Journal of Thrombosis and Haemostasis, 14: 964–972 DOI: 10.1111/jth.

13279

ORIGINAL ARTICLE

Prevalence of deep vein thrombosis and pulmonary embolism

in patients with superficial vein thrombosis: a systematic
review and meta-analysis
M . N . D . D I M I N N O , * † P . A M B R O S I N O , ‡ F . A M B R O S I N I , § E . T R E M O L I , † G . D I M I N N O ‡ and
F. DENTALI§
*Division of Cardiology – Department of Advanced Biomedical Sciences, Federico II University, Naples; †Unit of Cell and Molecular Biology
in Cardiovascular Diseases, Centro Cardiologico Monzino, IRCCS, Milan; ‡Department of Clinical Medicine and Surgery, Federico II
University, Naples; and §Department of Clinical and Experimental Medicine, University of Insubria, Varese, Italy

To cite this article: Di Minno MND, Ambrosino P, Ambrosini F, Tremoli E, Di Minno G, Dentali F. Prevalence of deep vein thrombosis and
pulmonary embolism in patients with superficial vein thrombosis: a systematic review and meta-analysis. J Thromb Haemost 2016; 14:
964–72.

and 11 studies (2484 patients) evaluated the prevalence of

Essentials
 The association of superficial vein thrombosis (SVT)
with venous thromboembolism (VTE) is variable.
 We performed a meta-analysis to assess the prevalence
of concomitant VTE in patients with SVT.
 Deep vein thrombosis was found in 18.1%, and pul-
monary embolism in 6.9%, of SVT patients.
 Screening for VTE may be worthy in some SVT
patients to plan adequate anticoagulant treatment.
Summary. Background: Some studies have suggested that
patients with superficial vein thrombosis (SVT) have a
non-negligible risk of concomitant deep vein thrombosis
(DVT) or pulmonary embolism (PE) at the time of SVT
diagnosis. Unfortunately, the available data on this asso-
ciation are widely variable. Objectives: To perform a sys-
tematic review and meta-analysis of the literature in order
to evaluate the prevalence of concomitant DVT/PE in
patients with SVT of the lower limbs. Methods: Studies
reporting on the presence of DVT/PE in SVT patients
were systematically searched for in the PubMed, Web of
Science, Scopus and EMBASE databases. The weighted
mean prevalence (WMP) of DVT and PE was calculated
by use of the random effect model. Results: Twenty-one
studies (4358 patients) evaluated the prevalence of DVT
Correspondence: Matteo Nicola Dario Di Minno, Department of
Advanced Biomedical Sciences Division of Cardiology, Federico II
University, Via S. Pansini 5, 80131 Naples, Italy.
Tel./fax: +390817464323.
E-mail: dario.diminno@hotmail.it
Received 11 September 2015
Manuscript handled by: M. Cushman
Final decision: F. R. Rosendaal, 26 January 2016
PE in patients with SVT. The WMP of DVT at SVT
diagnosis was 18.1% (95%CI: 13.9%, 23.3%) and the
WMP of PE was 6.9% (95%CI: 3.9%, 11.8%). Hetero-
geneity among the studies was substantial. Selection of
studies including outpatients only gave similar results
(WMP of DVT, 18.2%, 95% CI 12.2–26.3%; and WMP
of PE, 8.2%, 95% CI 3.3–18.9%). Younger age, female
gender, recent trauma and pregnancy were inversely asso-
ciated with the presence of DVT/PE in SVT patients.
Conclusions: The results of our large meta-analysis sug-
gest that the prevalence of DVT and PE in patients pre-
senting with SVT is not negligible. Screening for a major
thromboembolic event may be worthwhile in some SVT
patients, in order to allow adequate anticoagulant treat-
ment to be planned. Other high-quality studies are war-
ranted to confirm our findings.
Keywords: deep vein thrombosis; phlebothrombosis;
pulmonary embolism; superficial vein thrombosis;
thrombophlebitis.
Introduction
Superficial vein thrombosis (SVT) is a common disease
that most often affects the veins of the lower limbs,
but that can also be found in other sites. The great
saphenous vein is involved in 60–80% of cases, and the
small saphenous vein in 10–20% [1]. Little information
is available on the epidemiology of SVT of the lower
limbs. It has been estimated that there is a prevalence
of 3–11% in the general population [2], which is
approximately two-fold higher than that of deep vein
thrombosis (DVT) and pulmonary embolism (PE) com-
bined [3]. Moreover, in a recent community-based
study, the annual incidence of SVT appeared to be ~ 6/

