We and many others have previously published on the potential association of venous
thromboembolism (VTE) with systemic inflammation and viral/bacterial infections in
this journal.[1]
[2]
[3] In particular, several studies demonstrated that hepatitis viruses such as cytomegalovirus
(CMV), Epstein–Barr virus (EBV), hepatitis A, B, and C viruses (HAV, HBV, HCV) are
able to induce hemostatic abnormalities and increase the risk of venous thromboses,
including splanchnic vein thrombosis (SVT).[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9] Episodes of SVT, defined as the occlusion of veins of the portal venous system (including
the mesenteric, splenic, and portal veins) or the hepatic veins (Budd–Chiari's syndrome,
BCS), usually occur in patients with intra-abdominal inflammatory processes, cirrhosis,
cancer and/or prothrombotic disorders.[10]
Acute CMV infection has been reported to associate with VTE in immunocompromised patients
(HIV-positive or posttransplant patients). However, VTE during acute CMV infection
is increasingly reported even in immunocompetent adults. The sites of CMV-associated
thrombosis in immunocompetent subjects include pulmonary embolism (PE), lower limb
deep vein thrombosis (DVT), and SVT (portal vein thrombosis [PVT], mesenteric vein
thrombosis [MVT], and BCS).[5]
[9] Other reported VTE sites include inferior vena cava, and ovarian, iliac, and jugular
veins.[7]
CMV is a herpesvirus that may be asymptomatic or may cause few symptoms. After the
first infection, CMV often becomes latent, but immunosuppression may allow the virus
to reactivate and become symptomatic. Acute CMV infection manifests with a mononucleosis-like
syndrome, with fever, cervical lymphadenitis and arthralgia, more rarely with pneumonia,
hepatitis, myocarditis, pericarditis, colitis, or hemolytic anemia.[5]
CMV increases the thrombotic risk in immunocompetent adults especially when the acute
infection is associated with concomitant prothrombotic conditions such as oral contraceptive
(OC) use, prolonged immobilization, malignancies, trauma, or inherited thrombophilic
conditions.[7]
[8]
We therefore wish to report two recent cases of young immunocompetent patients referred
to our tertiary care center who experienced SVT associated with acute CMV infection.
We follow these case descriptions with a brief narrative review focused on the available
literature evidence especially with respect to: (1) clinical presentation and risk
factors of events, (2) mechanisms involved in the thrombotic process in immunocompetent
patients with acute CMV infection, and (3) management of these thrombotic events.
Case I: A 40-year-old man was admitted to our center because of fever, abdominal pain, and
diarrhea. He had been treated ineffectively with antibiotic drugs in the previous
weeks. His past clinical history included arterial hypertension and obesity, and he
was a carrier of β-thalassemia. On admission, the patient had persistent fever (>38°C)
and pain in the right hip. No superficial lymphadenopathy on physical examination
was detected. Blood tests revealed leukopenia with relative lymphocytosis and elevated
liver cytolysis indices, as well as increased values of lipase, amylase, lactate dehydrogenase,
and D-dimer. Chest X-ray and electrocardiogram were negative. A computed tomography
(CT) scan of his abdomen demonstrated an extensive thrombus of splenic vein and superior
MVT. The screening for inherited/acquired thrombophilia, including the search for
factor V Leiden (FVL) and prothrombin (FII) G20210A gene polymorphisms, the detection
of plasma anticardiolipin antibodies and lupus anticoagulant, and measurements of
plasma antithrombin, protein C, protein S, and homocysteine were negative. Hematological
(negative bone marrow aspirate and JAK2 mutation) or clear-cut neoplastic diseases were excluded. Serological tests for HIV,
hepatitis (A, B, C, and E), toxoplasma, and EBV were normal, whereas CMV serology
revealed high-titer immunoglobulin M (IgM) antibodies (100 IU/mL, normal values < 12
IU/mL) with negative immunoglobulin G (IgG) antibodies, and a low IgG avidity index:
0.122 (reference range < 0.2: low), thus suggesting acute infection. The patient was
put on anticoagulation with therapeutic doses of low molecular weight heparin (LMWH)
(enoxaparin 100 IU/kg twice daily) and then warfarin. The systemic symptoms gradually
resolved without the need for ganciclovir treatment. After 6 months, a complete resolution
of the thrombus was shown at the CT scan, and warfarin therapy was discontinued.
