Cytomegalovirus infection with pulmonary embolism, splenic vein thrombosis and monoclonal gammopathy of undetermined significance: a case and systematic review

Abstract

A 62-year-old immunocompetent woman was admitted with cytomegalovirus (CMV) infection, pulmonary embolism, splenic vein thrombosis and monoclonal gammopathy of undetermined significance (MGUS). Anticoagulation therapy was started. Two months later, seroconversion of CMV IgM to IgG was observed, while the monoclonal protein was no longer detectable. This suggests a relationship between acute CMV infection, transient MGUS and thrombosis. In accordance with current best practice guidelines for provoked venous thromboembolism (VTE), anticoagulation therapy could be discontinued after 3 months instead of 6 for unprovoked VTE, thereby reducing unnecessary time at risk of bleeding complications. While the relationships between CMV and both MGUS and thrombosis have been described independently, we are first to describe these three conditions occurring simultaneously. Furthermore, we provide a systematic review on the relation between CMV, MGUS and thrombosis.

Background

Cytomegalovirus (CMV) is a common pathogen to which 40%–100% of the world population has been exposed.¹ Depending on the immune status of the host, CMV can manifest itself in many ways, ranging from an asymptomatic infection to severe morbidity affecting multiple organ systems. While CMV infection is relatively common in immunodeficient patients, organ-specific involvement in immunocompetent hosts is rare. However, there have been cases reported of CMV-associated colitis, hepatitis, myocarditis and encephalitis in immunocompetent patients.^{2 3}

Case presentation

A 62-year-old woman with an unremarkable medical history arrived at the emergency department with dry cough and sternal pain that worsened during inspiration. She had already been suffering from headache, nausea, nocturnal sweating and fever of up to 39.4°C for 16 days before presentation. There had been no response to antibiotic treatment with azithromycin and doxycycline prescribed by her general practitioner. The patient did not smoke, drank one or two glasses of wine daily and did not use recreational drugs. There were no family members with similar symptoms. On physical examination, the patient was alert and fully oriented. She was feverous with a temperature of 38.6°C. Her blood pressure was 131/73 mm Hg, with a pulse of 108 beats per minute. The peripheral oxygen saturation was 96%, with a respiratory rate of 16 breaths per minute while breathing ambient air. During auscultation of the lungs, a pleural friction rub was heard in the left lower region with otherwise normal inhalation and exhalation sounds. On further physical examination no additional abnormalities were found.

Investigations

Laboratory testing showed an erythrocyte sedimentation rate within the normal range. There was no leucocytosis; however, there was lymphocytosis of 4.71 (normal values 1.00–3.50×10⁹/L) and an elevated C reactive protein of 28 (0–8 mg/L). Liver enzymes were elevated as well: aspartate transaminase was 93 (<31 U/L), alanine transaminase 169 (0–34 U/L), alkaline phosphatase 157 (40–120 U/L), gamma-glutamyl transpeptidase 174 (<38 U/L) and lactate dehydrogenase 417 (<248 U/L). The amylase and bilirubin levels were normal, as were the prothrombin time, activated partial thromboplastin time and albumin values. There was an elevated ferritin of 1592 (20–200 μg/L) with a normal transferrin saturation of 28%. Because of the coughing and sternal pain, a chest X-ray was performed, which showed no abnormalities. To rule out pulmonary embolism (PE), a CT angiography (CTA) of the thorax was performed subsequently, which indeed revealed a segmental lingual PE (figure 1A,B). The patient was admitted to the hospital for observation of the yet unexplained fever and treated for PE with low molecular weight heparin and a vitamin K antagonist.

Figure 1.

CT angiography (CTA) of the thorax showing a segmental lingual pulmonary embolism (A), as indicated with an arrow and coloured red (B) and CTA of the abdomen showing a small thrombus in the splenic vein (C), as indicated with an arrow and coloured blue (D).

