Active cytomegalovirus infection in patients with acute venous thrombosis: A case‐control study

Abstract

Infectious diseases have been associated with venous thrombosis (VT) [1]. Several case-studies described active cytomegalovirus infection in patients with VT [2-6]. In this case-control study, we detected five cases with active cytomegalovirus infection and VT, and none in the control group. All were female, below 37 years of age, and had another acquired risk factor for VT. This may imply that active cytomegalovirus infection contributes to a procoagulant state in young subjects with other mild risk factors for VT. Recent prospective cohort studies showed that an increase of the inflammatory marker C-reactive protein (CRP) was associated with a 1.5–2.5-fold increased risk of VT [7, 8]. Interestingly, this risk was highest when CRP levels were elevated closely before the onset of VT [7]. The latter observation supports the hypothesis that activation of the coagulation system is stimulated by an infectious process [9, 10]. It is possible that the procoagulant response induced by active cytomegalovirus infection [6, 11-13], in some persons exceeds the thrombosis threshold leading to VT [14]. Although active cytomegalovirus infection may play a role in VT onset, it cannot be ruled out that case-reports reporting on this issue emphasized the rare exception. Therefore, we decided to conduct a case-control study in consecutive patients with suspicion of VT to determine the prevalence of acute cytomegalovirus infection and the concurrence of procoagulant abnormalities in patients with VT. Of 397 patients in the study, 258 (65%) were cases and 139 (35%) were controls (Table I).Compared to controls, patients with VT were more often male, had less often a history of VT, more often experienced infectious signs in the preceding 4 weeks, and had higher CRP, leukocytes and factor VIII levels. Female patients were more frequent on oral contraceptives. Seroprevalence of cytomegalovirus IgG was not higher in the case group (n = 126, 55% in cases, and n = 77, 60% in controls). Levels of cytomegalovirus IgM were found in 4% of cases and 4% of controls. In five patients with VT (2%) a positive viral load of cytomegalovirus DNA was detected, compared to none in the controls. All were younger than 37 years of age and female. As 31 of 258 patients with VT (12%) were younger than 37 years and 5 of these 31 patients had active cytomegalovirus infection, means that 16% of all VT patients younger than 37 years had an active cytomegalovirus infection (Table II). None had a history of VT, and all had an additional acquired VT risk factor. Compared to VT patients without active cytomegalovirus infection or controls, they were younger (median 36 vs. 57 years, P = 0.003 and 36 vs. 53 years, P = 0.005), appeared to have higher levels of CRP (median 31 vs. 22 mg/L, P = 0.419 and 31 vs. 5 mg/L, P = 0.006), leukocytes (median 9.0 vs. 8.5 × 10E9 cells/L, P = 0.75 and 9.0 vs. 7.7 × 10E9 cells/L, P = 0.81) and factor VIII (217 vs. 194 IU/dL, P = 0.392 and 217 vs. 156 IU/dL, P = 0.028). Our study demonstrates that simultaneous occurrence of active cytomegalovirus infection and VT is rare (prevalence 2%). VT is a multicausal disease, often involving acquired or environmental risk factors as well as genetic predisposition [12, 14]. This may explain why all five patients in this study with acute VT and active cytomegalovirus infection had at least one concomitant acquired risk factor for VT. A systematic review of 32 studies showed that patients with active cytomegalovirus infection and VT had acquired risk factors in nearly 40% of cases [13]. The difference between that review and our observation could be explained by the fact that not all of the studies that were reported in the review classified oral contraceptive use-associated VT as acquired VT. We have no clear explanation why all our patients with active cytomegalovirus infection were women. It could be that women are more at risk for VT then men at younger age, due to oral contraception use and being in the reproductive part of their live [14]. Also, women may be more exposed to cytomegalovirus than men, as in a general Western (US) population women had a higher seroprevalence of cytomegalovirus infection than men (64% vs. 54%) [15]. We found no difference in cytomegalovirus IgG levels between groups, suggesting