HCV-associated thrombocytopenia: clinical characteristics and platelet response after recombinant α2b-interferon therapy
Dr Julio García-Suárez, Servicio de Hematología, Hospital Universitario Príncipe de Asturias, Carretera Alcalá-Meco S/N (Campus Universitario), 28805 Alcalá de Henares, Madrid, Spain. E-mail: email@example.com
Hepatitis C virus (HCV) has been proposed as a possible causative agent of chronic thrombocytopenia. We investigated HCV infection in a series of 51 unselected Spanish patients with chronic acquired thrombocytopenia. Anti-HCV and HCV viraemia were detected in 13/51 (22·5%) of cases; this prevalence was particularly significant when compared with HCV seropositivity in age-matched controls (0·4%). Anti-HCV-positive patients, four men and nine women with a median age of 59·3 years (range 36–72), had a mean platelet count of 55·8 × 109/l (range 12–96). Only one of our HCV-positive thrombocytopenic patients had hypersplenism. Platelet-associated IgG (PAIgG) was negative, as measured by immunofluorescent flow cytometric analysis in 11/13 HCV-positive thrombocytopenic patients. Thus, thrombocytopenia in our HCV-positive patients appeared to be non-autoimmune mediated. In six patients, a trial of recombinant α2b-interferon (IFN-α) given at a dose of 3 MU three times per week for 6–24 months gave a durable (> 1 year) and significant increase in platelet count in all six patients. The maximum increase occurred after 6 months of IFN-α therapy. In conclusion, the ability of IFN-α to increase platelet counts in HCV-positive thrombocytopenic patients supports mechanisms involving a direct role for HCV inhibiting platelet production.
Several reports ( Silva et al, 1992 ; Pawlotsky et al, 1995 ; Linares et al, 1996 ; Bauduer et al, 1998 ; Gerli et al, 1999 ), mostly from southern Europe, have shown an increased prevalence (5–35%) of hepatitis C virus (HCV) infection in patients with chronic acquired thrombocytopenia. The pathogenic mechanism of the thrombocytopenia in these patients has not been defined as symptoms of portal hypertension or splenomegaly were not evident in most of the cases. The proposed mechanisms for induction of thrombocytopenia in chronic HCV infection invoke the direct involvement of HCV. These mechanisms include (i) direct infection by HCV of megakaryocyte/platelet lineage, leading to underproduction of platelets ( Silva et al, 1992 ; Nagamine et al, 1996 ), and/or (ii) dysregulation of the host immune system, leading to destruction of platelets ( Wang et al, 1992 ; Alter, 1995; Nagamine et al, 1996 ; Bauduer et al, 1998 ).
Treatment of patients with HCV-associated thrombocytopenia is an important and unresolved problem particularly because the mechanism responsible for the occurrence of thrombocytopenia in these individuals remains unclear. The various treatment used, including steroids, intravenous immunoglobulin (IVIg) and splenectomy, have generally produced only transient responses ( Durand et al, 1994 ; Linares et al, 1996 ; Bauduer et al, 1998 ). In addition, steroids generally increase hepatitis C viraemia in these HCV-positive patients ( Magrin et al, 1994 ). Alpha-interferon (IFN-α) is a cytokine with pleiotropic activities that has been shown to be an effective drug for patients with chronic HCV infection. This therapeutic activity has been observed in liver disease as well as in the extrahepatic manifestations of chronic HCV infection ( Pozzato et al, 1994 ; Gordon & Schuchter, 1995; Dusheiko, 1997). Therefore, it may also be important to know whether IFN-α has the ability to increase platelet counts in patients with HCV-associated thrombocytopenia.
To address this issue, we first investigated the prevalence of HCV infection in an unselected series of 51 patients with chronic acquired thrombocytopenia. Subsequently, we also studied the clinical characteristics and the haematological response to IFN-α in our HCV-positive thrombocytopenic patients.