Outcome and follow-up

In the following days, the general clinical condition of the patient remained stable, but the liver enzymes increased further (figure 2). This, in combination with nocturnal sweating, persistent fever and unexplained PE, raised the suspicion of an underlying malignancy. Serum electrophoresis and immunofixation were ordered to investigate the presence of monoclonal protein (M-protein) in order to demonstrate a possible multiple myeloma or lymphoma. Furthermore, a CTA of the abdomen was performed, which showed a small thrombus in the splenic vein (figure 1C,D). No abnormalities of the liver or other organs were found. Alcoholic hepatitis was considered as the cause for the increase in liver enzymes, but the patient again emphasised her moderate alcohol consumption. The acetaminophen, which was started on admission, was stopped. Apart from the low molecular weight heparin and the vitamin K antagonist, the patient did not use any other medication. Because of the increased ferritin concentration, macrophage activation syndrome was considered. However, the absence of anaemia, thrombocytopaenia and neutropaenia and the stable condition of the patient allowed for conservative treatment instead of immunosuppressive therapy.⁴ Haemochromatosis was ruled out by the normal transferrin saturation. Autoimmune hepatitis was deemed unlikely because of normal levels of antinuclear antibodies, antineutrophil cytoplasmic antibodies and smooth muscle antibodies. Wilson’s disease and an alpha-1 antitrypsin deficiency were ruled out. While investigating the possibility of a viral hepatitis, serology for Epstein-Barr virus (EBV) and hepatitis A, B, C and E viruses all turned out to be negative. Because our hospital is situated in an area with high HIV prevalence, we tested for HIV as well, the result of which came back negative. However, a strong positive CMV IgM with a titre of 2.84 Au/mL and only weak positive IgG with a titre of 5 Au/mL and low avidity of 0.10 was found. A diagnosis of acute CMV infection with concurrent pulmonary and splenic thrombosis was made. As the condition of the patient was improving, treatment for the CMV infection was not initiated, and a liver biopsy to diagnose CMV hepatitis was not performed. The patient was discharged from the hospital in good condition and treated with anticoagulation therapy only.

Figure 2.

Graphs showing the liver enzyme levels through follow-up time, with upper and lower normal values in red and blue. AF, alkaline phosphatase; ALT, alanine transaminase; AST, aspartate transaminase; gGT, gamma-glutamyl transpeptidase; LDH, lactate dehydrogenase.

On outpatient follow-up 4 weeks later, the patient was in good health. Surprisingly, an M-protein of type IgG lambda, 1.8 g/L, was discovered when the results of the previous testing came in. There were no indications for multiple myeloma as a low-dose whole-body CT scan ruled out the presence of a plasmacytoma, Bence Jones proteinuria was absent, and serum electrolytes, including calcium and phosphate, remained normal. We thus diagnosed the patient with monoclonal gammopathy of undetermined significance (MGUS). In line with the current best practice, no therapy was initiated for the MGUS.⁵

Two months later, seroconversion of the CMV IgM to IgG with an increase in the IgG titre of 5 to 62 Au/mL was observed; the IgM titre was decreased markedly to 1.55 Au/mL. Simultaneously, the M-protein was no longer detectable with serum protein electrophoresis and immunofixation (figure 3). This observation indicated an association between the acute CMV infection and the transient MGUS. Because either the acute CMV infection or the occurrence of MGUS could have attributed to the thromboembolic events, anticoagulation therapy could be discontinued after 3 months instead of at least 3 months, in accordance with the “The American College of Chest Physicians” (“CHEST”) guidelines, for provoked and unprovoked PE, respectively.⁶ Thus far, after 24 months of follow-up, the M-protein remains undetectable, no thromboembolic event has recurred, and the patient remains in good health.

Figure 3.

Serum protein electrophoresis and immunofixation showing a monoclonal protein of type IgG lambda, 1.8 g/L at presentation (A) and normalisation 2 months thereafter (B), as indicated with arrows. A, IgA; ELP, serum protein electrophoresis; G, IgG; K, kappa light chains; L, lambda light chains; M, IgM.

Discussion

In the described case, the seroconversion of the CMV IgM to IgG with concomitant normalisation of the M-protein demonstrates the relationship between acute CMV infection and the occurrence of MGUS in an immunocompetent host. Confirmation of the CMV infection by PCR was not performed due to logistic reasons. However, a positive CMV IgM in combination with a low avidity and titre of IgG is considered reliable evidence for acute CMV infection.⁷ The more than tenfold increase of the CMV IgG and marked decrease in IgM titres during follow-up also strengthen this conclusion.⁸ Since the venous thromboembolism (VTE) could not be explained otherwise, our patient illustrates the associations between VTE and either CMV or MGUS. In order to summarise previous reports and existing evidence on these associations, we conducted a systematic review (search method and flow chart provided in the online supplementary material). An overview of the results of the included studies is presented in tables 1–3.