Case II: A 24-year-old female presented to the emergency department with fever and abdominal
pain for the past week. She was a carrier of β-thalassemia, was obese, and reported
a history of uterine myomas and gallbladder stones. On this basis, a biliary colic
was first hypothesized, not confirmed at ultrasound evaluation, while an urgent CT
scan of abdomen demonstrated thrombosis of the inferior mesenteric vein, without extension
to portal and splenic veins. Use of OCs was denied, hematological and neoplastic causes
were excluded, and no other cardiovascular risk factors were present. Serological
tests for HIV, hepatitis (A, B, C, and E), toxoplasma, and EBV were negative, whereas
CMV serology revealed high-titer IgM antibodies (129 IU/mL, normal values < 12 IU/mL)
with negative IgG antibodies, consistent with acute infection. She was treated with
low-dose LMWH (enoxaparin 4,000 IU once daily) for 2 to 3 weeks. Four years later,
the patient was referred to us for counseling about pregnancy, and at this time, a
heterozygous FII G20210A polymorphism was identified. A CT scan of her abdomen showed
resolution of the previous thrombosis. Due to the presence of thrombophilia and obesity,
she received an intermediate-dose enoxaparin prophylaxis (100 IU/kg once daily) during
pregnancy, without thromboembolic recurrences.
CMV infection was the most likely precipitating factor for SVT in these two patients
because no other cause/disease could explain the clinical and laboratory presentation.
In both the cases, fever and abdominal pain were detectable. Anticoagulant therapy,
although conducted in the second patient at low dose and only for 2 to 3 weeks, resulted
in complete resolution of thrombosis.
[Table 1] summarizes the largest studies and reviews reflecting on the association between
CMV infection and SVT. Risk factors and clinical characteristics of hospitalized patients
with acute CMV infection-associated thrombosis were evaluated in 2010 in a retrospective
case–control study[11] including 140 consecutive patients (mean age: 37.3 ± 17.4 years) diagnosed with
the acute infection in a tertiary medical center and 140 matched controls without
CMV infection. Among the control group, none of the patients had thrombotic events,
whereas among the study group, nine (6.4%; p = 0.003) patients had thromboses: five arterial thromboses (four splenic infarcts
and one renal infarct) and four venous thromboses (one PE, one lower limb DVT, one
upper limb DVT, and one jugular vein thrombosis). Concomitant predisposing conditions
for thrombosis (OCs, immobilization, recent surgery), other than the infection, were
present in 6/9 (66.6%) cases. Binary logistic regression analysis showed that acute
CMV infection was independently associated with thrombosis in the whole cohort (p = 0.004), and the use of OCs/hormones or pregnancy were independently associated
with thrombosis among patients with acute infection (p = 0.043).[11]
Table 1
The largest studies contributing information about the association between CMV infection
and SVT
|
Authors, year
|
Design
|
Number of cases with acute CMV and thrombosis
|
Number of female patients with thrombosis
|
Age, mean
(range)
|
Thrombosis sites
|
Risk factors
|
Antithrombotic treatment (when described)
|
Antiviral treatment
|
|
Atzmony et al.,
2010[11]
|
Retrospective
case–control
|
9/140 (6.4%) patients with acute CMV infection (3 immunocompetent and 6 immunocompromised)
vs. 0/140 controls
|
2/9
|
47.1 (29–65)
|
1 PE
1 lower limb DVT
1 upper limb DVT
1 jugular vein
4 splenic infarcts
1 renal infarct
|
Presence of acquired or inherited conditions in 66.6%
|
All anticoagulated
|
Five patients treated
|
|
Tichelaar et al.,
2011[12]
|
Case–control
|
5/258 (1.9%) cases with active CMV vs. 0/139 controls
|
5/5
|
32.4 (26–36)
|
3 DVT
2 PE
|
3 OC
1 surgery
1 puerperium
|
NA
|
NA
|
|
Schimanski et al.,
2012[13]
|
Case–control
|
CMV IgM positivity in 7/166 (4.2%) immunocompetent VTE patients vs. 0.6% in controls
(no VTE)
|
6/7
|
43.4 (29–61)
|
6 DVT
1 DVT + PE
|
Presence of acquired RF in 3/7 (42.89%) cases