Materials and methods
Patients Blood samples were from a well-characterized series of 51 patients with chronic acquired thrombocytopenia, defined as isolated thrombocytopenia (< 100 × 109/l) that had persisted for a minimum of 6 months [excluding ethylenediamine tetra-acetic acid (EDTA)-related thrombocytopenia]. All the patients were negative for human immunodeficiency virus antibodies. A clinical diagnosis of autoimmune thrombocytopenic purpura (AITP) was made on the basis of a normal or increased number of bone marrow megakaryocytes, absence of myelodysplastic features on bone marrow, absence of drug intake or coagulopathy and absence of other causes of thrombocytopenia (for example antiphospholipid syndrome, connective tissue disease or neoplastic disorder). A diagnosis of HCV-associated thrombocytopenia required all the criteria for AITP and also positive anti-HCV antibodies and detectable HCV viraemia. Informed consent was obtained from each patient before entry into the study, and before IFN-α treatment. During the same period, 1000 consecutive blood donors from our geographical area, aged 35–65 years, served as control group for this study.
Laboratory investigations All patients were evaluated for routine blood chemistry according to standard procedures. Anti-HCV testing was performed in duplicate by means of second-generation enzyme-linked immunosorbent assay (ELISA) (EIA 2-HCV; Ortho Diagnostic Systems, Raritan, NJ, USA), according to the manufacturer's instructions. To confirm HCV infection, all anti-HCV-positive patients were tested for HCV-RNA by reverse transcriptase polymerase chain reaction (RT-PCR), as previously described ( Okamoto et al, 1990 ). All anti-HCV-positive patients were also tested for type 1 anti-liver kidney microsomal (LKM-1) autoantibodies by means of immunofluorescence on rat liver and kidney cryostat sections ( Rizzeto et al, 1973 ). Standard assays were used to detect antinuclear (ANA) and smooth muscle antibodies. All serum samples were negative for HIV and hepatitis B virus serology using a standard enzyme immunoassay (EIA) technique.
Other investigations Abdominal ultrasound was performed in all patients at presentation to rule out hypersplenism. Liver biopsy (taken percutaneously) was only obtained in 4 of 13 HCV-positive thrombocytopenic patients. Platelet-associated immunoglobulin G (PAIgG) levels were determined in 11/13 HCV-positive thrombocytopenic patients by immunofluorescent flow cytometric analysis, as previously described by Goodall & Macey (1994). Briefly, each patient's platelets were obtained by centrifugation of the platelet-rich plasma at 640 g for 10 min at room temperature. Washed platelets from the patient were resuspended in EDTA phosphate-buffered saline to 1–2 × 107/ml. The platelets were incubated with fluorescein isothiocyanate (FITC)-conjugated rabbit F(ab′)2 anti-human IgG (Tago, CA, USA) for 20 min at room temperature. After washing twice, the sample was analysed on a FACScan (Becton Dickinson, CA, USA) using lysis ii software. Only single intact platelets were gated for analysis of fluorescence. Platelets were incubated with 5 µl FITC rabbit anti-mouse immunoglobulins and with 5 µl FITC anti-GpIb or GpIIIa for negative and positive controls respectively. Background immunofluorescence was established on the negative control for each patient. A normal amount of PAIgG was considered as negative. To assess the performance of the PAIgG assay in our laboratory, platelets from nine patients with well-defined AITP and 10 patients with non-immune thrombocytopenia (nine with inherited thrombocytopenia and one with incidental thrombocytopenia of pregnancy) were also tested. Elevated PAIgG was observed in 7/9 AITP patients and 4/10 patients with non-immune thrombocytopenia. Therefore, non-immune thrombocytopenic patients were more likely to give negative results.
