- Top of page
- Case reports
- Author contribution
Immune thrombocytopenic purpura (ITP) is characterized by decreased platelet numbers secondary to platelet destruction and reduced platelet production. Even if the ITP persists, it typically responds to ‘ITP-specific’ therapies, such as intravenous immunoglobulin, steroids, intravenous anti-D, and splenectomy. Several reports, including our previous study, have implicated cytomegalovirus (CMV) in the pathogenesis of infrequent cases of ITP that were not severe in nature. A recent study from China suggested that CMV is the aetiology of some cases of acute ITP of childhood and may require different treatment. We report two adult and two paediatric patients with refractory, severe, symptomatic thrombocytopenia, who were diagnosed with ITP and found to have active CMV infection. Their presentations included fever, transaminitis, neutropenia, and atypical lymphocytosis, but in particular, treatment-refractory, severe ITP. Treatment with steroids appeared to worsen the CMV-ITP. All four cases showed improvement in platelet counts within two weeks of starting ganciclovir and cytogam and tapering steroids. Based on the four patients described here, we believe that, in certain cases, CMV infection will result in symptomatic, severe, refractory ITP, which may be indistinguishable from typical ITP. Eradication of CMV with antiviral therapy improved the ITP in these cases.
Immune thrombocytopenic purpura (ITP) is a common cause of acquired thrombocytopenia. It is generally believed to be caused by auto-anti-platelet antibodies that are thought to not only destroy platelets peripherally, but also to damage megakaryocytes and/or inhibit platelet production in the marrow (Bussel et al, 2006;Xiao et al, 2006). The aetiology of ITP is not known in most instances and a highly sensitive and specific test for its diagnosis has yet to be developed.
ITP in childhood has been estimated to be post-infectious in approximately two thirds of instances. Many different agents have been implicated in the aetiology of this form of ITP, including not only cytomegalovirus (CMV), but also varicella, rubella, mumps, Epstein-Barr virus, and parvovirus B19, among others (Chanarin & Walford, 1973; Fiala & Kattlove, 1973;Ip & Corner, 1973; Harris et al, 1975; Muntendam, 1975; Dor et al, 1977; Sahud & Bachelor, 1978; Shimm et al, 1980; Aguado et al, 1984; Anton et al, 1985; Amitai & Granit, 1986; Landonio et al, 1992; Wright, 1992; Eisenberg & Kaplan, 1993; Murray et al, 1994; Mizutani et al, 1995; Von Spronsen & Breed, 1996; Wright et al, 1996; Arruda et al, 1997; Miyahara et al, 1997; Swanobori et al, 1997; Gasbarrini et al, 1998; Gural et al, 1998; Lopez et al, 1998; Heegaard et al, 1999; Sakata et al, 1999; Aboulafia et al, 2000; Crapnell et al, 2000; Garcia-Suarez et al, 2000; Hida et al, 2000; Hsiao, 2000; Emilia et al, 2001; Scaradavou, 2002; Yenicesu et al, 2002; Ichiche et al, 2003; Aktepe et al, 2004; Alliot & Barrios, 2005; Jackson et al, 2005; Li et al, 2005; Nomura et al, 2005; Rajan et al, 2005; Asahi et al, 2006; Cooper & Bussel, 2006; Sayan et al, 2006; Suvajdzic et al, 2006). These infections trigger an autoimmune process against platelets, even though the infections themselves are transient and seemingly neither atypical or severe in nature. However, in the past several years, cases of apparently idiopathic ITP have been found to be secondary to an unsuspected, persistent infection. Optimal management of the ITP in these cases, secondary to human immunodeficiency virus (HIV), H. Pylori or hepatitis C infection, often requires recognition and treatment of the underlying infection.
CMV has been reported to both cause and perpetuate ITP (Chanarin & Walford, 1973; Fiala & Kattlove, 1973; Ip & Corner, 1973; Harris et al, 1975; Muntendam, 1975; Dor et al, 1977; Sahud & Bachelor, 1978; Shimm et al, 1980; Aguado et al, 1984; Anton et al, 1985; Wright, 1992; Eisenberg & Kaplan, 1993; Mizutani et al, 1995; Von Spronsen & Breed, 1996; Arruda et al, 1997; Miyahara et al, 1997; Swanobori et al, 1997; Gural et al, 1998; Lopez et al, 1998; Sakata et al, 1999; Crapnell et al, 2000; Ichiche et al, 2003; Alliot & Barrios, 2005; Nomura et al, 2005). These cases and their treatments are summarized in Tables I and II, respectively. The majority of the small number of previously described cases were newly diagnosed, acute ITP in adults that, nonetheless, appeared to be susceptible to standard treatments. Our previous study found that only three out of 28 paediatric patients (11%) and three out of 80 adult patients (4%) presenting to Cornell University Weill Medical Centre for the management of thrombocytopenia were positive for CMV in the urine by shell vial assay. None of the viruric cases were linked to refractory disease and none had their CMV treated (Levy & Bussel, 2004).
