0.36 to 1.02)
and 0.81 (CI, 0.32 to 2.06) for those with secondary DVT. Major
bleeding occurred in 2 (0.7%) patients in the fixed-duration group
and 4 (1.5%) patients in the flexible-duration group (P ⫽ 0.67).
Limitations: The trial lacked a double-blind design. The sample size
was not powered to detect differences in bleeding between groups
and to detect effectiveness of the intervention in the subgroups of
patients with unprovoked and secondary DVT. Patients with previ-
ous thromboembolism, permanent risk factors for thrombosis, and
thrombophilic abnormalities other than factor V Leiden and pro-
thrombin mutation were excluded.
Conclusion: Tailoring the duration of anticoagulation on the basis
of ultrasonography findings reduces the rate of recurrent VTE in
adults with proximal DVT.
Primary Funding Source: None.
Ann Intern Med. 2009;150:577-585. www.annals.org
For author affiliations, see end of text.
ClinicalTrials.gov registration number: NCT00380120.
* For additional AESOPUS investigators, see the Appendix (available at
www.annals.org).
discontinued is associated with a substantial increase in the
risk for subsequent recurrent thromboembolism (8 –10).
Therefore, assessment of residual thrombosis may play an
important role in tailoring the duration of anticoagulation.
We did a randomized study to test the value of tailor-
ing the duration of anticoagulation according to recanali-
zation or persistence of residual thrombosis. We randomly
See also:
Print
Editors’ Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 578
Editorial comment. . . . . . . . . . . . . . . . . . . . . . . . . . 644
Summary for Patients. . . . . . . . . . . . . . . . . . . . . . . I-44
Web-Only
Appendix
Conversion of graphics into slides
© 2009 American College of Physicians 577
Article Residual Thrombosis and Duration of Anticoagulation
Context
Optimal anticoagulation strategies for deep venous throm-
bosis (DVT) are evolving.
Contribution
The investigators randomly assigned 538 patients with
proximal DVT who had completed 3 months of anti-
coagulation to fixed-duration or flexible-duration,
ultrasonography-guided continued therapy. Patients with
unprovoked DVT in the first group received 3 additional
months of warfarin. The second group had no further
therapy if ultrasonography showed recanalized veins and
continued to receive anticoagulation for 9 to 21 months if
repeated ultrasonography showed persistent thrombi. Dur-
ing follow-up, 17% and 12% of the fixed-duration and
ultrasonography-guided groups had recurrent thrombo-
embolism, whereas 0.7% and 1.5% had major bleeding
events.
Implication
Ultrasonography findings may help tailor the duration of
anticoagulant therapy for DVT.
—The Editors
assigned patients with a first episode of proximal DVT
who completed 3 months of anticoagulation to receive ei-
ther fixed-duration anticoagulation or flexible-duration,
ultrasonography-guided anticoagulation. All patients had
long-term follow-up.
METHODS
Design
Our study was a randomized, multicenter, open-label
trial, with independent and blinded assessment of study
outcomes. We designed this study to assess whether tailor-
ing the duration of anticoagulation on the basis of the
ultrasonographic persistence of residual thrombi in patients
with proximal DVT reduces the rate of recurrent venous
thromboembolism (VTE) over 3 years of follow-up com-
pared with the administration of conventional fixed-
duration anticoagulation.
Trial recruitment started in January 1999 and ended
in July 2003. We ended the study when the last recruited
patient reached 3-year follow-up (July 2006). The institu-
tional review board of each participating center approved
the study protocol.
Setting and Participants
We conducted the study at 9 university or hospital
centers in Italy with particular skills in the long-term man-
agement of VTE disorders.
Consecutive patients with symptomatic proximal
DVT who had completed 3 months of anticoagulation
were eligible. We excluded patients who met any of the
following criteria: previous thromboembolism, known can-
578 5 May 2009 Annals of Internal Medicine Volume 150 • Number 9
cer, immobilization resulting from chronic irreversible
medical diseases (such as previous stroke or other neuro-
logic diseases, severe congestive heart failure, rheumatic
diseases, advanced atherosclerosis, or dementia), need for
indefinite anticoagulation for atrial fibrillation or valve re-
placement, inability to attend follow-up visits, carriage
of deficiencies of natural anticoagulants or lupus-like anti-
coagulants (including participants with anticardiolipin
antibody titers ⬎15 UPL/mL) based on assessment done
before anticoagulation (11), short (⬍3 months) life expect-
ancy, pregnancy, and age younger than 18 years. In addi-
tion, we excluded patients who developed recurrent throm-
boembolism, major bleeding, or cancer during the first 3
months of anticoagulation and those who declined to par-
ticipate. One of the trial physicians assessed the eligibility
criteria. We asked patients who met the inclusion criteria
to participate in the study after they received detailed writ-
ten information about study hypotheses and procedures.
We asked study patients for written informed consent to
participate. In addition, we discharged these patients with
a letter for their family physician indicating that they had
accepted to participate in a randomized study assessing the
value of tailoring the duration of anticoagulation according
to the persistence of residual thrombi.
We considered patients with thrombosis occurring in
association with recent (⬍3 months) trauma, surgical in-
tervention or puerperium, prolonged (⬎7 days) immobili-
zation from any cause or use of hormonal treatment as
having secondary DVT. We defined all other patients as
having unprovoked DVT.
Randomization and Intervention
We stratified randomization for secondary versus un-
provoked DVT and for center. We did this through a
central computer-generated list, which was accessible only
to a trial nurse who informed study physicians on treat-
ment allocation once patients had provided informed
consent. We allocated each patient (n ⫽ 100) in balanced
blocks of 10.
After patients completed an uneventful 3 months of
oral anticoagulant therapy, we randomly assigned them to
fixed or flexible durations of warfarin therapy. In the fixed-
duration group, patients with unprovoked DVT received 3
additional months of treatment (for a total treatment du-
ration of 6 months) and patients with secondary DVT
discontinued treatment (for a total treatment duration of
3 months). In the flexible-duration group, we tailored the
length of anticoagulation according to the ultrasono-
graphic persistence of residual thrombi. If veins had re-
canalized, we discontinued anticoagulation, along with fur-
ther ultrasonography. If veins had not recanalized, we
invited patients to have further ultrasonography after 3 and
9 months in patients with secondary DVT and after 3, 9,
15, and 21 months in those with unprovoked DVT. An-
ticoagulation was discontinued when the veins had recana-
lized, along with further ultrasonography. We adjusted
www.annals.org