© 2016 International Society on Thrombosis and Haemostasis


1000 patients, which is approximately three times higher
than that of DVT [4].
Besides sharing prothrombotic risk factors with DVT
and PE, including a personal or family history of venous
thromboembolism, active malignancy, recent surgery or
trauma, immobilization, inherited thrombophilia, use of
oral contraceptives, infectious diseases, obesity, and car-
diac or respiratory failure [5,6], SVT is often accompa-
nied by the presence of varicose veins, which are
documented in up to 80% of SVT patients [7]. SVT has
long been considered to be a benign entity, with more
local than systemic implications. However, in more
recent years, it has become clearer that SVT may be a
manifestation of a systemic tendency to thrombosis, with
a non-negligible risk of recurrence or concomitant DVT
or PE at the time of SVT diagnosis [8]. Unfortunately,
the available data on this association are widely vari-
able, ranging between 6% and 36% for concomitant
DVT and up to 33% for PE [1,9]. To better address this
issue, we performed a systematic review and meta-analy-
sis of the literature to evaluate the prevalence of con-
comitant DVT/PE in patients with SVT of the lower
limbs.
Materials and methods
A protocol for this review was prospectively developed,
detailing the specific objectives, the criteria for study
selection, the approach to assessing the outcomes, and the
statistical methods.
Aim of the study
The primary aim of our study was the assessment of the
prevalence of DVT and PE at the time of SVT diagnosis.
Search strategy and study selection
To identify all available studies, a detailed search per-
taining to the association of lower limb SVT with
DVT/PE was conducted according to Preferred Report-
ing Items for Systematic reviews and Meta-Analyses
(PRISMA) guidelines [10]. A systematic search was per-
formed in the electronic databases (PubMed, Web of
Science, Scopus, and EMBASE) up to May 2015, with
the following search terms (as medical subjects headings
and text words) in all possible combinations: superficial
vein thrombosis, superficial venous thrombosis, phle-
bothrombosis, thrombophlebitis, thrombosis of superfi-
cial veins and venous thromboembolism, venous
thrombosis, deep vein thrombosis, pulmonary embolism,
and deep venous system. The last search was performed
on November 2015. Research was supplemented by
manually reviewing abstract books from congresses of
the ISTH, the American Society of Angiology, and the
International Society for Vascular Surgery (2003–2014).
© 2016 International Society on Thrombosis and Haemostasis
Superficial vein thrombosis and VTE 965
The search strategy was developed without any lan-
guage restriction.
In addition, the reference lists of all retrieved articles
were manually reviewed. Two independent authors
(M.N.D.D.M. and P.A.) analyzed each article and per-
formed the study selection independently. In cases of dis-
agreement, a third investigator was consulted (F.D.).
Discrepancies were resolved by consensus. The selection
results showed high inter-reader agreement (j = 0.98),
and have been reported according to the PRISMA flow-
chart (Fig. S1).
Data extraction and quality assessment
According to the prespecified protocol, all studies report-
ing on DVT and/or PE prevalence in patients with an
objectively documented SVT of the lower limb (by ultra-
sound, phlebography, plethysmography, or venography)
were included.
Case reports, reviews and animal studies were excluded.
For each study, data regarding sample size and major
clinical and demographic variables were extracted.
Formal quality score adjudication was not used, as pre-
vious investigations failed to demonstrate its usefulness
[11].
Statistical analysis and risk of bias assessment
Statistical analysis was carried out with COMPREHENSIVE
META-ANALYSIS (Version 2; Biostat, Englewood, NJ,
USA).
The prevalence of DVT and/or PE was expressed as
weighted mean prevalence (WMP) with pertinent 95%
confidence intervals (CIs). Statistical heterogeneity
between studies was assessed with the chi-square Cochran
Q-test and with the I2 statistic, which measures the incon-
sistency across study results and describes the proportion
of total variation in study estimates that is attributable to
heterogeneity rather than sampling error. In detail, I2 val-
ues of 0% indicate no heterogeneity, 25% low hetero-
geneity, 25–50% moderate heterogeneity, and 50% high
heterogeneity [12].
Publication bias was assessed with Egger’s test and rep-
resented graphically by funnel plots of the standard dif-
ference in means versus the standard error. Visual
inspection of funnel plot asymmetry was performed to
account for possible small-study effects, and Egger’s test
was used to assess publication bias, over and above any
subjective evaluation. A P-value of < 0.10 was considered
to be statistically significant [13]. In case of a significant
publication bias, the Duval and Tweedie trim and fill
method was used to allow for the estimation of an
adjusted effect size [14]. In order to be as conservative as
possible, the random-effect method was used for all anal-
yses to take into account the variability among the
included studies.
966 M. N. D. Di Minno et al

tioned covariates as independent variables (x). This

Sensitivity analyses
In the framework of a sensitivity analysis, all analyses
were repeated with selection of studies including outpa-
tients only. In addition, the results have been stratified
according to the study design (prospective or retrospec-
tive) and according to the methodology used to screen for
the presence of DVT (ultrasound from groin to ankle, or
ultrasound limited to the thigh) and of PE (evaluation of
symptomatic patients or systematic screening of both
symptomatic and asymptomatic patients).
Meta-regression analyses
To assess the possible effects of demographic variables
(mean age and female gender) and of some major risk
factors for venous thrombosis (the prevalence of vari-
cose veins, involvement of the saphenofemoral junction
[SFJ], active cancer, recent surgery, trauma, and preg-
nancy) on the prevalence of concomitant DVT/PE, we
planned to perform meta-regression analyses after
implementing a regression model with DVT/PE preva-
lence as the dependent variable (y) and the above-men-
analysis was performed with COMPREHENSIVE META-ANA-
LYSIS (Version 2).
Results
After exclusion of duplicate results, the search retrieved
23 147 articles. Of these studies, 17 402 were excluded
after scanning of the title and/or the abstract because they
were off the topic, and 5699 because they were reviews/
comments/case reports or they lacked of data of interest.
Twenty-four further studies were excluded after full-
length paper evaluation (Fig. S1).
Thus, 22 studies [4,5,9,15–33] with a total of 4379 SVT
patients were included in the meta-analysis, 21 [4,5,15–33]
reporting data on the prevalence of DVT, and 11
[4,5,9,16,17,21,23,24,30,31,33] reporting data on the pres-
ence of PE.
Study characteristics
The main characteristics of the 22 included studies are
shown in Table 1.

Table 1 Demographic and clinical data of studies on the prevalence of deep vein thrombosis/pulmonary embolism in patients with superficial
vein thrombosis (SVT)

SVT Origin of Mean age M/F Varicose Malignancy Obesity* Surgery Trauma Pregnancy
Author (n) patients (years) (n) veins (%) (%) (%) (%) (%) (%)

