Devil’s Stavros
Mussurakis,
MD
Compression US Venous Thrombosis’
I
N a recent
issue al (1) described with compression
be misleading
compliance
of their
of the
study
with
the authors concentrated venous thrombosis, they
ropopliteal test
the
true-positive
spurious
If
in
results.
fidence limits, which range from 55% to 98% . The lower limit overlaps the values reported in studies published earlier, and it is not at all clear whether the new approach is an improvement; it may even be worse. The problem is in the small number of patients with deep venous thrombosis who were studied, a
minimizes
the statistical
power
of the study. The criteria used in patient selection raise the question whether the sample
US studies,
935.1298,
‘
From
venous, 935.751, 935.751, 936.751 936.1298
Academic
Department
of Radiol-
ogy, University of Sheffield, Royal Hallamshire Hospital, Sheffield SlO 2JF, England. Received July 8, 1991; accepted August 7. Address reprint requests to the author. C
RSNA,
1991
example,
there
is
is
caif thrombi
extensive,
and
there-
small asympsamdifficult
venous
seems highly the reliability evaluated, reported
sonographic
it is uncertain sensitivity and
whether specificity
surements nomenon
other’s
the
of the spurious have
review
of bias encourages
bias.
This
concordance
between the findings on US scans and venograms and inflates the level of both sensitivity and specificity. The ability of US to reveal calf
has not yet met the expectaauthorities, who believe the venous system of the calf is too of most
complex
to yield
able results
that
Since not the
and
calf deep
reli-
thrombo-
topic in the deep problem (3). Previstudies have shown
venous
thrombosis
is self-
limiting, without risk of pulmonary embolism or postphlebitic syndrome. Other investigators believe that calf
deep
venous
high cant
recurrence probability
thrombosis
tion.
Since
serious
does
rate, as well of proximal
have
a
as a signifipropaga-
complications
could
such propagation, a reasonable strategy would be to perform serial sonographic follow-up examinations follow
the popliteal vein stead of increasing
for a few days, the examination
of
in-
time exponentially by trying to exclude disease in the numerous meandering
calf veins. Compression
US is a powerful
nonin-
vasive alternative to venography, though still insufficiently sensitive for calf deep venous thrombosis. Authors expecting to challenge this cumulative evidence on firmer
should ground,
a new technique in overstatement
can sonogskillful and
cost-effective
at US. Calf venous
sis is a controversial venous thrombosis ously published
base their conclusions as the enthusiasm
for
may sometimes result of the true potential.
by the average
of a relatively are reproducible.
References I.
mea-
stable pheReliability
is
assessed by the examination being repeated by at least one additional operator, with the operators being blinded to
each
diagnostic
that
technique
operator dependent. of the method was
of interpreting
introduced
tions
thrombosis.
described
practice
with full knowledge result-”to avoid US studies”-may
thrombi
extent of claim initially
rapher or by only the most experienced sonographer. Reliability refers to the extent to which serial
1991; 181:351-353
the
For
(2). The
venograms sonographic false-positive type
that they scanned the gastrocnemius, soleal, peroneal, and posterior tibia! veins of every patient, but they later state that once they had established the diagnosis of calf deep venous thrombosis, they did not attempt to identify every site of thrombosis to avoid prolonging the examinations. While this approach may be permitted in routine, everyday practice, it is unacceptable in the formal prospective evaluation of a diagnostic technique. All sonographically visible calf veins should have been examined in a consistent manner, and the results should have been stratified by location to reveal potential pitfalls across the full anatomic spectrum of calf
be accomplished Radiology
only
overestimation of sensitivity. Extrapolation of the results to a more general population is not possible. Discrepancies also appear in the de-
The
Veins,
by Yu-
to quantify
deep
Index terms: Thrombosis, 936.751 #{149} Veins, extremities,
studied including
symptomatic
more
tion
calf symptoms normal femo-
scription of the anatomic thrombosis. The authors
Notwithstanding, sensitivity of 85% is still an impressive figure, until one examines the 95% con-
fact that
that
in in ac-
fore much easier to detect than thrombi in the soleal sinuses of tomatic high-risk patients. This pling bias has probably made it
on only deep would achieve
increase
sample biased,
system.
doubt
no
are bulkier,
a sensitivity of 85% (ii of 13 cases). Superficial thrombophlebitis should have been included in the analysis of conditions mimicking deep venous thrombosis to avoid
populations is needed
patients with unilateral and a sonographically
basic
which in fact represents the amalgamation of two different entities, superficial and deep caff venous thrombosis.
Calf
The
cel et al is highly
suboptimal
standards of methods for diagnostic research. The first problem with compliance applies to the type of disease studied, the “isolated calf venous thrombosis,”
practice.
tual
thrombosis. Howthey reached may
because
in Isolated
representative of the which US examination
of Radiology, Yucel et their experience ultrasound (US) in
isolated calf venous ever, the conclusions
Advocate
findings
during
interpreta-
2.
3.
Yucel EK, FisherJS, Egglin TX. Geller Sd. Waitman Ad. Isolated calf venous thrombosis: diagnosis with cornpression US. Radiology 1991; 179:443-446. Becker DM, Philbrick JT, Abbitt PL Real-time ultrasonography for the diagnosis of lower extremity deep venous thrornbosis: the wave of the future? Arch Intem Med 1989; 149:1731-1734. Philbrick Fr, Becker DM. dalf deep venous thrornbosis: a wolf in sheep’s clothing? Arch Intern Med 1988; 148: 2134-2138.