warfarin therapy to a target international normalized ratio
of 2.0 to 3.0 in both groups.
Outcomes and Measurements
We followed patients for 33 months to document the
incidence of symptomatic recurrent thromboembolism. In
addition, we recorded major bleeding events occurring
during anticoagulation and 1 additional month.
Recurrent thromboembolism was diagnosed by com-
pression ultrasonography, ventilation–perfusion lung scan-
ning, or helical tomography, as appropriate (7, 8). If recur-
rent thrombosis was suspected in a previously unaffected
leg, the sole diagnostic criterion was incompressibility of a
proximal vein. Ultrasonographic criteria for recurrent ipsi-
lateral thrombosis were incompressibility of a proximal
vein segment initially free of thrombi or incompressibility
of a proximal vein that had completely recanalized. We
defined nonfatal pulmonary embolism by a segmental or
subsegmental ventilation–perfusion mismatch on lung
scanning or an intraluminal filling defect on spiral com-
puted tomography of the chest. We diagnosed fatal pulmo-
nary embolism if it was confirmed at autopsy, was pre-
ceded immediately before death by objectively confirmed
pulmonary embolism or venous thrombosis, or was a sud-
den death that could not be explained by a disease or con-
dition other than pulmonary embolism.
We defined bleeding as major if it was clinically overt
and associated with a decrease in hemoglobin levels of at
least 20 g/L or transfusion of at least 2 units of red blood
cells, or if it was retroperitoneal or intracranial.
An independent adjudication committee, whose mem-
bers were unaware of treatment assignment, reviewed all
outcome events.
Follow-up Procedures and Monitoring
We scheduled a follow-up visit at 3, 9, 15, 21, and 33
months after randomization for all patients. At each
follow-up visit, we examined patients and inquired about
health status, clinical symptoms or signs of recurrent VTE,
bleeding, postthrombotic manifestations, adherence to
treatment, and concomitant analgesic or anti-inflammatory
therapy (including aspirin). We scheduled ultrasonography
up to a maximum of 21 months for all study patients
(including those randomly assigned to fixed-duration anti-
coagulation) whose veins had not recanalized.
Because we lacked widely accepted criteria for the def-
inition of vein recanalization, all investigators agreed a pri-
ori (during a consensus meeting held before the beginning
of the study) on standard procedures for study purposes.
Vein diameters were measured during maximal compres-
sion and were considered recanalized if the diameter was
less than 2.0 mm in a single determination or less than 3.0
mm in 2 consecutive determinations. This procedure is
based on findings obtained in separate prospective cohorts
of patients with proximal DVT from the coordinating cen-
ter in Padua, Italy, and were found to be highly reproduc-
ible, fast, and easy to obtain (8, 12). One or more trained
www.annals.org
Residual Thrombosis and Duration of Anticoagulation Article
physicians who agreed on the study procedures and were
unaware of clinical details and of previous ultrasonographic
findings assessed the common femoral and popliteal veins
(transverse plane) according to a standardized procedure.
Asymptomatic variations of the vein diameter recorded at
the scheduled follow-up visits did not qualify as a study
outcome.
Patients received a card with the telephone numbers of
the thrombosis centers and were instructed to return to the
study center if clinical manifestations suggestive of recur-
rent venous thrombosis in either leg (edema, redness, ten-
derness, pain, or swelling), or pulmonary embolism (dys-
pnea, chest pain, or tachycardia) occurred. If this was the
case, we invited patients to the study center for further
diagnostic procedures. In addition, we instructed patients
to report to the study center on an emergency basis if any
bleeding occurred. Physicians involved in the study re-
corded all information potentially useful for adjudication
from the central independent committee.
We managed clinical and ultrasonographic data lo-
cally, and they were periodically transmitted through an
electronic system to the coordinating center in Padua
(which took care of data quality). We requested further
information whenever necessary.
Statistical Analysis
On the basis of the assumption that the overall 33-
month rate of recurrent thromboembolism would be about
20% in the fixed-duration group, approximately 280 pa-
tients would be required in each group to give a power of
0.90 and a type I error of 0.05 for the detection of 50%
risk reduction with flexible-duration anticoagulation.
We conducted our analysis on an intention-to-treat
basis. We calculated Kaplan–Meier estimates and their
95% CIs. To assess the risk for recurrent thromboembo-
lism in patients allocated to flexible-duration compared
with fixed-duration anticoagulation, we used Cox propor-
tional hazards modeling to calculate hazard ratios (HRs)
and their 95% CIs after adjustment for age, sex, thrombo-
philic status, and the presence of clinical symptoms of pul-
monary embolism at the time of the patient’s referral and
after stratification for secondary versus unprovoked DVT.
In addition, we calculated HRs for these variables to iden-
tify factors that were independently associated with an in-
creased risk for recurrent thromboembolism. We checked
the assumptions of the Cox proportional hazards model by
visual inspection of the log–log survivor function-by-time
curve.
We used the chi-square test with Yates correction to
compare the rate of persistent residual venous thrombosis
at the last ultrasonography assessment between the 2
groups. We did all calculations with SPSS, version 15.0
(SPSS, Chicago, Illinois).
Role of the Funding Source
This study had no external source of funding.
5 May 2009 Annals of Internal Medicine Volume 150 • Number 9 579
Article Residual Thrombosis and Duration of Anticoagulation