Ascher, 2009 [15] 32 Outpatients 57.8 14/18 59 0 NA NA 0 0

Bergqvist, 1986 [16] 56 Outpatients 58.0 21/35 67.8 3.6 NA NA NA NA
Binder, 2009 [17] 46 Outpatients 65.0 14/32 100 7.0 NA NA NA NA
Bl€
attler, 1993 [18] 25 Outpatients 47.0 NA 76.0 12.0 NA NA NA NA
Blumenberg, 1997 [19] 232 Outpatients NA NA NA NA NA NA NA NA
Bounameaux, 1997 [20] 554 Outpatients NA NA NA NA NA NA NA NA
Decousus, 2010 [5] 844 Outpatients/ 65.0 296/548 81.8 6.0 28.8 4.3 5.4 4.5
inpatients
Frappe, 2014 [4] 171 Outpatients/ 68.0 60/111 82.8 7.9 25.3 3.7 5.6 3.8
inpatients
Galanaud, 2011 [21] 788 Outpatients/ 65.0 283/505 64.8 8.2 16.1 7.7 NA 3.2
inpatients
Gorty, 2004 [22] 60 Outpatients 52.0 23/37 80.0 7.0 NA 20.0 7.0 NA
Hirmerova, 2013 [23] 138 Outpatients 61.4 50/88 89.8 6.5 NA NA NA NA
Lutter, 1991 [24] 187 Outpatients 58.4 87/99 51.9 14.9 35.8 24.6 13.4 NA
Noppeney, 2009 [25]† 114 Outpatients 60 42/72 81.3 NA NA NA NA NA
Pomero, 2015 [26] 494 Outpatients 56.3 314/180 32.2 9.4 19 9.1 11.1 NA
Prountjos, 1991 [27] 57 Inpatients 58.9 23/34 NA NA NA NA NA NA
Pulliam, 1991 [28] 20 Outpatients NA 8/12 NA NA NA NA NA NA
Rathbun, 2012 [29] 302 Outpatients/ NA NA NA NA NA NA NA NA
inpatients
Skillman, 1990 [30] 42 Outpatients/ NA NA 92.8 4.8 NA 21.4 4.8 11.9
inpatients
Sobreira, 2009 [31] 60 Outpatients 50.2 20/40 85.0 18.3 NA 30.0 10.0 NA
Verlato, 1999 [9] 21 Outpatients 61.3 11/10 80.9 4.8 NA 4.8 4.8 NA
Yucel, 1992 [32] 5 Outpatients 79.0 1/4 NA NA NA NA NA NA
Wichers, 2008 [33] 131 Outpatients NA NA NA NA NA NA NA NA

M/F, male/female; NA, not available. *Obesity was defined as a body mass index of > 30 kg m 2 in the studies by Decousus (2010), Frappe
(2014), Galanaud (2011), and Pomero (2015). The criteria for defining obesity are missing in the study by Lutter (1991). †Only abstract avail-
able.

© 2016 International Society on Thrombosis and Haemostasis

 Superficial vein thrombosis and VTE 967

Fifteen studies had a prospective design [4,5,9,15–
18,21,23,27–32], and seven were retrospective studies
[19,20,22,24–26,33].
One study specifically enrolled in-hospital patients [27],
five studies [4,5,21,29,30] included both in-hospital patients
and outpatients with SVT, and all of the other studies
specifically included outpatients with SVT. In 21 of 22
studies [5,9,15–33], patients with a clinically suspected or
objectively diagnosed SVT were evaluated for the concur-
rent presence of DVT/PE, whereas only one study [4] sys-
tematically screened asymptomatic patients for SVT and
DVT/PE presence. As detailed in Table S1A, among the 21
studies reporting on the presence of DVT, in 11 a system-
atic ultrasound examination from groin to ankle was per-
formed [4,5,15,17,19,21–25,33], and in the other 10 studies
screening for DVT was performed with venography,
plethysmography, or an ultrasound assessment limited to
the thigh. PE was searched for by means of chest radiogra-
phy, ventilation–perfusion lung scanning or computed
tomography scan in seven studies [4,5,9,21,23,24,31], four
of which evaluated only symptomatic patients [4,5,21,23],
and three of which [9,24,31] also systematically screened
for PE in asymptomatic patients. In two studies [17,30], the
diagnosis of PE was based on clinical assessment, and for
two other studies [16,33] information about the criteria and
imaging techniques used was missing.
The number of patients varied from five to 844, the mean
age varied from 47 years to 79 years, and the prevalence of
female gender varied from 47.6% to 80%. Among major
risk factors for venous thrombosis, an active malignancy
was reported by 0–18.3%, obesity by 16.1–35.8%, recent
surgery by 3.7–30%, trauma by 0–13.4%, and pregnancy
by 0–11.9%. The presence of varicose veins ranged between
32.2% and 100% of patients with SVT, with a WMP of
76.7% (95% CI 66.2–84.6%). However, only two studies
[21,26] reported specific criteria for varicose vein diagnosis.
All of the other studies only classified varicose veins as ‘pre-
sent’ or ‘absent’, without any specific criteria.
One study [29] excluded patients with SFJ involvement,
whereas 12 studies [5,15,17,19,21,22,24,26,28,31,32]
showed that the WMP of SFJ involvement in SVT
patients was 17.5% (95% CI 10.4–27.9%). In more
detail, in four studies [5,15,19,32] SFJ thrombosis was
considered as a ‘high-risk SVT’ or as a DVT, whereas in
eight studies [9,17,21,22,24,26,28,31] it was analyzed with
all of the other cases of SVT.
Prevalence of DVT in patients presenting with SVT
Twenty-one studies [4,5,15–33] reported on 4358 SVT
patients screened for the presence of DVT. As reported in
Fig. 1, the prevalence of concomitant DVT at the

Study name Statistics for each study Event rate and 95% CI

Event Lower Upper

rate limit limit z-Value P-value

Ascher, 2009 [15] 0.656 0.479 0.798 1.737 0.082

Bergqvist, 1986 [16] 0.161 0.086 0.281 –4.543 0.000
Binder, 2009 [17] 0.239 0.138 0.382 –3.349 0.001
Blättler, 1993 [18] 0.440 0.263 0.634 –0.599 0.549
Blumenberg, 1997 [19] 0.060 0.036 0.099 –9.958 0.000
Bounameaux, 1997 [20] 0.056 0.040 0.078 –15.286 0.000
Decousus, 2010 [5] 0.235 0.207 0.264 –14.558 0.000
Frappé, 2014 [4] 0.246 0.187 0.316 –6.316 0.000
Galanaud, 2011 [21] 0.288 0.258 0.321 –11.502 0.000
Gorty, 2004 [22] 0.133 0.068 0.245 –4.929 0.000
Hirmerova, 2013 [23] 0.304 0.233 0.386 –4.468 0.000
Lutter, 1991 [24] 0.283 0.223 0.352 –5.716 0.000
Noppeney, 2009 [25] 0.096 0.054 0.166 –7.052 0.000
Pomero, 2015 [26] 0.160 0.130 0.195 –13.514 0.000
Prountjos, 1991 [27] 0.193 0.110 0.316 –4.263 0.000
Pulliam, 1991 [28] 0.300 0.141 0.527 –1.736 0.082
Rathbun, 2012 [29] 0.063 0.040 0.097 –11.397 0.000
Skillman, 1990 [30] 0.119 0.050 0.256 –4.201 0.000
Sobreira, 2009 [31] 0.217 0.130 0.338 –4.101 0.000
Wichers, 2008 [33] 0.031 0.012 0.079 –6.809 0.000
Yucel, 1992 [32] 0.200 0.027 0.691 –1.240 0.215
Overall effect 0.181 0.139 0.233 –9.382 0.000
–1.00 –0.50 0.00 0.50 1.00

Fig. 1. Prevalence of deep vein thrombosis in patients with superficial vein thrombosis. CI, confidence interval.