that cytomegalovirus infections may only increase venous thrombotic risk at time of active infection, and not afterwards. This is in contrast with other studies, which suggested that previous cytomegalovirus infections may enhance atherosclerosis progression, even years after active infection [16]. The high prevalence of deep vein thrombosis in our study is a noteworthy finding itself; of 289 consecutive patients that were referred by their general practitioner to our emergency department with suspicion of deep vein thrombosis, 150/289 (52%) were diagnosed with this disease. This used to be around 30% in the Netherlands [17]. Our higher prevalence might be due to the fact that we are a university hospital, i.e. patients with a higher probability of VT may be more often referred to a university hospital than to a secondary care hospital. It might also reflect reality that clinical decision rules of ruling out deep vein thrombosis become more and more applied by family physicians within the Netherlands [18]. In conclusion, active cytomegalovirus infection is a rare phenomenon in patients with acute VT, but may occur more frequently in younger patients with VT. Seroprevalence of cytomegalovirus IgG was not associated with an increased risk of VT. From April 2008 till January 2010, consecutive patients with suspicion of deep vein thrombosis visiting our emergency department were included in our study by a physician of Hematology, after informed consent was taken. Cases were patients with objectively confirmed proximal deep vein thrombosis by compression ultrasound. Calf-vein thrombosis and thrombophlebitis were not considered deep vein thrombosis. Controls were patients with suspected deep vein thrombosis but who did not have the disease at (repeated) compression ultrasound. We also included consecutive patients with objectively confirmed pulmonary embolism by spiral computerized tomography, referred by physicians of the department of Pulmonology. Hence, we have no control subjects in which the suspicion of pulmonary embolism was raised and ruled out. A standardized questionnaire was used to ask about patients' characteristics [9]. Blood was taken by peripheral venipuncture at time of presentation. Viral load of cytomegalovirus was detected using polymerase chain reaction (PCR) for quantification of cytomegalovirus DNA [11]. A result of >100 copies/ml was defined as a positive viral load and as an active cytomegalovirus infection. Serology (IgM and IgG) of cytomegalovirus was measured in citrated plasma samples using colorimetric assays on an automated immunological analyzer (Abott Diagnostics, IL) according to standard instructions of the manufacturer. IgG levels of cytomegalovirus were categorized according to percentiles of the controls. We had missing data for IgG levels in 28 cases and 11 controls. CRP levels >5 mg/L were defined as elevated. Factor VIII:C levels were considered increased when >150 IU/dL. Factor V Leiden mutation and prothrombin G20210A mutation were detected with PCR methods. Data management and statistical analysis were performed using SPSS (version 16, SPSS inc. Illinois, Chicago). The authors thank Prof. Jan van der Meer, MD, PhD, who passed away in January 2009, for his important contribution to both the intellectual and practical side of the study. They also thank H. Marieke Knol, MD, Marion Noordzij, MD, and Karina Meijer, MD, PhD, for their contribution to data collection; Ruiqi Xue, PhD student and Annelies Riezebos-Brilman, MD, for their work on PCR and serology. The study was originally designed by Jan van der Meer and W.L. Y.T., A.M., and W.L. included patients in this study. Data analysis was performed by Y.T. and W.L. Y.T. and H.S. wrote the manuscript. H.K.-N. and W.L. reviewed the manuscript. H.N. performed the serology and PCR tests and contributed intellectually to the study design. Vladimir Y.I.G. Tichelaar [e-mail]*, Herman G. Sprenger , Anja B.U. Mäkelburg*, Bert G. M. Niesters , Hanneke C. Kluin-Nelemans*, Willem M. Lijfering*, * Department of Hematology, Division of Hemostasis and Thrombosis, University Medical Center Groningen, Groningen, The Netherlands, Department of Internal Medicine, Division of Infectious Diseases, University Medical Center Groningen, Groningen, The Netherlands, Department of Microbiology, University Medical Center Groningen, Groningen, The Netherlands.