RESULTS (n ⫽ 4), major bleeding (n ⫽ 3), or development of cancer

Patients
From January 1999 to June 2003, we evaluated 1020
potentially eligible patients with symptomatic proximal
DVT treated for 3 months. Of these patients, 466 were
excluded because of previous thromboembolism (n ⫽
152), known cancer (n ⫽ 112), immobilization from
chronic medical diseases (n ⫽ 88), need for indefinite an-
ticoagulation (n ⫽ 45), inability to attend follow-up visits
(n ⫽ 35), deficiency of natural anticoagulants or carriage of
lupus-like anticoagulants (n ⫽ 18), short life expectancy
(n ⫽ 10), pregnancy (n ⫽ 5), or age younger than 18 years
(n ⫽ 1). Of the remaining 554 patients, an additional 16
were excluded because of recurrent thromboembolism
(n ⫽ 3) during the first 3 months of anticoagulation, and 6
declined informed consent. Therefore, we randomly as-
signed 538 patients to fixed-duration (n ⫽ 268) or flexible-
duration (n ⫽ 270) anticoagulation. Figure 1 shows the
study flow, and Table 1 shows the main characteristics of
the study patients.
Fixed-Duration Treatment Group
Of the 268 patients allocated to the fixed-duration
group, 1 with unprovoked DVT had anticoagulation pre-
maturely stopped; 2 with secondary DVT had prolonged
treatment beyond the mandated protocol duration for 3

Figure 1. Study flow diagram.

Acute proximal DVT

(n = 1020)

Excluded (n = 466)
Previous VTE: 152
Active cancer: 112
Chronic medical diseases: 88
Need for indefinite anticoagulant
therapy: 45
Inaccessibility for visits: 35
Thrombophilia: 18
Short life expectancy: 10
Pregnancy or young age: 6

Eligible (n = 554)

Excluded (n = 16)
VTE in the first 3 mo: 4
Bleeding event in the first 3 mo: 3
Cancer diagnosis in the first 3 mo: 3
Declined to consent: 6

Randomly assigned
(n = 538)

Flexible OAT Fixed OAT

(n = 270) (n = 268)

Idiopathic DVT Secondary DVT Idiopathic DVT Secondary DVT

(n = 155) (n = 115) (n = 151) (n = 117)

Complete follow-up Complete follow-up Complete follow-up Complete follow-up

(n = 153) (n = 113) (n = 149) (n = 115)
Recurrent VTE (n = 24) Recurrent VTE (n = 8) Recurrent VTE (n = 36) Recurrent VTE (n = 10)
Bleeding event (n = 2) Bleeding event (n = 2) Bleeding event (n = 2) Bleeding event (n = 0)

DVT ⫽ deep venous thrombosis; OAT ⫽ oral anticoagulant therapy; VTE ⫽ venous thromboembolism.
580 5 May 2009 Annals of Internal Medicine Volume 150 • Number 9 www.annals.org
 Residual Thrombosis and Duration of Anticoagulation Article

Table 1. Baseline Characteristics of Study Patients

Characteristic Patients With Unprovoked DVT* Patients With Secondary DVT*

Flexible OAT Fixed OAT Flexible OAT Fixed OAT

(n ⴝ 155) (n ⴝ151) (n ⴝ 115) (n ⴝ117)
Median age (range), y 70 (20–90) 69 (21–89) 60 (18–91) 58 (23–92)

Men, n (%) 93 (60.0) 87 (57.6) 42 (36.5) 48 (41.0)

PE symptoms at referral, n (%) 17 (11.0) 20 (13.2) 11 (9.6) 14 (12.0)

DVT localization, n (%)

Popliteal 85 (54.8) 83 (55.0) 61 (53.0) 60 (51.3)
Common femoral 9 (5.8) 23 (15.2) 11 (9.6) 14 (12.0)
Both locations 61 (39.4) 45 (29.8) 43 (37.4) 43 (36.8)

Risk factors, n (%)

Recent trauma – – 41 (35.7) 35 (29.9)
Recent surgery – – 33 (28.7) 36 (30.8)
Hormonal treatment – – 22 (19.1) 22 (18.8)
Puerperium – – 1 (0.9) 3 (2.6)
Medical diseases – – 11 (9.6) 10 (8.5)
Combined factors – – 7 (6.1) 11 (9.4)

FVL or PTM, n (%)

Underwent determination of status 150 (96.8) 146 (96.7) 110 (95.7) 115 (98.3)
Carriers 34 (22.7) 28 (19.2) 25 (22.7) 27 (23.5)
FVL† 23 17 19 12
PTM‡ 6 10 4 9
Both mutations 5 1 2 6

Duration of anticoagulation after randomization, n (%)

No further therapy 34 (21.9) 1 (0.7) 38 (33.0) 115 (98.3)
ⱕ3 mo 49 (31.6) 150 (99.3) 39 (33.9) 2 (1.7)
ⱕ9 mo 27 (17.4) 0 38 (33.0) 0
ⱕ15 mo 13 (8.4) 0 0 0
ⱕ21 mo 32 (20.6) 0 0 0

DVT ⫽ deep venous thrombosis; FVL ⫽ factor V Leiden; OAT ⫽ oral anticoagulant therapy; PE ⫽ pulmonary embolism; PTM ⫽ prothrombin G20210A mutation.
* None of the observed differences between groups were statistically significant.
† Homozygous in 1 patient per group.
‡ Homozygous in 1 patient with secondary DVT who was randomly assigned to flexible-duration anticoagulation.