© 2016 International Society on Thrombosis and Haemostasis

968 M. N. D. Di Minno et al

moment of SVT diagnosis ranged from 3.1% to 65.6%,
with a pooled WMP of 18.1% (95% CI 13.9–23.3%) and
with significant heterogeneity among the studies
(I2 91.8%, P < 0.001). Heterogeneity was not reduced by
the exclusion of one study at a time. When the six
studies [4,5,21,27,29,30] that also included in-hospital
patients were excluded, the WMP of DVT resulted
unchanged (WMP 18.2%, 95% CI 12.2–26.3%, I2 91.6%,
P < 0.001).
When we specifically analyzed the 14 studies with a
prospective design [4,5,15–18,21,23,27–32], we found that
the WMP of concomitant DVT at the moment of SVT
diagnosis was 24.0% (95% CI 18.9–30.0%, I2 85.4%,
P < 0.001), whereas the seven retrospective studies
[19,20,22,24–26,33] showed a pooled WMP of 10.0%
(95% CI 5.6–17.2%, I2 92.9%, P < 0.001). The WMP of
DVT was 20.5% (95% CI 15.3–26.8%, I2 90.6%,
P < 0.001) when we analyzed the 11 studies in which a
complete ultrasound examination from groin to ankle
was performed to identify SVT and DVT
[4,5,15,17,19,21–25,33], and 15.0% (95% CI 3.9–43.2%,
I2 83.9%, P = 0.002) in the three studies[28,29,32] that
limited the ultrasound examination to the thigh.
Prevalence of PE in patients presenting with SVT
Eleven studies [4,5,9,16,17,21,23,24,30,31,33] with a total
of 2484 SVT patients evaluated the prevalence of con-
comitant PE at the moment of SVT diagnosis. This ran-
ged from 0.9% to 33%, with a pooled WMP of 6.9%
(95% CI 3.9–11.8%; Fig. 2), and with significant hetero-
geneity among the studies (I2 87.9%, P < 0.001). Hetero-
geneity was not reduced by the exclusion of one study at
a time. When the four studies [4,5,21,30] that also
included in-hospital patients were excluded, the WMP of
PE was 8.2% (95% CI 3.3–18.9%, I2 87.2%, P < 0.001).
When we specifically analyzed the nine studies with a
prospective design [4,5,9,16,17,21,23,30,31], we found that
the WMP of concomitant PE at the moment of SVT
diagnosis was 8.2% (95% CI 4.3–14.9%, I2 89.4%,
P < 0.001), whereas the two retrospective studies [24,34]
showed a pooled WMP of 3.6% (95% CI 2.0–6.4%,
I2 0%, P = 0.347). The WMP of PE was 9.6%
(95% CI 5.3–16.9%, I2 91.6%, P < 0.001) when we ana-
lyzed the seven studies in which PE was searched for by
means of chest radiography, ventilation–perfusion lung
scanning, or computed tomography scan
[4,5,9,21,23,24,31]. The four studies in which PE was
searched for only in symptomatic patients [4,5,21,23]
showed a WMP of PE of 6.4% (95% CI 3.9–10.4%,
I2 84.4%, P < 0.001), whereas the three studies [9,24,31]
that systematically evaluated both symptomatic and
asymptomatic patients showed a WMP of PE of 17.0%
(95% CI 4.4–48.1%, I2 92.5%, P < 0.001).
Meta-regression analyses
A meta-regression approach (Fig. 3) showed that female
gender, recent trauma and ongoing pregnancy were associ-
ated with a lower prevalence of concomitant involvement
of the deep venous system at the time of SVT diagnosis. In
contrast, increasing age was a predictor of DVT/PE in SVT
patients (Fig. 3). Interestingly, when we specifically ana-
lyzed data from studies enrolling outpatients, female gen-
der and recent trauma were confirmed as negative
predictors of concomitant involvement of the deep venous
system at the time of SVT diagnosis (respectively: z-score
of 3.57, P < 0.001; and z-score of 3.33, P < 0.001).
Moreover, we found that involvement of the SFJ was asso-
ciated with the concomitant involvement of the deep

Study name Statistics for each study Event rate and 95% CI

Event Lower Upper

rate limit limit z-Value P-value

Bergqvist, 1986 [16] 0.009 0.001 0.125 –3.328 0.001

Binder, 2009 [17] 0.011 0.001 0.149 –3.188 0.001
Decousus, 2010 [5] 0.039 0.028 0.054 –18.030 0.000
Frappé, 2014 [4] 0.047 0.024 0.091 –8.324 0.000
Galanaud, 2011 [21] 0.069 0.053 0.088 –18.507 0.000
Hirmerova, 2013 [23] 0.130 0.084 0.198 –7.506 0.000
Lutter, 1991 [24] 0.043 0.022 0.083 –8.601 0.000
Skillman, 1990 [30] 0.012 0.001 0.160 –3.123 0.002
Sobreira, 2009 [31] 0.283 0.184 0.409 –3.239 0.001
Verlato, 1999 [9] 0.333 0.168 0.553 –1.497 0.134
Wichers, 2008 [33] 0.023 0.007 0.069 –6.426 0.000
Overall effect 0.069 0.039 0.118 –8.606 0.000
–1.00 –0.50 –0.00 0.50 1.00

Fig. 2. Prevalence of pulmonary embolism in patients with superficial vein thrombosis. CI, confidence interval.