Treatment strategies In this study, two groups of patients were delineated by treatment received: (i) patients who received IFN-α treatment and (ii) patients who were managed with conventional therapy for classic AITP. As immune-mediated thrombocytopenia may be worsened by IFN-α treatment ( Cavanna et al, 1994 ; Dusheiko, 1997), before starting IFN-α we excluded an autoimmune-mediated mechanism as being responsible for thrombocytopenia in our HCV-positive patients. All patients in the IFN-α group were given recombinant α2b-interferon (Intron A, Schering-Plough) at 3 million units (MU) s.c. three times per week. Patients were premedicated with acetaminophen 650 mg orally 30 min before IFN-α. Patients receiving conventional AITP therapy received various forms of treatment, including prednisone (1 mg/kg/d for 3–6 weeks), intravenous immunoglobulins (IVIg; 400 mg/kg/d for 5 d), danazol (600 mg/d for 8–14 weeks), splenectomy or no therapy and kept under observation (because conventional treatment was not indicated).
Anti-HCV prevalence in thrombocytopenic patients
Between January 1996 and December 1998, 51 consecutive patients with chronic acquired thrombocytopenia (39 men, 12 women; mean age 47·8 years, range 16–72) were prospectively studied at the haematology clinic of our hospital (Table I). Anti-HCV was detected in the serum of 13/51 patients (22·5%; four men and nine women). The age of these HCV-positive thrombocytopenic patients ranged from 36 to 72 years (median 59·3). The 13 anti-HCV-positive patients were also HCV-RNA positive by RT-PCR. None of these patients presented anti-LKM-1, ANA or smooth muscle antibodies. The prevalence of HCV infection in the age-matched healthy controls was 4/1000 (0·4%).
Among the 13 HCV positive patients, five had been transfused before the 1990s (before the implementation of blood donor screening for HCV antibodies). These five HCV-positive patients had received a median of four (range 2–10) blood product transfusions. The remaining eight HCV-positive patients did not have any risk factor for HCV infection. None of the HCV-negative thrombocytopenic patients had a history of blood transfusions.
Clinical and laboratory data of HCV-positive thrombocytopenic patients
The complete records of the 13 patients with chronic thrombocytopenia and HCV infection were analysed (Table II). At the time of admission to our department, the mean platelet count was 55·8 × 109/l (range 12–96). Four patients (cases 1, 4, 6 and 8) had severe thrombocytopenia (platelet count < 40 × 109/l). Three of these had bleeding symptoms. The remaining nine patients had moderate thrombocytopenia (platelet count > 40 × 109/l) were without bleeding symptoms. The duration of thrombocytopenia ranged from 1 to 240 months. In all HCV-positive patients, thrombocytopenia occurred as the presenting manifestation of HCV infection.
Among the five HCV-positive thrombocytopenic patients who had a history of blood transfusion, the interval between transfusion of the suspected infectious donation and diagnosis of thrombocytopenia ranged from 10 to 25 years. Serum alanine aminotransferase (ALT) levels were persistently increased in 12/13 patients. Among the four patients who underwent liver biopsy, histological findings were severe chronic active hepatitis (case 1), minimal chronic active hepatitis (cases 8 and 13) and chronic persistent hepatitis (case 10). In addition, mild hypersplenism was only demonstrated in one HCV-positive patient (case 1). This patient presented with bleeding symptoms, a platelet count of 18 × 109/l, a haemoglobin level of 13 g/dl and a white blood cell count of 4·8 × 109/l. Abdominal ultrasonography revealed a mild splenomegaly (13·1 × 6·9 cm) with portal hypertension. Because of extremely low platelet counts and the normal range of haemoglobin and leucocyte counts, it was thought that in this patient hypersplenism was not the major mechanism of thrombocytopenia at presentation. The remaining 12 patients had no symptoms of portal hypertension or splenomegaly. Serum protein electrophoresis showed a monoclonal protein in two HCV-positive thrombocytopenic patients (cases 2 and 13). No other systemic disorders related to HCV infection were found in our series. In the 11 HCV-positive thrombocytopenic patients in whom surface PAIgG was performed, the levels of PAIgG were not elevated and were considered as negative.