Presence of thrombophilia in 5/7 (71.4%) cases
|
NA
|
NA
|
|
Justo t al.,
2011[6]
|
Meta-analysis
|
97 (64 immunocompetent and 33 immunocompromised)
|
46/97
|
39.7 ± 14.9 (2–83)
|
52 DVT/PE
25 SVT
12 splenic infarction
Others veins (4 internal jugular, 2 CVT, 1 ovarian, 1 extra-hepatic, 1 brachial vein,
1 azygos)
2 renal artery thrombosis
1 MI
1 digital ischemia
|
16.5% OC
14.4% APLA
37.1% acquired conditions
17.5% inherited RFs
|
Heparin/warfarin in most cases (numbers not available)
Three thrombolysis
|
28
(28.9%) patients treated
|
|
Paran et al.,
2013[14]
|
Community prospective study
|
6 mo incidence of VTE: 19 of 6,205 (0.3%) patients;
6 mo incidence of arterial thrombosis: 7 of 6,222 (0.11%) patients after CMV infection
vs. 84,310 controls
|
63.7% in the whole VTE cohort
64.4% in the whole arterial cohort
|
42 ± 11 (mean of the entire cohort)
(range 30–98)
|
12 lower limb DVT
4 PE
2 upper limb DVT
1 superficial VT
7 arterial thromboses
|
NA
|
NA
|
NA
|
|
Sherman et al.,
2014[15]
|
Narrative review
|
113
|
Female:male: 1/1
|
41.7 ± 14.6
(mean)
(17–83)
|
63 DVT/PE
31 SVT
2 renal artery
2 strokes
1 MI
1 digital ischemia
|
60.2% of cases presented transient or chronic RF
15% presented OC
10.6% presented factor V Leiden
|
All anticoagulated
|
34 patients treated
|
|
Yildiz et al.,
2016[16]
|
Retrospective
|
10 immunocompetent
|
9/10
|
37.5
(mean)
|
3 superficial VT
5 DVT (2 lower limb)
4 with PE associated
3 SVT
|
100% of cases presented at least one thrombophilic condition
50% presented OC use
|
All anticoagulated
|
NA
|
|
Atzmony et al.,
2010[18]
|
Review
|
39 immunocompetent
|
22/39
|
34.9 ± 10.8 (17–67)
|
21 SVT
16 VTE
1 ovarian vein
1 digital ischemia
|
65.7% presented one or more RF (acquired or inherited)
34.3% presented no RF associated
|
NA
|
NA
|
|
Deconinck et al.,
2016[17]
|
Review
|
69 immunocompetent
|
41/69
|
35 (median)
(17–83)
|
30 SVT
18 lower limbs
Others (CVT, ovarian, internal jugular, vena cava)
(25 cases associated with PE)
1 TIA
1 SMA thrombosis
1 digital ischemia
|
76%: presence of RF (58% thrombophilia, 42% acquired risk factors, 38% with several
RF)
22%: no RF associated
|
48 anticoagulation (heparin and/or VKA)
4 thrombolysis
2 aspirin
2 no treatment
|
3 ganciclovir
3 valganciclovir
6 ganciclovir followed by valganciclovir
|
Abbreviations: APLA, antiphospholipid antibody; CMV, cytomegalovirus; CVT, cerebral
vein thrombosis; DVT, deep venous thrombosis; IgM, immunoglobulin M; MI, myocardial
infarction; NA, not available; OC, oral contraceptive; PE, pulmonary embolism; RF,
risk factor; SMA, superior mesenteric artery; SVT, splanchnic vein thrombosis; TIA,
transient ischemic attack; VKA, vitamin K antagonist; VT, venous thrombosis.
In another case–control study, five cases of VTE (three DVT and two PE) and active
CMV infection were detected in hospitalized patients, all females, aged below 37 years
and showing at least one concomitant VTE acquired risk factor (three during OC treatment,
one after surgery, and one in puerperium).[12] In a case–control study that investigated the potential role of CMV in hospitalized
immunocompetent patients, a higher frequency of positive CMV–IgM and CMV–IgG was observed
in VTE patients than in controls.[13] Seven patients with VTE had CMV–IgM positivity (six women and one man); in two cases,
the thrombosis occurred during OC therapy and in one case during pregnancy.[13]
Clinical characteristics of patients with CMV-associated thrombosis were also evaluated
in a meta-analysis, published in 2011, that included 97 reports.[6] Overall, 64 immunocompetent patients and 33 immunocompromised patients were described,
with a mean age of 39.7 ± 14.9 years and a female-to-male ratio of 1:1, although females
were prevalent among the immunocompetent cases. The most prevalent clinical presentations
of infection were CMV mononucleosis and CMV colitis. The most frequent sites of thrombosis
were lower limbs (DVT) and/or pulmonary circulation (PE) (52/97; 53.6%), followed
by SVT (PVT, superior and inferior MVT, colic vein thrombosis) and splenic infarction.