© 2016 International Society on Thrombosis and Haemostasis

 Superficial vein thrombosis and VTE 969

Regression of female gender on Logit event rate

Regression of age on Logit event rate
0.80 0.80
z-score: 2.18, P = 0.028 z-score: –2.22, P = 0.026
0.42 0.42
0.04 0.04
–0.34 –0.34

Logit event rate

Logit event rate

–0.72 –0.72
–1.10 –1.10
–1.48 –1.48
–1.86 –1.86
–2.24 –2.24
–2.62 –2.62
–3.00 –3.00
43.80 47.64 51.48 55.32 59.16 63.00 66.84 70.68 74.52 78.36 82.20 44.36 48.25 52.14 56.02 59.91 63.80 67.69 71.58 75.46 79.35 83.24
Age Female gender

Regression of pregnancy on Logit event rate Regression of trauma on Logit event rate
0.80 0.80
z-score: –4.76, P < 0.001 z-score: –2.61, P = 0.009
0.42 0.42
0.04 0.04
–0.34 –0.34

Logit event rate

Logit event rate

–0.72 –0.72
–1.10 –1.10
–1.48 –1.48
–1.86 –1.86
–2.24 –2.24
–2.62 –2.62
–3.00 –3.00
–1.19 0.24 1.67 3.09 4.52 5.95 7.38 8.81 10.23 11.66 13.09 –1.34 0.27 1.88 3.48 5.09 6.70 8.31 9.92 11.52 13.13 14.74
Pregnancy Trauma

Fig. 3. Meta-regression analysis. Effect of age, female gender, recent trauma and pregnancy on the prevalence of deep vein thrombosis/pul-
monary embolism in patients with superficial vein thrombosis.

venous system (z-score of 2.63, P = 0.008). None of the patients is not an uncommon clinical finding
other meta-regression analyses provided significant results [4,5,16,17,34–36], and that the risk of subsequent DVT or
(Fig. S2). PE during follow-up is not negligible [8,37]. However, the
prevalence of deep venous system involvement in patients
with SVT was highly variable in different studies, as
Publication bias
many of these studies had limited samples, and such high
Visual inspection of funnel plots of effect size versus stan- variability suggested the need for a meta-analysis includ-
dard error for studies evaluating the prevalence of DVT ing all available studies [21]. Thus, in this meta-analysis,
and PE suggested a symmetric distribution of studies by pooling together data from 22 studies, we have been
around the effect size, and Egger’s test confirmed the lack able, for the first time, to provide an aggregate estimation
of a significant publication bias (P = 0.196 and of the prevalence of DVT/PE in SVT patients.
P = 0.846, respectively; Fig. S3). A number of risk factors, including advanced age, obe-
sity, active cancer, previous thromboembolic episodes,
pregnancy, oral contraceptives, hormone replacement
Discussion
therapy, recent surgery, and autoimmune diseases, are
The results of the present meta-analysis, which included common to both DVT/PE and SVT [38,39].
22 studies with a total of > 4300 SVT patients, suggest The coexistence of SVT and DVT/PE might be princi-
that DVT is present in ~ 18% of patients and that PE is pally explained by the migration of the SVT towards the
present in ~ 7% of patients at the time of SVT diagnosis. deep venous system, via the SFJ, the saphenopopliteal
Interestingly, these results were confirmed when we specif- junction, or a perforating vein [40]. As the SFJ is the
ically evaluated studies including outpatients. On the point where the great saphenous vein, which is the pre-
other hand, the study design seems to have impacted on dominant location of SVT, joins the common femoral
the results. Indeed, the prevalence of DVT and of PE was vein, extension of an SVT within 3 cm of the junction
significantly higher in prospective studies than in retro- may be considered to be as serious as a proximal DVT
spective ones. Although SVT of the lower limbs has been [41]. In our study, we found that involvement of the SFJ
traditionally considered to be a relatively benign disease, was reported by 17.5% of SVT patients, and we have also
this hypothesis has been challenged by a number of recent documented its direct impact on the prevalence of deep
studies, suggesting that the presence of DVT/PE in SVT venous system involvement.