After a careful literature search, we have identified a further five well-documented reports of HCV-associated thrombocytopenia (Table III). In these series, a high prevalence of HCV infection was also observed in thrombocytopenic patients. Although it is difficult to draw conclusions on the basis of a literature review of five small series, we found that this disorder was associated with a female sex preponderance, middle age, highly variable platelet levels, mild elevation of ALT and a relatively common history of past exposure to blood products. This was in line with our data.
Platelet response of patients receiving IFN-α or conventional therapy
As shown in Table II, six patients (cases 1, 2, 3, 10, 11 and 13) received IFN-α, three patients (cases 4, 6 and 8) received conventional AITP therapy and four patients (cases 5, 7, 9 and 12) did not receive any treatment at all (because they refused IFN-α treatment and had moderate thrombocytopenia). The six IFN-α-treated HCV-positive thrombocytopenic patients were women, median age 58·6 years (range 48–66). At admission, thrombocytopenia duration ranged from 1 to 240 months (median 54·6). Patients had not received prior specific treatments. Median platelet count was 61 × 109/l (range 18–87) at the onset of IFN-α therapy. The median duration of IFN-α treatment was 14 months (range 6–24 months). After 6 months, two patients (cases 1 and 3) discontinued IFN-α treatment because of side-effects. The remaining four patients have received 11–24 months of IFN-α therapy and are still on treatment. Patients were evaluated periodically during and after therapy. Changes in the platelet counts of IFN-α-treated patients are detailed in Table IV. After initiation of IFN-α treatment, all patients had a significant increase in platelet counts. Five patients (cases 2, 3, 10, 11 and 13) had > 100 × 109/l platelets by the 6th month of therapy with IFN-α. In patient 1, however, the increase in platelet count was minor, with a maximum increase to 38 × 109/l platelets after 6 months of IFN-α therapy. All of our IFN-α-treated patients had a durable (> 1 year) platelet responses. Even two patients (cases 1 and 3), who were treated with a short course of IFN-α, had a durable platelet response. Follow-up of these six IFN-α-treated patients ranged from 11 to 36 months (median 23 months). Interestingly, one patient (case 2) presented with an acute relapse (platelet count 17 × 109/l) 12 months after starting IFN-α therapy. In this patient, the platelet count rapidly returned to 130 × 109/l after a short course of low-dose prednisone (0·25 mg/kg/d for 2 weeks), although IFN-α treatment was continued. PAIgG was not determined at this time. Currently (at 24 months of IFN-α treatment), her platelet count is maintained above 100 × 109/l on a standard dose of IFN-α. Serum ALT levels were normalized in one patient (case 2), reduced to less than twice the upper normal limit in two patients (cases 3 and 10) and did not change in three patients (cases 1, 11 and 13).
In two patients (cases 1 and 3), treatment with IFN-α had to be stopped after 6 months because of depression. In the other four patients, side-effects were moderate, generally limited to asthenia. No clinical symptoms of anaemia or leucopenia were observed during follow-up and no autoimmune diseases occurred.
On the other hand, three patients with HCV-positive thrombocytopenia were managed with conventional therapy for AITP (Table II). Patient 4 had no platelet response after treatment with prednisone and IVIg. Then, she was splenectomized and achieved a good long-term response (platelet count 180 × 109/l), which persisted for longer than 1 year. Patient 6 was treated with prednisone without platelet response. She was subsequently treated with splenectomy and danazol without effect (platelet count 20 × 109/l). Because of bleeding episodes during follow-up, she was treated with IVIg, which produced a very transient response (platelet count 58 × 109/l). In patient 8, treatment with prednisone had no impact on the platelet count, then he responded to splenectomy and maintains a stable platelet count (platelet count 168 × 109/l) after 3 years.