351
Mussurakis,
MD
Compression US Venous Thrombosis’
I
N a recent
issue al (1) described with compression
be misleading
compliance
of their
of the
study
with
the authors concentrated venous thrombosis, they
ropopliteal test
the
true-positive
spurious
If
in
results.
fidence limits, which range from 55% to 98% . The lower limit overlaps the values reported in studies published earlier, and it is not at all clear whether the new approach is an improvement; it may even be worse. The problem is in the small number of patients with deep venous thrombosis who were studied, a
minimizes
the statistical
power
of the study. The criteria used in patient selection raise the question whether the sample
US studies,
935.1298,
‘
From
venous, 935.751, 935.751, 936.751 936.1298
Academic
Department
of Radiol-
ogy, University of Sheffield, Royal Hallamshire Hospital, Sheffield SlO 2JF, England. Received July 8, 1991; accepted August 7. Address reprint requests to the author. C
RSNA,
1991
example,
there
is
is
caif thrombi
extensive,
and
there-
small asympsamdifficult
venous
seems highly the reliability evaluated, reported
sonographic
it is uncertain sensitivity and
whether specificity
surements nomenon
other’s
the
of the spurious have
review
of bias encourages
bias.
This
concordance
between the findings on US scans and venograms and inflates the level of both sensitivity and specificity. The ability of US to reveal calf
has not yet met the expectaauthorities, who believe the venous system of the calf is too of most
complex
to yield
able results
that
Since not the
and
calf deep
reli-
thrombo-
topic in the deep problem (3). Previstudies have shown
venous
thrombosis
is self-
limiting, without risk of pulmonary embolism or postphlebitic syndrome. Other investigators believe that calf
deep
venous
high cant
recurrence probability
thrombosis
tion.
Since
serious
does
rate, as well of proximal
have
a
as a signifipropaga-
complications
could
such propagation, a reasonable strategy would be to perform serial sonographic follow-up examinations follow
the popliteal vein stead of increasing
for a few days, the examination
of
in-
time exponentially by trying to exclude disease in the numerous meandering
calf veins. Compression
US is a powerful
nonin-
vasive alternative to venography, though still insufficiently sensitive for calf deep venous thrombosis. Authors expecting to challenge this cumulative evidence on firmer
should ground,
a new technique in overstatement
can sonogskillful and
cost-effective
at US. Calf venous
sis is a controversial venous thrombosis ously published
base their conclusions as the enthusiasm
for
may sometimes result of the true potential.
by the average
of a relatively are reproducible.
References I.
mea-
stable pheReliability
is
assessed by the examination being repeated by at least one additional operator, with the operators being blinded to
each
diagnostic
that
technique
operator dependent. of the method was
of interpreting
introduced
tions
thrombosis.
described
practice
with full knowledge result-”to avoid US studies”-may
thrombi
extent of claim initially
rapher or by only the most experienced sonographer. Reliability refers to the extent to which serial
1991; 181:351-353
the
For
(2). The
venograms sonographic false-positive type
that they scanned the gastrocnemius, soleal, peroneal, and posterior tibia! veins of every patient, but they later state that once they had established the diagnosis of calf deep venous thrombosis, they did not attempt to identify every site of thrombosis to avoid prolonging the examinations. While this approach may be permitted in routine, everyday practice, it is unacceptable in the formal prospective evaluation of a diagnostic technique. All sonographically visible calf veins should have been examined in a consistent manner, and the results should have been stratified by location to reveal potential pitfalls across the full anatomic spectrum of calf
be accomplished Radiology
only
overestimation of sensitivity. Extrapolation of the results to a more general population is not possible. Discrepancies also appear in the de-
The
Veins,
by Yu-
to quantify
deep
Index terms: Thrombosis, 936.751 #{149} Veins, extremities,
studied including
symptomatic
more
tion
calf symptoms normal femo-
scription of the anatomic thrombosis. The authors
Notwithstanding, sensitivity of 85% is still an impressive figure, until one examines the 95% con-
fact that
that
in in ac-
fore much easier to detect than thrombi in the soleal sinuses of tomatic high-risk patients. This pling bias has probably made it
on only deep would achieve
increase
sample biased,
system.
doubt
no
are bulkier,
a sensitivity of 85% (ii of 13 cases). Superficial thrombophlebitis should have been included in the analysis of conditions mimicking deep venous thrombosis to avoid
populations is needed
patients with unilateral and a sonographically
basic
which in fact represents the amalgamation of two different entities, superficial and deep caff venous thrombosis.
Calf
The
cel et al is highly
suboptimal
standards of methods for diagnostic research. The first problem with compliance applies to the type of disease studied, the “isolated calf venous thrombosis,”
practice.
tual
thrombosis. Howthey reached may
because
in Isolated
representative of the which US examination
of Radiology, Yucel et their experience ultrasound (US) in
isolated calf venous ever, the conclusions
Advocate
findings
during
interpreta-
2.
3.
Yucel EK, FisherJS, Egglin TX. Geller Sd. Waitman Ad. Isolated calf venous thrombosis: diagnosis with cornpression US. Radiology 1991; 179:443-446. Becker DM, Philbrick JT, Abbitt PL Real-time ultrasonography for the diagnosis of lower extremity deep venous thrornbosis: the wave of the future? Arch Intem Med 1989; 149:1731-1734. Philbrick Fr, Becker DM. dalf deep venous thrornbosis: a wolf in sheep’s clothing? Arch Intern Med 1988; 148: 2134-2138.
351