© 2016 International Society on Thrombosis and Haemostasis

970 M. N. D. Di Minno et al
Patients with an isolated SVT and patients with con-
comitant involvement of the deep venous system or of the
pulmonary tree have significantly different clinical histo-
ries. Patients with a major thromboembolic event not
treated with anticoagulants show a very high risk of
DVT/PE recurrence during the acute phase, and thus
require at least 3 months of antithrombotic therapy [42].
On the other hand, SVT patients have a lower rate of
major thromboembolic complications, and a shorter per-
iod of antithrombotic therapy at a lower dosage appeared
to be effective and safe [43–45]. In the POST study,
90.5% of patients with isolated SVT received anticoagu-
lant treatment. Low molecular weight heparin was admin-
istered at therapeutic doses to 62.9% of patients, or at
prophylactic doses to 36.7%, for a median of 11 days in
both cases. In addition, vitamin K antagonists were
administered to 16.8% of patients for a median of
81 days. Overall, during the 3-month follow-up, 2.8% of
patients experienced symptomatic DVT (46.7% being
proximal DVT) and 0.5% symptomatic PE during a
3-month follow-up [5].
Moreover, a nationwide population-based study on
10 973 Danish patients with SVT not treated with anti-
coagulants showed that, during a median follow-up of
7 years, the incidence rate of DVT/PE was 18.0/
1000 person-years (95% CI 17.2–18.9). As compared
with a non-SVT cohort of 515 067 subjects selected
from the general Danish population, the hazard ratio
of developing a major thromboembolic event was 71.4
(95% CI 60.2–84.7) during the first 3 months, steadily
decreasing to 5.1 (95% CI 4.6–5.5) in the 5 years after
the SVT [8].
Compression ultrasonography (CUS) is an easy-to-per-
form and widely available diagnostic technique with high
sensitivity and specificity for the diagnosis of DVT [46].
Because of its relatively low cost and the lack of potential
contraindications, CUS may be proposed for screening of
the deep venous system in patients presenting with SVT.
The ICARO Study Group has recently proposed a simple
score based entirely on clinical variables (active malig-
nancy, limb edema, rope-like sign, age ≥ 50 years, and
unprovoked nature of SVT) to define the pretest proba-
bility of concomitant DVT in SVT patients [26]. In line
with our results, the ICARO study concluded that the
rate of the concomitant presence of DVT is not negligible
in patients with SVT (> 30% in the high-probability cate-
gory) [26].
Computed tomographic pulmonary angiography has
greatly improved the diagnostic approach to patients with
suspected PE, becoming the reference imaging test for this
disease [47]. However, it has a non-negligible biological
risk, and it should be used with caution in some patient
groups, such as patients with severe renal insufficiency,
those with known allergy to contrast media, and pregnant
women [47]. Thus, the role of extensive screening for PE
in these patients is more debatable.
When we evaluated risk factors potentially predisposing
to involvement of the deep venous system or of the pul-
monary tree, we found that the prevalence of DVT/PE in
SVT patients is lower in younger subjects, in female
patients, in the presence of a recent trauma, or during
pregnancy. These results are in agreement with data from
previous large studies [5,21,26], further supporting a
potential role for these clinical variables in the prediction
of deep venous system involvement in SVT patients.
In contrast, we did not find a significant impact of the
presence of varicose veins on the concomitant presence of
DVT/PE in SVT patient. This association was suggested
by the results of the OPTIMEV and POST studies [5,21],
but it was not confirmed in the ICARO study [26]. How-
ever, only two studies [21,26] reported specific criteria for
varicose vein diagnosis; all of the others only classified
varicose vein as ‘present’ or ‘absent’, without any specific
criteria. This might suggest the need for further studies
on this topic.
Some studies [5,21] have suggested that the presence of
lower limb edema, dull pain, age > 75 years, active cancer,
inpatient status and a personal history of DVT/PE are also
predictive factors for the presence of concurrent DVT/PE
in SVT patients. Thus, evaluation of the risk factors for a
major thromboembolic event may be useful to assess the
pretest probability and the opportunity for screening for
DVT and PE in the clinical workup of SVT patients
[2,26,37,38,48]. In addition, some anatomic factors might
explain the association between SVT and DVT. Indeed, as
compared with SVTs affecting the thigh region, those
affecting the saphenofemoral/popliteal junctions and the
calf show significantly greater associations with the pres-
ence of a concomitant DVT [17,49]. This can be explained
by greater proximity to the deep venous system and by a
higher number of perforating veins at the calf level, respec-
tively [17,49]. Further supporting this hypothesis, we found
an approximately 5% higher prevalence of DVT in studies
in which a complete ultrasound examination from groin to
ankle was performed than in those that limiting the exami-
nation to the thigh. Similarly, different approaches used for
PE diagnosis might represent a relevant source of hetero-
geneity. Interestingly, we found that the prevalence of PE
in SVT patients was 6.4% when only symptomatic patients
were screened. The prevalence increased to 17% in the
studies that also systematically evaluated asymptomatic
patients. However, the number of studies included in this
subanalysis is very limited, and caution is necessary in the
interpretation of these results.
Our study has some potential limitations. First, the stud-
ies included in our meta-analysis had different inclusion
and exclusion criteria, and there was significant heterogene-
ity among the studies. Although we were not always able to
ascertain the source of heterogeneity, and the exclusion of
one study at a time did not reduce the heterogeneity, we
performed several sensitivity and subgroup analyses to
address this issue. Interestingly, we found that study design
© 2016 International Society on Thrombosis and Haemostasis