The present study confirms the high prevalence (22·5%) of HCV infection in patients with chronic acquired thrombocytopenia. This association is particularly significant when compared with the prevalence of HCV in our age-matched healthy controls (0·4%). The clinical and laboratory data of our HCV-positive thrombocytopenic patients showed trends similar to those of most other reports of HCV-associated thrombocytopenia ( Silva et al, 1992 ; Pawlotsky et al, 1995 ; Linares et al, 1996 ; Bauduer et al, 1998 ; Gerli et al, 1999 ). In our series, thrombocytopenia was the presenting manifestation of HCV infection in all cases. The degree of thrombocytopenia was generally moderate, however a severe reduction to levels < 20 × 109/l was also observed in two cases. Mild elevation of serum ALT was detected in 92% of cases. In our study, we identified a risk factor for HCV infection (previous exposure to blood products) in 5 of the 13 HCV-positive thrombocytopenic patients. In these five cases, we observed that the time interval between potential HCV infection and thrombocytopenia was about 10–25 years, a result that is in accordance with earlier observations ( Bauduer et al, 1998 ). It would be reasonable to infer that the thrombocytopenia occurs in the late stages of HCV infection (when viral burden is increased) and before the primary symptoms of liver failure have been established.
Treatment of patients with HCV-associated thrombocytopenia remains controversial. IFN-α therapy is the current treatment of choice for chronic hepatitis C, but some side-effects (e.g. myelosuppression, induction of immune thrombocytopenia) limit the use of this cytokine in patients with thrombocytopenia and HCV infection ( Lopez Morante et al, 1992 ; Murakami et al, 1994 ; Dusheiko, 1997; Toccaceli et al, 1998 ). At present, only one patient with HCV-associated thrombocytopenia has been reported as treated with IFN-α and a platelet count response observed ( Durand et al, 1994 ). In the present study, after excluding that thrombocytopenia was due to anti-platelet antibodies, we started a trial of low-dose of IFN-α therapy in six patients with HCV-associated thrombocytopenia. IFN-α therapy achieved a long-term platelet response in all six cases treated. Our findings clearly suggest that IFN-α therapy is not contraindicated in patients with chronic HCV infection and thrombocytopenia.
The ability of IFN-α, an inhibitor of viral replication, to increase platelet counts in HCV-positive thrombocytopenic patients indicates that underproduction of platelets (i.e. infection of megakaryocytes and/or deranged cytokine production) appears to play a dominant role among the complex mechanisms that may be responsible for thrombocytopenia. In this sense, it will be particularly relevant to know whether a relationship exists between platelet response and HCV-RNA levels in HCV-positive thrombocytopenic patients treated with IFN-α. On the other hand, increased splenic platelet pooling mechanisms might be also operative in HCV-associated thrombocytopenia as splenectomy is also effective in some of these patients. It is possible that HCV could promote an intrinsic damage of platelets (i.e. at the level of platelet membrane), leading to increased sequestration of platelets by the spleen. In agreement with other studies ( Durand et al, 1994 ; Kosugi et al, 1997 ; Gerli et al, 1999 ), we found no elevated titres of PAIgG among patients with chronic thrombocytopenia and HCV infection, suggesting that thrombocytopenia is not due to antibodies which target and destroy platelets. The discrepancy with other reports ( Nagamine et al, 1996 ; Bauduer et al, 1998 ) may lie in the different conditions of infection (for example viral load, HCV genotype, stage of the disease) as well as in the pattern of the host's immune response to HCV infection.
In summary, our findings confirm that routine testing for HCV is recommended in all patients with chronic acquired thrombocytopenia. The salutary effect of IFN-α therapy on the platelet count in all treated patients is consistent with the view that megakaryocyte/platelet lineage may be directly or indirectly affected by the virus. To our knowledge, this is the first study to evaluate the efficacy of IFN-α in a cohort of patients with HCV-associated thrombocytopenia. Further studies will be required to determine the real value of this form of therapy.