additional months (after indication from their general
practitioner); and 7 resumed anticoagulation during follow-
up because of development of atrial fibrillation (n ⫽ 5),
the Budd–Chiari syndrome (n ⫽ 1), and cardiomyopathy
(n ⫽ 1). Follow-up was complete in 263 (98.1%) patients.
Flexible-Duration Treatment Group
Of the 270 patients allocated to flexible anticoagula-
tion, 7 discontinued treatment before vein recanalization
occurred and 6 continued anticoagulation despite vein re-
canalization (on indication from their general practitioner).
In an additional patient, we did not interrupt anticoagula-
tion because of development of cancer. We resumed anti-
coagulation during follow-up in 8 patients because of de-
velopment of atrial fibrillation (n ⫽ 7) or cancer (n ⫽ 1).
Follow-up was complete in 267 (98.5%) patients.
Compared with the fixed-duration group, anticoagu-
lation was prolonged in the flexible-duration group for a
mean of 3.7 months in patients with secondary DVT and
a mean of 5 months in those with unprovoked DVT.
www.annals.org
Recurrent Venous Thromboembolism
During follow-up, 78 (14.5%) of 538 patients devel-
oped recurrent thromboembolism, which was fatal in 9
(case-fatality rate, 11.5% [mostly accounted for by 6 sud-
den, otherwise unexpected deaths]). Two episodes (1 in
each group) occurred spontaneously during warfarin ther-
apy while patients were receiving the correct anticoagula-
tion. Of the 78 events, 46 occurred in the 268 (17.2%)
patients allocated to the fixed-duration group and 32 oc-
curred in the 270 (11.9%) allocated to the flexible-
duration group (adjusted HR, 0.64 [95% CI, 0.39 to
0.99]). Among patients with unprovoked DVT and those
with secondary DVT, the adjusted HR was 0.61 (CI, 0.36
to 1.02) and 0.81 (CI, 0.32 to 2.06), respectively (P for
interaction ⫽ 0.70). Patients randomly assigned to fixed-
duration anticoagulation had a 0.8% (CI, 0% to 1.8%)
cumulative incidence of recurrent thromboembolism after
3 months, 7.9% (CI, 4.7% to 11.2%) after 9 months,
11.0% (CI, 7.2% to 14.8%) after 15 months, 13.0% (CI,
8.9% to 17.0%) after 21 months, and 17.8% (CI, 13.1%
to 22.5%) after 33 months. The corresponding figures for
5 May 2009 Annals of Internal Medicine Volume 150 • Number 9 581
Article Residual Thrombosis and Duration of Anticoagulation
Table 2. Characteristics of Recurrent Thromboembolism
Characteristic Patients With Unprovoked DVT
Flexible OAT (n ⴝ 24) Fixed OAT (n ⴝ 36)
Type of recurrence, n
Ipsilateral DVT 6 17
Contralateral DVT 10 12
Nonfatal PE 4 4
Fatal PE 4 3
Relation to vein status, n/n (%)
Residual thrombosis – 16/50 (32.0)
Early recanalization – 20/101 (19.8)
DVT ⫽ deep venous thrombosis; OAT ⫽ oral anticoagulant therapy; PE ⫽ pulmonary embolism.
patients randomly assigned to flexible-duration anticoagu-
lation were 0.4% (CI, 0% to 1.1%), 4.1% (CI, 1.7% to
6.5%), 6.4% (CI, 3.4% to 9.3%), 7.5% (CI, 4.4% to
10.7%), and 12.3% (CI, 8.3% to 16.3%), respectively.
Table 2 shows results specified for patients presenting with
secondary and unprovoked DVT. Figure 2 shows the dis-
tribution of events over time.
Bleeding Complications
During anticoagulation, major bleeding occurred in 2
(0.7%) patients with unprovoked DVT who were ran-
domly assigned to fixed-duration therapy (1 with gastro-
intestinal and 1 with cerebral bleeding) and in 4 (1.5%)
patients with secondary DVT (n ⫽ 2) and unprovoked
thrombosis (n ⫽ 2) allocated to flexible-duration therapy
(3 with gastrointestinal and 1 with urinary tract bleeding)
(difference, 0.8 percentage points [CI, ⫺1.0 to 2.5 per-
centage points; P ⫽ 0.68]). All 6 patients developed bleed-
ing complications while receiving correct anticoagulation,
and no complication was fatal.
Death
Twenty-eight patients died during follow-up. In the
fixed-duration group, 11 patients died of pulmonary em-
bolism (n ⫽ 4), cancer (n ⫽ 3), myocardial infarction
(n ⫽ 1), and other causes (n ⫽ 3). In the flexible-duration
group, 17 patients died of pulmonary embolism (n ⫽ 5),
cancer (n ⫽ 5), stroke (n ⫽ 2), myocardial infarction (n ⫽
1), or other causes (n ⫽ 4). This difference was not statis-
tically significant (P ⫽ 0.33).
Additional Observations
In the multivariable analysis adjusted for randomized
allocation, factors that were independently associated with
a significantly increased risk for recurrent thromboembo-
lism were thrombophilia (HR, 2.75 [CI, 1.71 to 4.43])
and unprovoked presentation (HR, 2.24 [CI, 1.29 to
3.89]), whereas clinical symptoms suggestive of pulmonary
embolism at patients’ referral (HR, 1.50 [CI, 0.84 to
2.77]) and male sex (HR, 1.24 [CI, 0.78 to 1.96]) were
not. Factor V Leiden alone or associated with prothrombin
mutation was predictive of recurrent thromboembolism,
whereas prothrombin mutation alone was not.
582 5 May 2009 Annals of Internal Medicine Volume 150 • Number 9
Patients With Secondary DVT
Flexible OAT (n ⴝ 8) Fixed OAT (n ⴝ 10)
4 3
3 4
0 2
1 1
– 3/29 (10.3)
– 7/88 (8.0)
Persistent residual venous thrombosis at the last ultra-
sonography was shown in 79 (29.5%) of 268 patients in
the fixed-duration group (50 of 151 patients with unpro-
voked DVT and 29 of 117 patients with secondary DVT)
compared with 54 (20.0%) of 270 patients in the flexible-
duration group (34 of 155 with unprovoked DVT and 20
of 115 patients with secondary DVT) (P ⫽ 0.02).
Patient adherence to the study protocol was high. A
total of 93% of patients in the flexible-duration group
completed the scheduled ultrasonography assessments
compared with 91% of patients in the fixed-duration
group. The adequacy of ultrasonographic data was fully
satisfactory in all cases.
During the study, 15 patients in the flexible-duration
group and 19 patients in the fixed-duration group used
analgesic or anti-inflammatory drugs for variable periods
(ranging from a few days to 3 weeks).
DISCUSSION
One of the most pressing unresolved contemporary
issues in the management of VTE is improving risk strat-
ification to determine which patients with unprovoked
DVT can safely discontinue anticoagulation. These pa-
tients have a particularly high risk for VTE recurrence (5,
6). Less pressing, but still relevant, is the issue of whether
some patients with secondary DVT might benefit from
longer courses of anticoagulant therapy than the currently
recommended 3 months (7, 13).
Several years ago, our group demonstrated that the
persistence of residual vein thrombi, as shown by repeated
ultrasonography over time, is a powerful and independent
risk factor of recurrent VTE (8). After this analysis, we
decided to randomly assign a wide cohort of consecutive
patients with proximal DVT to receive fixed- or flexible-
duration anticoagulation on the basis of ultrasonography
findings. Patients allocated to fixed-duration therapy dis-
continued anticoagulation at randomization (3 months
after the qualifying episode) if they had secondary DVT
and received 3 additional months of therapy if they had
unprovoked DVT. Patients in the flexible-duration group
www.annals.org