In the current study, four patients with treatment unresponsive, clinically severe cases of ITP were discovered to have active CMV infection. Two of the patients had an intracranial haemorrhage and one died secondary to complications of ITP. These cases, as well as their treatments, are summarized in Tables III and IV, and Fig 1. Their treatment of CMV combined with discontinuation of immunosuppression, in particular prednisone, lead to marked increases in the platelet counts within 1–2 weeks. Although the ITP did not go completely into remission in the four cases, it improved markedly and became amenable to management with conventional agents.
Table III. Diagnostic and other clinical features in the four cases of CMV-related ITP in this study.
|Clinical features/Laboratory findings||Case 1||Case 2||Case 3||Case 4|
|Increase in platelet number as CMV PCR normalizes||+||+||+||+|
|Steroids result in worsening ITP/ no benefit||+||+||+||+|
|Nadir platelet count (×109/l)||2||2||5||1|
|Improvement in platelet count with ganciclovir/cytogam||+||+||+||+|
|Need for splenectomy||−||+||−||+|
|Presence of ICH||+||−||−||+|
|Bone marrow characteristics||Megakaryocytic hyperplasia||No significant megakaryocytic hyperplasia||No significant megakaryocytic hyperplasia||Megakaryocytic hyperplasia|
|Presented with CMV like illness-myalgia, fever, headache, malaise, throat pain||+||−||−||−|
|Severity of bleeding worse than others with ITP with similar degree of thrombocytopenia||+||+||+||+|
|HIV testing negative||+||+||+||+|
|Episodes of intermittent neutropenia||+||−||−||−|
|Atypical lymphocytes on smear||−||−||+||−|
Table IV. Proposed guidelines for treatment of CMV-related ITP.
|Initiate Treatment of CMV:|
| Cytogam (100–200 mg/kg) 1–3 times/week|
| Ganciclovir (5 mg/kg) 2 times/day|
|If patient experiences myelosupression from ganciclovir, switch to foscarnet or cidofavir|
|Once CMV PCR becomes undetectable for approximately 2 weeks, wean ganciclovir to daily and then cytogam|
|Reduce immunosuppressive agents, including steroids, as rapidly as feasible|
|Further discontinuation of therapy depends upon the clinical setting and the platelet response|
|If platelet-specific treatment is required while awaiting a response to anti-CMV therapy, use IVIG to avoid immunosuppression (the first indication of improvement may be an improved response to IVIG)|
Figure 1. Relationship of platelet count to CMV titres and anti-CMV treatment. (A) Case 1 was found to be thrombocytopenic with an elevated CMV PCR. Standard therapies for ITP did not cause an increase in her platelet count until anti-CMV therapy and a thrombopoietic agent were initiated. (B) Case 2 had no clinical or laboratory response to standard agents (IVIG) prior to the eradication of the active CMV infection. After the CMV load by PCR was <200 DNA copies/ml, the patient showed some response to standard agents even though it was not durable. In this case, eradication of CMV was not curative but it was an essential first step in allowing the patient to respond to standard therapies, such as splenectomy and IVIG. (C) The temporal association of suppression of CMV and rising platelet counts is illustrated for Case 3. In April 2004, the patient experienced an asthma attack and needed steroids, which resulted in presumed CMV reactivation and worsening of the platelet count. The ITP exacerbation required ganciclovir IV BID. Ganciclovir was continued for two months post the first non-detectable CMV PCR. (D) Case 4 was initially treated with high dose steroids without a significant rise in platelet counts. Once he was started on anti-CMV therapy and the CMV PCR became undetectable, his platelet count began to rise and remained stable despite simultaneously rapidly tapering the steroids.