Table 1.

Overview of existing literature on the associations between acute CMV infection and MGUS, resulting from our systematic review

Reports on the association between CMV and MGUS
Reference	Population	Study type	Results
Badley et al 9	201 liver transplant recipients.	Prospective cohort	RR=2.0 (95% CI 1.1 to 3.5, p<0.019) for M-protein development in symptomatic acute CMV infection.
Medd et al 10	92 allogeneic stem cell transplantation recipients.	Retrospective cohort	29 (32%) developed gammopathies. HR=5.74 (95% CI 5.21 to 6.27, p=0.001) for M-protein incidence in CMV reactivation. OR=3.69 (95% CI 1.38 to 9.86, p=0.01) for M-protein development in CMV seropositivity.
Bühler et al 21	Group 1: 25 immunocompetent patients with acute or recent (1–3 months) CMV infection. Group 2: 24 immunocompetent patients with acute or recent EBV infection. Group 3: 8 immunodeficient patients with acute or recent CMV infection.	Prospective cohort	Group 1: M-component in 10 (40%). Group 2: M-component in 0 (0%). Group 3: M-component in 5 (20%). Follow-up data available in 9 patients: M-component remained detectable in 4 patients: for 2 weeks n=2, 4 weeks n=1 and 6 weeks n=1 after the CMV infection.
Chakalarovski et al 11	192 HIV-negative renal transplant recipients.	Prospective cohort	25 (13%) gammopathies, of which 11 (44%) were monoclonal and 14 (56%) were polyclonal, all transient. Primary CMV infection in 10 of 22 (45%) patients with MIg tested, and in 28 of 98 (29%) patients without MIg tested.
Myara et al 12	76 heart transplant patients.	Prospective cohort	24 (32%) with CMV infection, of which 19 (79%) had monoclonal or oligoclonal banding.
Ginevri et al 13	80 paediatric kidney transplant recipients.	Retrospective cohort	30 (38%) CMV infection, of which 17 (56%) with MGUS.
Touchard et al 14	84 renal transplant recipients.	Prospective cohort	20 (24%) CMV infection, of which 19 (95%) with monoclonal IgG.
Hebart et al 15	47 allogeneic bone marrow transplant recipients.	Prospective cohort	36 (77%) with acute CMV, of which 11 (30%) with MGUS.
Drouet et al 16	84 renal transplant recipients.	Prospective cohort	21 (25%) oligoclonal or monoclonal immunoglobulins, of which 20 (95%) double CMV/EBV infection and 1 (5%) disseminated EBV infection.
Cakalaroski et al 17 No full text	115 HIV-negative renal transplant recipients.	Prospective cohort	MGUS in 16 patients (14%). Prior CMV infection was a risk factor for MGUS occurrence. Inability to convert IgM to IgG CMV antibodies was a risk factor for MGUS persistence.
Alliot et al 18	Male, 56 years, immunocompetent.	Case report	Acute CMV infection and polyclonal gammopathy with seroconversion of CMV IgM after 3 weeks and normal immunofixation after 1 month.
Anniccherico-Sánchez et al 19 Meeting abstract	Male, 66 years, immunocompetent.	Case report	Acute CMV infection with IgM lambda gammopathy, which disappeared after 3 months.
Kramer et al 20 No full text	Female, 67 years, immunocompetent.	Case report	Acute CMV infection and monoclonal IgG with cryoglobulinaemia type II.
Vodopick et al 32	Female, 13 years, leukaemia in remission.	Case report	Acute CMV infection with transient occurrence of IgA lambda and Bence Jones proteinuria, cleared before normalisation of CMV antibody titre.

CMV, Cytomegalovirus; EBV, Epstein-Barr virus; HR, hazard ratio; MGUS, monoclonal gammopathy of undetermined significance; MIg, monoclonal immunoglobulins; OR, odds ratio; RR, relative risk; VTE, venous thromboembolisms.

Table 2.

Overview of existing literature on the associations between acute CMV infection and VTE, resulting from our systematic review

Report on the association between CMV and VTE
Reference	Population	Study type	Results
Ceccarelli et al 22	Cases of immunocompetent patients with CMV infection and thrombosis.	Systematic review	79 reports including 115 cases were identified. 53 of 81 (65%) patients tested with a coagulation disorder. VTE locations: 29 (25%) pulmonary embolism, 26 (23%) portal vein, 16 (14%) splenic infarction, 15 (13%) superior mesenteric vein and 10 (9%) lower limbs.