received 9 additional months of therapy if they had sec-
ondary DVT and 21 additional months if they had unpro-
voked DVT (unless ultrasonography showed recanalization
of leg veins).
Our results show that assessment of residual thrombo-
sis by ultrasonography improves the identification of pa-
tients who are at highest risk for recurrence, and it allowed
decisions on anticoagulation that led to a substantial reduc-
tion in recurrent events, whereas only a limited proportion
of patients received prolonged treatment (Table 2). Tailor-
ing the duration of oral anticoagulant therapy on the basis
of findings on repeated leg vein ultrasonography reduced
the risk for recurrent VTE by 35% compared with the
administration of conventional, fixed doses without an ap-
preciable increase in hemorrhagic risk. These findings are
consistent with those recently obtained in a smaller group
of patients by Siragusa and coworkers (14), who adopted a
different study design. Those investigators randomly as-
signed 180 patients with residual thrombosis, as shown by
ultrasonography 3 months after the acute episode, to re-
ceive 9 additional months of anticoagulation or therapy
discontinuation, whereas anticoagulation was discontinued
in the 78 patients whose veins had already recanalized. All
patients were followed for 2 years after the index episode.
Of the latter group, only 1 patient (1.3%) had recurrence.
The recurrence rate in the former group was 19.3% in
patients who continued and 27.2% in those who discon-
tinued anticoagulation, resulting in an HR of 0.42, which
is consistent with our findings (HR, 0.64).
Our results suggest that the absolute benefit was great-
est in patients with unprovoked DVT, in whom 12 throm-
boembolic events were prevented at the cost of 65 person-
years of extra treatment and no additional major bleeding
events. In contrast, only 2 thromboembolic events were
prevented at the cost of 36 person-years of extra treatment
and 2 additional major bleeding events in patients with
secondary DVT. Thus, factors other than residual venous
thrombosis (such as carriage of thrombophilia) are likely to
Figure 2. Cumulative incidence of recurrent thromboembolism.
20
18 Fixed OAT
16 Flexible OAT
Incidence of Recurrent VTE, %
14
12
10
8
6
4
2
0 lation was not successful because it was less protective than
0 5 10
Patients at risk, n
Flexible OAT 270 256
Fixed OAT 268 241
OAT ⫽ oral anticoagulant therapy; VTE ⫽ venous thromboembolism.
www.annals.org
Residual Thrombosis and Duration of Anticoagulation Article
play a preponderant role in determining the risk for re-
current events in patients with secondary DVT. Of note,
anticoagulation in unprovoked DVT was also discontinued
after 21 months in 32 patients who still had residual
thrombosis. Prolonging anticoagulation beyond 21 months
in these patients (who represent approximately 20% of the
initial group) might have further improved the results ob-
tained in this study. However, it might have increased the
rate of a major bleeding event as well; thus, we cannot
predict an advantage in terms of the benefit–risk ratio of
indefinitely prolonging anticoagulation in all such patients.
The use of serial ultrasonography to guide the dura-
tion of anticoagulation may increase the ultrasonography
burden, which may not be desirable for patients and oper-
ators. However, based on information that has accumu-
lated in recent years, increasing evidence supports repeated
ultrasonography in patients with proximal DVT whose
veins have not recanalized in order to obtain baseline val-
ues that can allow diagnosis of recurrent ipsilateral DVT if
new symptoms occur after the index episode (12, 15, 16).
We repeated ultrasonography in patients with residual
thrombosis in both study groups, and the reduction in the
rate of recurrent ipsilateral thrombosis in patients allocated
to the flexible duration—whose diagnosis was essentially
allowed by the comparison with previously available ultra-
sonographic findings—was the factor that contributed
most to the study results (Table 2).
Of note, we showed that prolonging anticoagulation
may accelerate vein recanalization. Persistent vein obstruc-
tion at the last ultrasonographic assessment was shown in a
significantly smaller proportion of patients randomly as-
signed to the flexible-duration group (20.0%) than in the
fixed-duration group (29.5%). These findings can help in-
terpret the mechanism behind the success of our individu-
alized approach to the long-term prevention of recurrent
VTE. They are consistent with the view that residual
thrombosis is a marker of hypercoagulability (8, 9) and
that it may predict late cardiovascular death (10).
After searching MEDLINE and carefully scrutinizing
all pertinent literature published in English until October
2008, we believe that our results expand on those obtained
so far in the field of long-term prevention of recurrent
thromboembolism in patients with unprovoked VTE. Pro-
longing anticoagulation for 6 to 24 months in all patients
was found to reduce recurrent thromboembolism during
treatment but did not offer long-standing benefit (13, 17–
19). On the other hand, prolonging anticoagulation indef-
initely is associated with an unacceptably high rate of ma-
jor bleeding events (4 – 6). An attempt to improve the
benefit–risk ratio by decreasing the intensity of anticoagu-
15 20 25 30 35
Month
the standard intensity and did not decrease the bleeding
236 218 risk (20, 21). Thus, the key purpose of scientific research is
221 207
to discriminate between patients who may need prolonga-
tion of their anticoagulation and those who may not. The
design in which serial ultrasonography is done and deci-
5 May 2009 Annals of Internal Medicine Volume 150 • Number 9 583
Article Residual Thrombosis and Duration of Anticoagulation
sions made about continuing anticoagulation in response
to ultrasonography results reflects increasing understanding
that postbaseline variables may be as (or more) important
for predicting risk for VTE recurrence as baseline charac-