Download figure to PowerPoint
- Top of page
- Case reports
- Author contribution
Nineteen cases of CMV-associated thrombocytopenia have been reported in the English language (Tables I and II; Alliot & Barrios, 2005; Ichiche et al, 2003; Von Spronsen & Breed, 1996; Sahud & Bachelor, 1978; Wright, 1992; Mizutani et al, 1995; Arruda et al, 1997; Gural et al, 1998; Swanobori et al, 1997; Crapnell et al, 2000; Chanarin & Walford, 1973; Harris et al, 1975; Ip & Corner, 1973; Eisenberg & Kaplan, 1993; Miyahara et al, 1997; Nomura et al, 2005; Shimm et al, 1980; Fiala & Kattlove, 1973; Sakata et al, 1999; Muntendam, 1975;Dor et al, 1977; Aguado et al, 1984; Lopez et al, 1998; Anton et al, 1985), but only two described severe, refractory cases (Eisenberg & Kaplan, 1993; Gural et al, 1998). Several important trends are evident from these 19 cases. Thirteen presented with a CMV-like illness, including symptoms of fever, myalgia, malaise, headache and sore throat. A work-up of the thrombocytopenia revealed atypical lymphocytes in 13 cases and transaminitis in 10; bone marrow aspirates revealed megakaryocytic hyperplasia in 10 cases. Steroids were the main treatment in the majority of reports; however, they proved to be of minimal or no benefit in six of the cases. Three of these six cases did not respond consistently to conventional therapy and ultimately required splenectomy. Ganciclovir was instituted in only two of these cases.
There are few reported paediatric cases of CMV-associated ITP (Fiala & Kattlove, 1973; Mizutani et al, 1995; Sakata et al, 1999). This present report describes a 3-year-old female and 9-year-old male with CMV-ITP. A recent study in China demonstrated that in 81 children with ITP at diagnosis and a mean age of 2·75 years, the majority had evidence of CMV infection. Fifty cases were myeloid CMV antigen positive with a positive rate of 61·73%, while 17 cases were blood CMV IgM positive (20·99%), and 70 cases were CMV IgG positive (86·42%). The CMV-infected patients were more likely to experience an exacerbation of ITP, be refractory, and experience chronic disease, and the authors concluded that CMV is an important risk factor for more severe, persistent childhood ITP (Ding et al, 2007).
In marked contrast to the reported cases of CMV infection in patients with ITP, the four cases described here with CMV are notable for their refractory nature. The cases proved largely unresponsive to front line ITP treatments; any responses at all were very short-lived. In these cases, as illustrated in Fig 1, it appeared that the CMV must be suppressed (or eradicated if possible) before standard therapy will be effective. For instance, in Case 2, although the thrombocytopenia was still significant after eradication of CMV, the patient was responsive to standard agents; prior to CMV treatment, he had had no response. In fact, the aggressive immunosuppression used to treat refractory ITP may exacerbate the primary CMV infection and worsen the ITP, or at least prevent it from improving. Treatment with steroids led to CMV reactivation and exacerbation/relapse of ITP in two of the patients. Given that steroids, along with other immunosuppressive therapies, may reactivate CMV, patients on immunosuppresive therapies who experience severe, refractory, thrombocytopenia should be screened with a PCR for CMV infection. Fortunately, this appears to be a relatively infrequent phenomenon. This exacerbation by steroids is similar to reports of hepatitis C-associated ITP, in which steroid treatment of the thrombocytopenia worsened the hepatitis and failed to increase the platelets (Rajan et al, 2005).
In the reported cases here, the patients presented with a transaminitis, experienced improved platelet counts once ganciclovir and cytogam were instituted and immunosuppression was reduced, and, in two of the cases, had megakaryocytic hyperplasia upon bone marrow examination. Only one patient presented with a CMV-like illness or showed atypical lymphocytes on smear, highlighting the need to have a very low threshold in suspecting CMV infection as the underlying cause of thrombocytopenia (Table V).