The episodes of DVT or PE were more frequent in immunocompromised patients, whereas
SVTs were more prevalent among immunocompetent cases. Arterial thromboses (renal infarct,
myocardial infarction, digital ischemia) were rare.[6] In the meta-analysis, the use of OCs, transient/permanent antiphospholipid antibody
(APLA) positivity, and FVL heterozygosis were the most common concomitant predisposing
conditions for thrombosis. Other predisposing conditions were surgery, active malignancy,
and pregnancy.[6] Interestingly, the prevalence of inherited thrombophilia was significantly higher
among immunocompetent patients compared with immunocompromised patients,[6] thus suggesting a prothrombotic state associated with the condition of immunosuppression
of these patients to be sufficient to induce the thrombosis.
Paran et al in 2013 investigated, for the first time, the long-term effect (6-month
incidence) on VTE and/or arterial thrombosis in 6,205 outpatients who tested positive
for CMV–IgM antibodies in a large health maintenance organization in Israel.[14] During 6 months of follow-up from index date (the earliest CMV–IgM testing date),
the incidence rates of VTE among CMV–IgM seropositive were higher compared with CMV–IgM
seronegative patients (odds ratio [OR]: 2.25; 95% confidence interval [CI]: 1.38–3.66;
p = 0.003). The CMV–IgM seropositivity was independently associated with VTE in multivariable
logistic regression analysis (OR: 2.49; 95% CI: 1.53–4.06; p < 0.0001), but it was not associated with arterial thrombosis.[14]
Clinical characteristics of CMV-related thrombosis have been reviewed by Sherman et
al,[15] who described 78 reports concerning 113 patients. The mean age of patients was 41 ± 14.6
with a female-to-male ratio of 1:1. The female-to-male ratio was reanalyzed separately
in controlled studies, including consecutive patients, and was 1.6.[15]
Recently, Yildiz et al identified 10 patients with synchronous acute CMV infection
among 1,007 consecutive patients with VTE. Patients with coexistent VTE and acute
CMV infection were younger (37.5 vs. 56.6 years; p = 0.0088) and exhibited a female predominance (90 vs. 56%; p = 0.026) in comparison to the whole cohort. Hereditary thrombophilia was identified
in 9 of 10 patients. Acquired risk factors for VTE such as OCs and pregnancy were
equally prevalent in patients with and without CMV infection.[16]
In another recent literature review,[17] 49 papers with a total of 69 patients (41 women and 28 men) with VTE complicating
CMV primary infection were analyzed. The main sites of venous thrombosis were SVT
(30 patients) and DVT of lower limbs (18 patients). One-third of patients presented
with PE (25 patients). Forty-nine (76%) of the 64 patients screened for thrombophilia
had at least one VTE risk factor, inherited or acquired thrombophilia for 37 (58%)
patients, and another risk factor for 27 (42%) patients.[17] Anticoagulant treatment was described for 48 patients, and treatment duration reported
in 20 cases was 3 to 6 months. Anticoagulation was discontinued without any radiological
evidence of thrombosis resolution for most patients.
The potential mechanisms of thrombosis in CMV infection deserve mention. Venous thrombosis
is a multifactorial disease that can be triggered by environmental factors and inherited
predisposition. Based on data reported in the literature, it has been suggested to
include CMV infection among VTE risk factors/triggers.[4]
[15]
[18]
[19] However, unlike immunosuppressed patients, in whom infection is often the only condition
associated with thrombosis, inherited thrombophilia and acquired risk factors for
thrombosis often coexist in immunocompetent patients with SVT.[6]
[18]
In a review of case reports and case series on CMV-associated thrombosis in immunocompetent
adults,[18] 39 patients were described (mean age: 34.9 ± 10.8 years). Most events (21/39; 54%)
were SVT. Overall, 14 (35.9%) patients had one or more acquired risk factors for thrombosis;
16/35 (45.7%) patients investigated for inherited thrombophilia had one or more inherited
predispositions for thrombosis. Only 12 (34.3%) patients did not show any acquired
or inherited predisposition for thrombosis other than CMV infection. The most common
(n = 13; 33.3%) acquired predisposition for thrombosis was the use of OCs, whereas FVL
was the most frequently detected inherited thrombophilia factor (n = 6; 17.1%)[18] ([Table 1]). The relatively higher frequency of SVT rather than DVT/PE found in immunocompetent
patients with CMV infection[6]
[18] may be explained by a role of acute CMV hepatitis as a local inflammatory predisposing
condition.[4]
The mechanisms by which CMV infection may induce thrombosis are varied ([Table 2]). The most accepted theories suggest that CMV infects endothelial cells, thus inducing
membrane alteration and expression of adhesion molecules causing platelet and leukocyte
adhesion and generating tissue factor (TF) exposure, factor X activation, and thrombin
formation.[5]
[19]
[20] Intrinsic CMV envelope procoagulant properties may be another cause of hemostatic
imbalance, and furthermore, TF is present on infected monocytes.[19] Acute CMV infection is reported to increase circulatory levels of von Willebrand
factor and factor VIII.[5]
[18]
[19] Finally, acute CMV infection has been associated with transient appearance of APLA.[19]
[20]
[21]
[22]
[23]
[24]
Table 2
CMV infection and thrombosis: underlying mechanisms
|
Endothelial dysfunction (increased tissue factor exposure)
|
|
Platelet and leukocyte adhesion
|
|
Increased tissue factor exposure in infected monocytes
|
|
Procoagulant properties of CMV envelope
|
|
Increased factor VIII and von Willebrand factor levels
|
|
Transient antiphospholipid antibodies
|
Abbreviation: CMV, cytomegalovirus.