teristics, such as sex, thrombophilia, and type of DVT. A
similar example conceptually is using D-dimer levels when
anticoagulation is discontinued (or shortly thereafter) to
make decisions about continuing anticoagulation (22).
Similarly, recent reports indicate that development of the
postthrombotic syndrome (a postbaseline variable) (23, 24)
and failure to achieve adequate anticoagulation in the first
3 months after the thrombotic episode (25) predict a
higher risk for recurrent VTE.
The rate of recurrence was 11.9% in the flexible-
duration group, indicating that a substantial number of
patients have recurrence despite vein recanalization. A
lower, yet unacceptably high, failure rate was shown in
persons who discontinued anticoagulation because of a
negative D-dimer result obtained 1 month after discontinu-
ing warfarin treatment (22). Other options should there-
fore be tested by incorporating residual venous thrombosis
in an algorithm, including the assessment of other high-
performing baseline or postbaseline variables.
Our study has limitations. First, it was not a placebo-
controlled, double-blind trial. However, our findings are
likely to be valid because several measures were taken to
avoid bias. We included consecutive patients; followed all
those included; adhered to the prespecified treatment reg-
imens; did a central, independent adjudication of all out-
come events while unaware of treatment allocation; and
included all randomly assigned patients in the analysis.
Second, the sample size was not powered to detect differ-
ences in major bleeding events between the 2 study groups
or to detect effectiveness of the intervention in patients
with unprovoked and secondary DVT. Third, because pa-
tients with unprovoked DVT randomly assigned to fixed-
duration anticoagulation received 6 months of warfarin, we
do not know how our result would compare with a longer
period of anticoagulation. Fourth, exclusion of patients
with primary pulmonary embolism; isolated calf-vein
thrombosis; previous thromboembolism; cancer; or other
permanent risk factors of recurrent thromboembolism,
thrombophilic abnormalities other than factor V Leiden,
and prothrombin mutation does not permit us to extend
the implications of our study results to these patients.
Fifth, the labeling of sudden, otherwise unexplained death
as potentially accounted for by pulmonary embolism may
have spuriously increased the case-fatality rate over that
reported in the literature (26). However, the case-fatality
rate in our patients (11.5%) is consistent with that
(10.2%) obtained by Douketis and coworkers (27) in their
recently published prospective cohort study in which they
adopted the same conservative approach. Finally, the ultra-
sonographic criteria that we adopted for defining residual
thrombosis are not widely used on a routine basis and
differ from those used in similar studies addressing the
584 5 May 2009 Annals of Internal Medicine Volume 150 • Number 9
value of residual thrombosis in predicting the risk for re-
currence (9, 10, 14). However, we based them on findings
obtained in separate prospective cohorts of patients with
proximal DVT, and they were found to be highly repro-
ducible, fast, and easy to obtain (8, 12).
Our results show that adjusting the duration of anti-
coagulation on the basis of ultrasonography in patients
with proximal DVT reduces the risk for long-term recur-
rent thromboembolism compared with a fixed duration
without increasing the hemorrhagic risk. Further studies
are needed to identify the optimal duration of treatment of
patients with DVT whose veins do not recanalize within
the first 2 years, to find clinical or biological markers that
can help identify patients who have recurrence despite vein
recanalization and to assess the additional value of incor-
porating residual venous thrombosis in algorithms that in-
clude the assessment of other high-performing baseline or
postbaseline variables for identifying patients who have the
highest risk for recurrent thromboembolic events.
From University of Padua, Padua, Italy; University of Maastricht, Maas-
tricht, and University of Amsterdam, Amsterdam, the Netherlands; and
Santa Maria Nuova Hospital, Reggio Emilia, University of Insubria,
Varese, Hospital of Piacenza, Piacenza, Treviso Healthcare Trust, Tre-
viso, Civic Hospital of Venice, Venice, Civic Hospital, Castelfranco
Veneto, Civic Hospital, Rovigo, Civic Hospital of Parma, Parma, and
University of Perugia, Perugia, Italy.
Acknowledgment: The authors thank all of their patients, without
whom this study would not have been possible.
Potential Financial Conflicts of Interest: None disclosed.
Reproducible Research Statement: Study protocol: Available from Dr.
Prandoni (e-mail, paoloprandoni@tin.it). Statistical code and data set:
Available from Dr. Prins (e-mail, mh.prins@epid.unimaas.nl).
Requests for Single Reprints: Paolo Prandoni, MD, PhD, Department
of Cardiothoracic and Vascular Sciences, Clinica Medica II, Thrombo-
embolism Unit, University of Padua, Via Giustiniani 2, 35128 Padua,
Italy; e-mail, paoloprandoni@tin.it.
Current author addresses and author contributions are available at www
.annals.org.
References
1. Segal JB, Streiff MB, Hofmann LV, Hoffman LV, Thornton K, Bass EB.
Management of venous thromboembolism: a systematic review for a practice
guideline. Ann Intern Med. 2007;146:211-22. [PMID: 17261856]
2. Ost D, Tepper J, Mihara H, Lander O, Heinzer R, Fein A. Duration of
anticoagulation following venous thromboembolism: a meta-analysis. JAMA.
2005;294:706-15. [PMID: 16091573]
3. Streiff MB, Segal JB, Tamariz LJ, Jenckes MW, Bolger DT, Eng J, et al.
Duration of vitamin K antagonist therapy for venous thromboembolism: a sys-
tematic review of the literature. Am J Hematol. 2006;81:684-91. [PMID:
16838335]
4. Schulman S, Granqvist S, Holmström M, Carlsson A, Lindmarker P, Nicol
P, et al. The duration of oral anticoagulant therapy after a second episode of
venous thromboembolism. The Duration of Anticoagulation Trial Study Group.
N Engl J Med. 1997;336:393-8. [PMID: 9010144]
www.annals.org