was a significant factor impacting on the heterogeneity of
results.
Furthermore, there are some major differences in the
prevalence of concomitant risk factors for venous throm-
bosis, and, with the meta-regression approach, we were
able to adjust our results for only some potential con-
founders. Thus, extreme caution is necessary in overall
result interpretation.
Finally, the prevalence of DVT found in our study
(18%) was slightly lower than in the three largest studies
on this topic (25%, 29%, and 24.6%, respectively)
[4,5,21]. The reason for the different results may originate
from differences in patient selection. Indeed, in the OPTI-
MEV, POST and STEPH studies, both inpatients and
outpatients were considered, whereas the large majority
of studies included in the present meta-analysis included
only ambulatory outpatients.
In conclusion, the results of our large meta-analysis
suggest that the prevalence of concomitant DVT and PE
in patients presenting with SVT is not negligible. Screen-
ing for a major thromboembolic event might be worth-
while in some ‘high-risk’ SVT patients, to allow adequate
anticoagulant treatment to be planned. However, further
high-quality studies are needed to confirm our findings,
and to identify clinical and demographic predictors of the
presence of DVT/PE in SVT patients.
Addendum
M. N. D. Di Minno, P. Ambrosino, and F. Dentali con-
ceived and designed the study, performed statistical analy-
sis, interpreted results, and drafted the manuscript.
F. Ambrosini, E. Tremoli, and G. Di Minno acquired
clinical data, drafted the manuscript, and performed criti-
cal revisions. All authors read and approved the final ver-
sion of the manuscript. M. N. D. Di Minno, P.
Ambrosino and F. Dentali had full access to all data in
the study, and take responsibility for the integrity of the
data and the accuracy of data analysis.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Fig. S1. PRISMA flow diagram.
Fig. S2. Meta-regression analysis. Effect of varicose vein,
malignancy, obesity and surgery on the prevalence of
deep vein thrombosis (DVT)/pulmonary embolism (PE)
in patients with superficial vein thrombosis (SVT).
Fig. S3. Funnel plots of effect size versus standard error
for studies evaluating the prevalence of deep vein throm-
© 2016 International Society on Thrombosis and Haemostasis
Superficial vein thrombosis and VTE 971
bosis (A) and pulmonary embolism (B) in patients with
superficial vein thrombosis.
Table S1. Techniques used for the detection of lower limb
venous thrombosis and pulmonary embolism in the
included studies.
References
1 Decousus H, Leizorovicz A. Superficial thrombophlebitis of
the legs: still a lot to learn. J Thromb Haemost 2005; 3: 1149–
51.
2 Decousus H, Epinat M, Guillot K, Quenet S, Boissier C, Tardy
B. Superficial vein thrombosis: risk factors, diagnosis, and treat-
ment. Curr Opin Pulm Med 2003; 9: 393–7.
3 Di Minno G, Mannucci PM, Tufano A, Palareti G, Moia M,
Baccaglini U, Rudelli G, Giudici GA; First Ambulatory Screen-
ing On Thromboembolism (FAST) Study Group.The first ambu-
latory screening on thromboembolism: a multicentre, cross-
sectional, observational study on risk factors for venous throm-
boembolism. J Thromb Haemost 2005; 3: 1459–66.
4 Frappe P, Buchmuller-Cordier A, Bertoletti L, Bonithon-Kopp
C, Couzan S, Lafond P, Leizorovicz A, Merah A, Presles E,
Preynat P, Tardy B, Decousus H; STEPH Study Group. Annual
diagnosis rate of superficial vein thrombosis of the lower limbs:
the STEPH community-based study. J Thromb Haemost 2014;
12: 831–8.
5 Decousus H, Quere I, Presles E, Becker F, Barrellier MT, Cha-
nut M, Gillet JL, Guenneguez H, Leandri C, Mismetti P, Pichot
O, Leizorovicz A; POST (Prospective Observational Superficial
Thrombophlebitis) Study Group. Superficial vein thrombosis and
venous thromboembolism: a large prospective epidemiological
study. Ann Intern Med 2010; 152: 218–24.
6 Milio G, Siragusa S, Mina C, Amato C, Corrado E, Grimaudo
S, Novo S. Superficial venous thrombosis: prevalence of common
genetic risk factors and their role on spreading to deep veins.
Thromb Res 2008; 123: 194–9.
7 Decousus H, Frappe P, Accassat S, Bertoletti L, Buchmuller A,
Seffert B, Merah A, Becker F, Quere I, Leizorovicz A. Epidemi-
ology, diagnosis, treatment and management of superficial-vein
thrombosis of the legs. Best Pract Res Clin Haematol 2012; 25:
275–84.
8 Cannegieter SC, Horvath-Puh o E, Schmidt M, Dekkers OM,
Pedersen L, Vandenbroucke JP, Sørensen HT. Risk of venous
and arterial thrombotic events in patients diagnosed with superfi-
cial vein thrombosis: a nationwide cohort study. Blood 2015;
125: 229–35.
9 Verlato F, Zucchetta P, Prandoni P, Camporese G, Marzola
MC, Salmistraro G, Bui F, Martini R, Rosso F, Andreozzi GM.
An unexpectedly high rate of pulmonary embolism in patients
with superficial thrombophlebitis of the thigh. J Vasc Surg 1999;
30: 1113–15.
10 Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred report-
ing items for systematic reviews and meta-analyses: the PRISMA
statement. PLoS Med 2009; 6: e1000097.
11 Juni P, Witschi A, Bloch R, Egger M. The hazards of scoring
the quality of clinical trials for meta-analysis. JAMA 1999; 282:
1054–60.
12 Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring
inconsistency in meta-analyses. BMJ 2003; 327: 557–60.
13 Sterne JA, Egger M, Smith GD. Systematic reviews in health
care: investigating and dealing with publication and other biases
in meta-analysis. BMJ 2001; 323: 101–5.
14 Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based
method of testing and adjusting for publication bias in meta-ana-
lysis. Biometrics 2000; 56: 455–63.
972 M. N. D. Di Minno et al
15 Ascher E, Hanson JN, Salles-Cunha S, Hingorani A. Lesser
saphenous vein thrombophlebitis: its natural history and implica-
tions for management. Vasc Endovascular Surg 2003; 37: 421–7.
16 Bergqvist D, Jaroszewski H. Deep vein thrombosis in patients
with superficial thrombophlebitis of the leg. Br Med J (Clin Res
Ed) 1986; 292: 658–9.
17 Binder B, Lackner HK, Salmhofer W, Kroemer S, Custovic J,
Hofmann-Wellenhof R. Association between superficial vein
thrombosis and deep vein thrombosis of the lower extremities.
Arch Dermatol 2009; 145: 753–7.
18 Bl€attler W, Frick E. [Complications of superficial throm-