No guideline on management of CMV-associated venous thrombosis is currently available.
In particular, there are no agreed recommendations on the duration of anticoagulation
and benefit of antiviral therapy. The majority of CMV-associated venous thromboses
reported in the literature were treated with LMWH and warfarin, with the duration
of oral anticoagulation ranging between 20 days and 9 months.[20] Therapy was stopped when the resolution of thrombosis was confirmed or, in some
cases, on the basis of reduction or absence of APLA.[6]
[20]
Three months of anticoagulant therapy seems to be the most appropriate strategy since
CMV infection is a transient and reversible risk factor, according to the current
guidelines of the American College of Chest Physicians.[25]
[26] However, the search for inherited thrombophilia and APLA should guide the choice
of extended treatment.
The benefit of primary anticoagulant prophylaxis in patients with acute CMV infection
is not defined and should be considered in the presence of multiple concomitant prothrombotic
conditions (i.e., hospitalized patients with immobilization or a history of VTE).[17]
As regards CMV antiviral therapy, in the recent review by Sherman et al,[15] 30.1% of the total population of patients studied were treated with antiviral agents
(ganciclovir or valganciclovir) and 73.5% of these had viremia diagnosed by CMV–DNA
PCR (deoxyribonucleic acid polymerase chain reaction) or antigen assays.[15] In the meta-analysis by Justo et al, only 17.2% of immunocompetent patients received
antiviral agents compared with 51.5% in the immunocompromised group.[6] In the previously cited literature review,[17] regarding 49 papers with a total of 69 patients with VTE complicating CMV primary
infection, only 11 patients (all cases with severe thromboembolic events) received
antiviral therapy (ganciclovir and/or valganciclovir). A positive outcome was observed
in all patients, with the exception of a case of BCS. The authors suggested that antiviral
therapy should be considered for patients presenting with severe VTE, VTE with a negative
outcome despite anticoagulation, and severe organ involvement or for patients managed
in the intensive care unit. They also suggested a practical antithrombotic strategy
for CMV- associated thrombosis[17] ([Table 3]).
Table 3
Suggested strategies for CMV-associated venous thromboembolism
|
Acute CMV infection + no RF
|
No anticoagulation
|
|
Acute CMV infection + ≥ 1 RF
|
Preventive anticoagulation
|
|
Acute CMV infection + VT
|
VKA, INR 2–3, 3 mo
|
|
Acute CMV infection + VT + ≥ 1 irreversible RF
|
VKA, INR 2–3, at least 3 mo, discuss life-long treatment
|
Abbreviations: CMV, cytomegalovirus; INR, international normalized ratio; mo, month;
RF, risk factors for venous thromboembolism; VKA, vitamin K antagonist; VT, vascular
thrombosis.
Note: Modified from Deconinck et al.[17]
In conclusion, the number of reported SVT in patients with acute CMV infections has
increased considerably in recent years, mainly due to the growing awareness of the
existence of an association between thrombotic events and this infection. The frequency
of SVT is higher than that of DVT/PE in immunocompetent patients with CMV-associated
thrombosis in whom concomitant acquired/inherited predisposing factors other than
the infection are often recognized.
On the basis of literature data, acute CMV infection should be considered as a possible
prothrombotic condition and searched for in adult immunocompetent patients with SVT
in the presence of symptoms and signs of the infection. The search for concomitant
thrombotic risk factors and inherited/acquired thrombophilic conditions may better
guide the optimal management strategy in the absence of specific recommendations or
guidelines.