5. Kearon C, Kahn SR, Agnelli G, Goldhaber S, Raskob GE, Comerota AJ;
American College of Chest Physicians. Antithrombotic therapy for venous
thromboembolic disease: American College of Chest Physicians Evidence-Based
Clinical Practice Guidelines (8th Edition). Chest. 2008;133:454S-545S. [PMID:
18574272]
6. Cardiovascular Disease Educational and Research Trust. Prevention and
treatment of venous thromboembolism. International Consensus Statement
(guidelines according to scientific evidence). Int Angiol. 2006;25:101-61.
[PMID: 16763532]
7. Prandoni P, Noventa F, Ghirarduzzi A, Pengo V, Bernardi E, Pesavento R,
et al. The risk of recurrent venous thromboembolism after discontinuing antico-
agulation in patients with acute proximal deep vein thrombosis or pulmonary
embolism. A prospective cohort study in 1,626 patients. Haematologica. 2007;
92:199-205. [PMID: 17296569]
8. Prandoni P, Lensing AW, Prins MH, Bernardi E, Marchiori A, Bagatella P,
et al. Residual venous thrombosis as a predictive factor of recurrent venous
thromboembolism. Ann Intern Med. 2002;137:955-60. [PMID: 12484710]
9. Piovella F, Crippa L, Barone M, Viganò D’Angelo S, Serafini S, Galli L,
et al. Normalization rates of compression ultrasonography in patients with a first
episode of deep vein thrombosis of the lower limbs: association with recurrence
and new thrombosis. Haematologica. 2002;87:515-22. [PMID: 12010666]
10. Young L, Ockelford P, Milne D, Rolfe-Vyson V, Mckelvie S, Harper P.
Post-treatment residual thrombus increases the risk of recurrent deep vein throm-
bosis and mortality. J Thromb Haemost. 2006;4:1919-24. [PMID: 16836658]
11. Simioni P. The molecular genetics of familial venous thrombosis. Baillieres
Best Pract Res Clin Haematol. 1999;12:479-503. [PMID: 10856982]
12. Prandoni P, Cogo A, Bernardi E, Villalta S, Polistena P, Simioni P, et al. A
simple ultrasound approach for detection of recurrent proximal-vein thrombosis.
Circulation. 1993;88:1730-5. [PMID: 8403319]
13. Pinede L, Ninet J, Duhaut P, Chabaud S, Demolombe-Rague S, Durieu I,
et al; Investigators of the “Durée Optimale du Traitement AntiVitamines K”
(DOTAVK) Study. Comparison of 3 and 6 months of oral anticoagulant therapy
after a first episode of proximal deep vein thrombosis or pulmonary embolism
and comparison of 6 and 12 weeks of therapy after isolated calf deep vein throm-
bosis. Circulation. 2001;103:2453-60. [PMID: 11369685]
14. Siragusa S, Malato A, Anastasio R, Cigna V, Milio G, Amato C, et al.
Residual vein thrombosis to establish duration of anticoagulation after a first
episode of deep vein thrombosis: the Duration of Anticoagulation based on Com-
pression UltraSonography (DACUS) study. Blood. 2008;112:511-5. [PMID:
18497320]
15. Prandoni P, Lensing AW, Bernardi E, Villalta S, Bagatella P, Girolami A,
et al; DERECUS Investigators Group. The diagnostic value of compression
ultrasonography in patients with suspected recurrent deep vein thrombosis.
Thromb Haemost. 2002;88:402-6. [PMID: 12353067]
16. Prandoni P, Tormene D, Dalla Valle F, Concolato A, Pesavento R.
D-dimer as an adjunct to compression ultrasonography in patients with suspected
www.annals.org
Residual Thrombosis and Duration of Anticoagulation Article
recurrent deep vein thrombosis [Letter]. J Thromb Haemost. 2007;5:1076-7.
[PMID: 17367489]
17. Schulman S, Rhedin AS, Lindmarker P, Carlsson A, Lärfars G, Nicol P,
et al. A comparison of six weeks with six months of oral anticoagulant therapy
after a first episode of venous thromboembolism. Duration of Anticoagulation
Trial Study Group. N Engl J Med. 1995;332:1661-5. [PMID: 7760866]
18. Agnelli G, Prandoni P, Santamaria MG, Bagatella P, Iorio A, Bazzan M,
et al; Warfarin Optimal Duration Italian Trial Investigators. Three months
versus one year of oral anticoagulant therapy for idiopathic deep venous throm-
bosis. Warfarin Optimal Duration Italian Trial Investigators. N Engl J Med.
2001;345:165-9. [PMID: 11463010]
19. Agnelli G, Prandoni P, Becattini C, Silingardi M, Taliani MR, Miccio M,
et al; Warfarin Optimal Duration Italian Trial Investigators. Extended oral
anticoagulant therapy after a first episode of pulmonary embolism. Ann Intern
Med. 2003;139:19-25. [PMID: 12834314]
20. Ridker PM, Goldhaber SZ, Danielson E, Rosenberg Y, Eby CS, Deitcher
SR, et al; PREVENT Investigators. Long-term, low-intensity warfarin therapy
for the prevention of recurrent venous thromboembolism. N Engl J Med. 2003;
348:1425-34. [PMID: 12601075]
21. Kearon C, Ginsberg JS, Kovacs MJ, Anderson DR, Wells P, Julian JA, et al;
Extended Low-Intensity Anticoagulation for Thrombo-Embolism Investiga-
tors. Comparison of low-intensity warfarin therapy with conventional-intensity
warfarin therapy for long-term prevention of recurrent venous thromboembo-
lism. N Engl J Med. 2003;349:631-9. [PMID: 12917299]
22. Palareti G, Cosmi B, Legnani C, Tosetto A, Brusi C, Iorio A, et al;
PROLONG Investigators. D-dimer testing to determine the duration of antico-
agulation therapy. N Engl J Med. 2006;355:1780-9. [PMID: 17065639]
23. Stain M, Schönauer V, Minar E, Bialonczyk C, Hirschl M, Weltermann A,
et al. The post-thrombotic syndrome: risk factors and impact on the course of
thrombotic disease. J Thromb Haemost. 2005;3:2671-6. [PMID: 16359506]
24. Rodger MA, Kahn SR, Wells PS, Anderson DA, Chagnon I, Le Gal G,
et al. Identifying unprovoked thromboembolism patients at low risk for recur-
rence who can discontinue anticoagulant therapy. CMAJ. 2008;179:417-26.
[PMID: 18725614]
25. Palareti G, Legnani C, Cosmi B, Guazzaloca G, Cini M, Mattarozzi S. Poor
anticoagulation quality in the first 3 months after unprovoked venous thrombo-
embolism is a risk factor for long-term recurrence. J Thromb Haemost. 2005;3:
955-61. [PMID: 15869591]
26. Douketis JD, Kearon C, Bates S, Duku EK, Ginsberg JS. Risk of fatal
pulmonary embolism in patients with treated venous thromboembolism. JAMA.
1998;279:458-62. [PMID: 9466640]
27. Douketis JD, Gu CS, Schulman S, Ghirarduzzi A, Pengo V, Prandoni P.
The risk for fatal pulmonary embolism after discontinuing anticoagulant therapy
for venous thromboembolism. Ann Intern Med. 2007;147:766-74. [PMID:
18056660]
5 May 2009 Annals of Internal Medicine Volume 150 • Number 9 585