bophlebitis]. Schweiz Med Wochenschr 1993; 123: 223–8.
19 Blumenberg RM, Barton E, Gelfand ML, Skudder P, Brennan J.
Occult deep venous thrombosis complicating superficial throm-
bophlebitis. J Vasc Surg 1998; 27: 338–43.
20 Bounameaux H, Reber-Wasem MA. Superficial thrombophlebitis
and deep vein thrombosis. A controversial association. Arch
Intern Med 1997; 157: 1822–4.
21 Galanaud JP, Genty C, Sevestre MA, Brisot D, Lausecker M,
Gillet JL, Rolland C, Righini M, Leftheriotis G, Bosson JL,
Quere I; OPTIMEV SFMV investigators. Predictive factors for
concurrent deep-vein thrombosis and symptomatic venous
thromboembolic recurrence in case of superficial venous throm-
bosis. The OPTIMEV study. Thromb Haemost 2011; 105: 31–9.
22 Gorty S, Patton-Adkins J, DaLanno M, Starr J, Dean S, Satiani
B. Superficial venous thrombosis of the lower extremities: analy-
sis of risk factors, and recurrence and role of anticoagulation.
Vasc Med 2004; 9: 1–6.
23 Hirmerova J, Seidlerova J, Subrt I. Deep vein thrombosis and/or
pulmonary embolism concurrent with superficial vein thrombosis
of the legs: cross-sectional single center study of prevalence and
risk factors. Int Angiol 2013; 32: 410–16.
24 Lutter KS, Kerr TM, Roedersheimer LR, Lohr JM, Sampson
MG, Cranley JJ. Superficial thrombophlebitis diagnosed by
duplex scanning. Surgery 1991; 110: 42–6.
25 Noppeney T, Noppeney J, Winkler M, Kurth I. Acute superficial
thrombophlebitis – therapeutic strategies. Zentralbl Chir 2006;
131: 51–6.
26 Pomero F, Di Minno MN, Tamburini Premunian E, Malato A,
Pasca S, Barillari G, Fenoglio L, Siragusa S, Di Minno G,
Ageno W, Dentali F. A clinical score to rule out the concomitant
presence of deep vein thrombosis in patients presenting with
superficial vein thrombosis: the ICARO study. Thromb Res 2015;
15: 30121–3.
27 Prountjos P, Bastounis E, Hadjinikolaou L, Felekuras E, Balas
P. Superficial venous thrombosis of the lower extremities co-
existing with deep venous thrombosis. A phlebographic study on
57 cases. Int Angiol 1991; 10: 63–5.
28 Pulliam CW, Barr SL, Ewing AB. Venous duplex scanning in
the diagnosis and treatment of progressive superficial throm-
bophlebitis. Ann Vasc Surg 1991; 5: 190–5.
29 Rathbun SW, Aston CE, Whitsett TL. A randomized trial of
dalteparin compared with ibuprofen for the treatment of superfi-
cial thrombophlebitis. J Thromb Haemost 2012; 10: 833–9.
30 Skillman JJ, Kent KC, Porter DH, Kim D. Simultaneous occur-
rence of superficial and deep thrombophlebitis in the lower
extremity. J Vasc Surg 1990; 11: 818–23.
31 Sobreira ML, Maffei FH, Yoshida WB, Rollo HA, Last oria S,
Griva BL, De Carvalho LR. Prevalence of deep vein thrombosis
and pulmonary embolism in superficial thrombophlebitis of the
lower limbs: prospective study of 60 cases. Int Angiol 2009; 28:
400–8.
32 Yucel EK, Egglin TK, Waltman AC. Extension of saphenous
thrombophlebitis into the femoral vein: demonstration by color
flow compression sonography. J Ultrasound Med 1992; 11: 285–
7.
33 Wichers IM, Haighton M, B€ uller HR, Middeldorp S. A retro-
spective analysis of patients treated for superficial vein thrombo-
sis. Neth J Med 2008; 66: 423–7.
34 Wichers IM, Di Nisio M, B€ uller HR, Middeldorp S. Treatment
of superficial vein thrombosis to prevent deep vein thrombosis
and pulmonary embolism: a systematic review. Haematologica
2005; 90: 672–7.
35 Galanaud JP, Bosson JL, Genty C, Presles E, Cucherat M,
Sevestre MA, Quere I, Decousus H, Leizorovicz A. Superficial
vein thrombosis and recurrent venous thromboembolism: a
pooled analysis of two observational studies. J Thromb Haemost
2012; 10: 1004–11.
36 Coon WW, Willis PW, Keller JB. Venous thromboembolism and
other venous disease in the Tecumseh community health study.
Circulation 1973; 48: 839–46.
37 Marchiori A, Mosena L, Prandoni P. Superficial vein thrombo-
sis: risk factors, diagnosis and treatment. Semin Thromb Hemost
2006; 32: 737–43.
38 Leon L, Giannoukas AD, Dodd D, Chan P, Labropoulos N.
Clinical significance of superficial vein thrombosis. Eur J Vasc
Endovasc Surg 2005; 29: 10–17.
39 Samama MM, Dahl OE, Quinlan DJ, Mismetti P, Rosencher N.
Quantification of risk factors for venous thromboembolism: a
preliminary study for the development of a risk assessment tool.
Haematologica 2003; 88: 1410–21.
40 Totten HP. The surgical treatment of acute ascending superficial
thrombophlebitis. Angiology 1965; 16: 37–42.
41 Chengelis DL, Bendick PJ, Glover JL, Brown OW, Ranval TJ.
Progression of superficial venous thrombosis to deep vein throm-
bosis. J Vasc Surg 1996; 24: 745–9.
42 Kearon C, Akl EA, Comerota AJ, Prandoni P, Bounameaux H,
Goldhaber SZ, Nelson ME, Wells PS, Gould MK, Dentali F,
Crowther M, Kahn SR; American College of Chest Physicians.
Antithrombotic Therapy for VTE Disease: Antithrombotic Ther-
apy and Prevention of Thrombosis, 9th ed: American College of
Chest Physicians. Evidence-Based Clinical Practice Guidelines.
Chest 2012; 141: e419S–94S.
43 Decousus H, Prandoni P, Mismetti P, Bauersachs RM, Boda Z,
Brenner B, Laporte S, Matyas L, Middeldorp S, Sokurenko G,
Leizorovicz A; CALISTO Study Group. Fondaparinux for the
treatment of superficial-vein thrombosis in the legs. N Engl J
Med 2010; 363: 1222–32.
44 Cosmi B. Management of superficial vein thrombosis. J Thromb
Haemost 2015; 13: 1175–83.
45 Decousus H, Bertoletti L, Frappe P. Spontaneous acute superfi-
cial vein thrombosis of the legs: do we really need to treat? J
Thromb Haemost 2015; 13(Suppl. 1): S230–7.
46 Goodacre S, Sampson F, Thomas S, van Beek E, Sutton A. Sys-
tematic review and meta-analysis of the diagnostic accuracy of
ultrasonography for deep vein thrombosis. BMC Med Imaging
2005; 5: 6.
47 Le Gal G, Righini M, Wells PS. Computed tomographic pul-
monary angiography for pulmonary embolism. JAMA 2015; 314:
74–5.
48 Quere I, Leizorovicz A, Galanaud JP, Presles E, Barrellier MT,
Becker F, Desprairies G, Guenneguez H, Mismetti P, Decousus
H; Prospective Observational Superficial Thrombophlebitis
(POST) Study Investigators. Superficial venous thrombosis and
compression ultrasound imaging. J Vasc Surg 2012; 56: 1032–8.
49 Gillet JL, Allaert FA, Perrin M. Superficial thrombophlebitis in
non-varicose veinsof the lower limbs. A prospective analysis in
42 patients. J Mal Vasc 2004; 29: 263–72.
© 2016 International Society on Thrombosis and Haemostasis