CMV, Cytomegalovirus; VTE, venous thromboembolisms.

Table 3.

Overview of existing literature on the associations between MGUS and VTE, resulting from our systematic review

Reports on the association between MGUS and VTE
Reference	Population	Study type	Results
Kristinsson et al 26	5326 patients with MGUS vs 20 161 matched controls.	Retrospective cohort	At 1, 5 and 10 years after MGUS diagnosis, the HR for VTE was 3.4 (95% CI 2.5 to 4.6), 2.1 (95% CI 1.7 to 2.5) and 2.1 (95% CI 1.8 to 2.4), respectively. The HR for arterial thrombosis was 1.7 (95% CI 1.5 to 1.9), 1.3 (95% CI 1.2 to 1.4) and 1.3 (95% CI 1.3 to 1.4), respectively.
Kristinsson et al 27	2374 veterans with MGUS vs 4 187 631 controls.	Retrospective cohort	RR of DVT after diagnosis of MGUS: 3.3 (95% CI 2.3 to 4.7). RR of DVT the first year after diagnosis of MGUS: 8.4 (95% CI 5.7 to 12.2).
Gregersen et al 28	1610 patients with MGUS vs 16 100 matched controls.	Retrospective cohort	IRR of VTE after MGUS diagnosis: 1.37 (95% CI 1.0 to 1.88).
Goebel et al 33	72 patients with MGUS vs 21 990 controls, all solid organ transplant recipients.	Retrospective cohort	Propensity-adjusted RR of VTE after MGUS diagnosis: 1.66 (95% CI 1.15 to 2.41).
Sallah et al 34	310 patients with MGUS.	Prospective cohort	VTE in 19 (6.1%) with a median of 44 months (12–67 months).
Srkalovic et al 35	174 patients with MGUS.	Retrospective cohort	VTE in 13 (7.5%) with a median of 4 months (0–67 months).
Za et al 36	1238 patients with MGUS without history of thrombosis <2 years prior to diagnosis.	Retrospective cohort	33 (2.7%) thrombosis: 2.5/1000 PY arterial events, 1.9/1000 PY VTE in >12 months of follow-up.
Cohen and Sarid37	166 male veterans with MGUS vs 465 matched controls.	Retrospective cohort	Multivariate analysis: HR of VTE after MGUS diagnosis: 1.31 (95% CI 0.58 to 2.99, p=0.52).
Lindqvist et al 38 Meeting abstract	299 patients with MGUS vs 52 patients with LC-MGUS and 5413 controls.	Retrospective cohort	OR for VTE: 0.75 (95% CI 0.50 to 1.12) in MGUS vs controls. OR for VTE: 1.81 (95% CI 0.92 to 3.58) in LC-MGUS vs controls.
Muslimani et al 39 No full text	112 patients with MGUS.	Retrospective cohort	VTE in 9 patients (8%).
Righini et al 40	Male, 39 years, no medical history, with extensive thrombosis in both legs and a pulmonary embolism.	Case report	MGUS: IgG lambda 2.97 g/dL. No genetic or acquired anomaly predisposing thrombophilia or solid or haematological neoplasia.
Ali and Mirrakhimov41	Female, 86 years, medical history of MGUS.	Case report	Cyanosis of the fifth digit of the left hand, with no thromboembolic source on imaging studies.

CMV, Cytomegalovirus; DVT, deep vein thrombosis; HR, hazard ratio; IRR, incidence rate ratio; LC, light chain; MGUS, monoclonal gammopathy of undetermined significance; OR, odds ratio; PY, person years; RR, relative risk; VTE, venous thromboembolisms.

The association between CMV and MGUS (table 1) has been demonstrated previously in cohort studies with immunodeficient patients.^9–17 Reports on this phenomenon in immunocompetent individuals are scarce and limited to three case reports and one small cohort study.^18–21 In this cohort study, 10 out of 25 (40%) immunocompetent patients with acute CMV infection developed concurrent MGUS, compared with none in a control group consisting of patients with acute EBV infection.²¹ Although seroconversion of the CMV IgM to IgG was not investigated in this study, M-protein was transient in all patients with available follow-up data. Simultaneous CMV seroconversion and disappearance of an M-protein has only been described once in a French case report.¹⁸ This concerned a hypergammaglobulinaemia with monoclonality of IgM lambda, IgG kappa and IgG lambda and is therefore not comparable with our case. Apart from one confined case report in a meeting abstract,¹⁹ VTE was not present in any of these articles.