Current Author Addresses: Dr. Prandoni: Department of Cardiotho-
racic and Vascular Sciences, Clinica Medica II, Thromboembolism Unit,
University of Padua, Via Giustiniani 2, 35128 Padua, Italy.
Dr. Prins: Department of Clinical Epidemiology and Technology Assess-
ment, University of Maastricht, PO Box 616, NL-6200 MD Maastricht,
the Netherlands.
Dr. Lensing: Center for Vascular Medicine, Academic Medical Center,
University of Amsterdam, Meibergdreef 9, Box 22660, Amsterdam, the
Netherlands.
Dr. Ghirarduzzi: Department of Internal Medicine, Angiology Unit, Ar-
cispedale Santa Maria Nuova, Viale Risorgimento 80, I-42100 Reggio
Emilia, Italy.
Dr. Ageno: Department of Clinical Medicine, University of Insubria,
Viale Borri 57, 21100 Varese, Italy.
Dr. Imberti: Thrombosis and Haemostasis Center, Emergency Depart-
ment, Hospital of Piacenza, Ospedale Civile, Via Taverna 49, I-29100
Piacenza, Italy.
Dr. Scannapieco: Committee for Thromboembolic Disease, Treviso
Healthcare Trust, 31100 Treviso, Italy.
Dr. Ambrosio: Department of Internal Medicine, Civic Hospital of
Venice, Venice, Italy.
Dr. Pesavento: Department of Angiology, Civic Hospital, Castelfranco
Veneto, Italy.
Dr. Cuppini: Department of Internal Medicine, Civic Hospital, SOC of
Medicine, ULSS 18-Rovigo, Viale Tre Martyrdoms 89, 45100 Rovigo,
Italy.
Dr. Quintavalla: Azienda Ospedaliera-Universitaria di Parma, Via Gram-
sci 14, 43100 Parma, Italy.
Dr. Agnelli: Division of Internal and Cardiovascular Medicine, Univer-
sity of Perugia, Santa Maria della Misericordia Hospital, Via Dottori 1,
06129 Perugia, Italy.
W-102 5 May 2009 Annals of Internal Medicine Volume 150 • Number 9
Annals of Internal Medicine
Author Contributions: Conception and design: P. Prandoni, G. Agnelli.
Analysis and interpretation of the data: M.H. Prins, A.W.A. Lensing.
Drafting of the article: P. Prandoni, M.H. Prins, A.W.A. Lensing, G.
Scannapieco, G.B. Ambrosio.
Critical revision of the article for important intellectual content: A.W.A.
Lensing, A. Ghirarduzzi, W. Ageno, D. Imberti, G. Scannapieco, G.B.
Ambrosio, R. Pesavento, S. Cuppini, R. Quintavalla, G. Agnelli.
Final approval of the article: P. Prandoni, M.H. Prins, A.W.A. Lensing,
A. Ghirarduzzi, W. Ageno, D. Imberti, R. Pesavento, S. Cuppini, R.
Quintavalla, G. Agnelli.
Provision of study materials or patients: P. Prandoni, A. Ghirarduzzi, W.
Ageno, D. Imberti, G. Scannapieco, G.B. Ambrosio, R. Pesavento, S.
Cuppini, R. Quintavalla.
Statistical expertise: M.H. Prins.
Administrative, technical, or logistic support: P. Prandoni.
Collection and assembly of data: P. Prandoni, A. Ghirarduzzi, W.
Ageno, D. Imberti.
APPENDIX: ADDITIONAL INVESTIGATORS OF THE
AESOPUS STUDY
In order of contributed patients: D. Tormene, M. Perlati, F.
Dalla Valle, I. Minotto, C. Becattini, P. Simioni (Padua, Italy;
229 patients); F. Landini, M. Iotti (Reggio Emilia, Italy; 135
patients); F. Dentali (Varese, Italy; 46 patients); G. Pierfranceschi
(Piacenza, Italy; 41 patients); S. Villalta, V. Pagliara (Treviso,
Italy; 32 patients); R. Parisi, C. Bortoluzzi (Venice, Italy; 18
patients); E. Barban, L. Lusiani, A. Visonà, A. Pagnan (Castel-
franco Veneto, Italy; 18 patients); A. Bonanome, G. Vescovo, S.
Zamboni (Rovigo, Italy; 13 patients); C. Pattacini (Parma, Italy;
6 patients).
www.annals.org