Table V. Signs and symptoms that are possibly suggestive of the diagnosis of CMV-related ITP.
|Fever of unknown aetiology|
|Atypical lymphocytes on smear|
|Severe symptomatic ITP with platelet counts less than 5–10 × 109/l and/or ICH|
|ITP unresponsive to standard therapies|
|ITP exacerbations with steroid or other immunosuppressive treatment|
It is not clear why the four patients presented here were so vulnerable to CMV. Immunoglobulin levels, T cell and B cell function, and antibody levels to pneumoccocus, were found to be normal in each case. None of the patients were ever hypogammaglobulinemic or had any significant history of other infections either before or after their episodes of CMV ITP. The 3-year-old patient only experienced other infections after being intubated in the paediatric intensive care unit, having a central line, and receiving antibiotics for fever that may have been caused by CMV. Her PCR had been negative for at least one month at the time of her death. In her case, she appeared to have had a very complex, multi-factorial disease in which CMV was just one manifestation. Furthermore, CMV infection and hepatitis preceded the development of her ITP. The other three patients were remarkably well before, and since diagnosis and treatment of the CMV.
No specific studies of the mechanism of the underlying thrombocytopenia were performed other than carefully tracking the CMV PCR and relating platelet improvement to the decreasing PCR level in all four cases. Although the exact mechanism by which viral or bacterial agents lead to ITP is not known, several hypotheses exist (see Table VI). While molecular mimicry and chronic infection leading to immune dysregulation are possible, CMV is able to directly infect megakaryocytes and thereby decrease platelet production (Mizutani et al, 1995; Crapnell et al, 2000; Xiao et al, 2006). This seems to be the best explanation for the failure to respond to any of the conventional ITP therapies, especially as their primary effects are to block platelet destruction, not to stimulate platelet production (Psaila & Bussel, 2008). If this is the case, short-term use of thrombopoietic agents may be useful in these patients in the future. Two of the four patients suffered intracranial haemorrhages, and all four cases involved bleeding symptoms that were severe, suggesting that vasculopathy or dysfunctional platelets may compound the underlying thrombocytopenia.
Table VI. Possible mechanisms of how viruses/bacteria cause ITP.
|Molecular mimicry (Wright et al, 1996)|
|Immune dysregulation (Cooper & Bussel, 2006)|
|Induction of platelet phagocytosis (Swanobori et al, 1997)|
|Direct infection of megakaryocytes (Amitai & Granit, 1986; Landonio et al, 1992; Mizutani et al, 1995; Crapnell et al, 2000; Xiao et al, 2006)|
|Production of hematopoietic inhibitory cytokines such as interferon-γ and tumour necrosis factor-α by CMV-infected leucocytes and stromal cells of bone marrow (Miyahara et al, 1997)|
|Suppression of haematopoiesis of human progenitor cells (Arruda et al, 1997)|
|Megakaryocyte production of dysfunctional platelets|
|Induction of vascular disease|
Treatment of the CMV requires two components: effective treatment of the CMV and discontinuation of immunosuppressive medications (see Table IV). The treatment protocol followed was arbitrary but included a combination of both IV ganciclovir and IV cytogam to optimize rapid suppression of the CMV. Simultaneous reduction of immunosuppression, with IVIG as needed to transiently support the platelet count, may be crucial during this phase. If an effect on the platelet count is seen over 2–4 weeks, tapering of the CMV-specific therapies may be instituted. Because of the potential for myelosuppression by prolonged IV ganciclovir, it is recommended to taper this first. In Case 1, long-term treatment with high-dose ganciclovir led to the eventual development of myelosuppression, forcing the use of foscarnet, despite its nephrotoxicity, and later, cidofavir. The main point of management is that when CMV is controlled by anti-viral therapies, normal treatments for ITP, such as IVIG and splenectomy, regain their effectiveness.
Several general points can be made based on these four reported cases. First, on presentation, the thrombocytopenia is essentially indistinguishable from classic ITP, characterized by decreased platelet numbers while other cell lines remain intact, and by (at best) a rather limited response to ‘ITP specific’ therapies, such as IVIG, anti-D, and steroids. Second, if there is any suspicion of CMV based on exposure to it, refractoriness or severity of the ITP, or due to similarities to the cases listed in the tables (transaminitis, atypical lymphocytes on smear, fever, etc), it is appropriate to test for CMV infection by PCR. The authors believe that PCR testing is superior to serology, given the likelihood of false positive results when using serology after a patient has received IVIG. There is no requirement for any type of a mono-like illness, which was seen in only one of the four cases. Third, if steroids appear to worsen or have no affect on the thrombocytopenia, CMV infection should be high on the differential diagnosis. Finally, in certain areas, if suited to the population epidemiology, such as possibly in parts of China, screening of all cases of de novo ITP for CMV may be appropriate so that treatment can be altered accordingly in infected cases.