The VTE in our case could be attributed to either the acute CMV infection or the MGUS, as both relations have been reported. The relationship between acute CMV infection and VTE (table 2) has recently been demonstrated in a systematic review including 79 articles containing 115 cases of immunocompetent patients in total.²² As we identified the same references based on our search strategy, we decided to affirm their conclusion that CMV should be taken into consideration in all cases of thrombosis with no apparent cause. Furthermore, another recent literature review identified the splanchnic veins and those of the lower limbs as main sites of thrombosis associated with acute CMV infection. A PE was reported in one-third of the cases.²³ High incidence of CMV-related thrombosis in the splanchnic veins, such as in the splenic vein in our case, could be explained by the local prothrombotic effect of acute CMV hepatitis.²⁴ Several theories discuss the mechanism of CMV-associated thrombosis, including the enhancement of platelet and leucocyte adhesion to infected endothelial cells, activation of factor X, elevation of circulating factor VIII (FVIII) levels and induction of antiphospholipid antibody production.²⁵

The correlation between MGUS and VTE (table 3) has been demonstrated in several large cohort studies.^26–28 For example, Gregersen et al found an incidence rate ratio for VTE of 1.37 (95% CI 1.0 to 1.88) for patients with MGUS compared with matched controls, indicating a 37% higher incidence of VTE in patients with MGUS as compared with the general population.²⁸ The exact aetiology of this relationship has yet to be determined, although suggestions have been offered. Auwerda et al found a significant increase in median von Willebrand factor (vWF) activity and FVIII levels in patients with MGUS.²⁹ Crowley et al also demonstrated a trend towards patients with MGUS having higher vWF and FVIII levels than normal controls, although this was non-significant.³⁰ Furthermore, Caimi et al found a significant increase in whole-blood viscosity at high shear rate and in plasma viscosity at low shear rate and a marked decrease in erythrocyte deformability as possible factors underlying the pathogenesis of MGUS-related VTE.³¹

In conclusion, the possible relationships between acute CMV infection, MGUS and VTE in immunocompetent patients have been described separately. To our knowledge, we are the first to demonstrate these three associations combined. The assumption that MGUS is a symptom of acute CMV infection may have considerable consequences in determining the length of follow-up and treatment of the patient. This also applies to the association between VTE and both acute CMV infection and MGUS. This knowledge provided an explanation for the PE in our case, on which anticoagulant therapy could be discontinued earlier, decreasing the time at risk of bleeding and other adverse events. Furthermore, by explaining that the MGUS was induced by the acute CMV infection, we could lessen our patients’ fear of developing a plasma cell malignancy. Therefore, we believe insight into the associations between acute CMV infection, transient MGUS and VTE in immunocompetent patients is valuable to avoid uncertainty and unnecessary therapy or follow-up, and thereby reduces morbidity of future patients.

Patient’s perspective.

The first two weeks of my disease (fever, nauseous, headache) I felt insecure and not understood because I had the feeling that nobody could help me. Finally, after multiple blood tests and a CT-scan, I was diagnosed with a pulmonary embolism. I felt a little relieved. During my stay in the hospital I again felt insecure and afraid because of all the examinations. They wanted to know where the pulmonary embolism came from. What was the reason?! They told me that cancer could be a possibility. A blood thinner was the only medicine they gave me. That wasn’t a problem for me, I even felt a little reassured. I had a constant headache, but I couldn’t have any acetaminophen. Tramadol was proposed, but I didn’t want to take that. After I was discharged from the hospital, I was still scared because of the observed presence of MGUS. I was scared because of the possible causes and consequences/impact the doctor had explained. Fortunately, I recovered very soon. The MGUS disappeared and after three months I could stop the blood thinners. I didn’t have a problem with that! These days I’m doing fine!

Learning points.

• Transient monoclonal gammopathy of undetermined significance (MGUS) can be caused by acute cytomegalovirus (CMV) infection, a phenomenon which can also occur in an immunocompetent patient.

• Unprovoked venous thromboembolic events can be attributed to both acute CMV infection and MGUS.

• Knowledge of these associations can alleviate uncertainties that patients may have on the disease origin and allow for earlier discontinuation of anticoagulant treatment, reducing risk of adverse events.