ITPA Polymorphisms Are Associated with Hematological Side Effects during Antiviral Therapy for Chronic HCV Infection

Background/Objective Genetic polymorphisms in the inosine triphosphatase (ITPA) gene have been associated with the protection from early ribavirin(RBV)-induced hemolytic anemia among patients with chronic hepatitis C virus (HCV) infection. The aim of the present study was to investigate the association between the functional ITPA variants and hematological side effects during antiviral therapy with pegylated interferon (PegIFN) and RBV. Patients and Methods This cohort study included all consecutive Caucasian patients treated for chronic HCV infection with PegIFN and RBV between 2000 and 2009 for whom a serum sample was available for genetic testing. The predicted inosine triphosphate pyrophosphatase (ITPase) activity was based on the genotypes of the SNPs rs1127354 and rs7270101. Decline in hemoglobin (Hb) during antiviral therapy, as well as dose reductions, blood transfusions and use of erythropoietin were assessed. Results In total, 213 patients were included. The predicted ITPase activity was normal among 152 (71%) patients; 61 (29%) patients had ITPase deficiency. By multivariable linear regression, RBV dose in mg per kilogram (Beta 0.09, 95%CI 0.04–0.13, p<0.001) and normal ITPase activity (Beta 0.89, 95%CI 0.64–1.14, p<0.001) were associated with more Hb decline at week 4 of treatment. Patients with normal ITPase activity underwent more dose adjustments of RBV than patients with ITPase deficiency (19(13%) vs 1(2%),p = 0.014) and received erythropoietin more frequently (12 (8%) vs 0 (0%),p = 0.024). Conclusion Genetic variants in the ITPA gene protected against RBV treatment-induced anemia among Caucasian patients with chronic HCV infection. Patients with normal ITPase activity underwent more dose reductions of RBV and received erythropoietin more frequently.


Introduction
HCV infection who were treated with PegIFN alfa-2a or -2b and RBV between 2000 and 2009 in our center [15,16]. The inclusion and exclusion criteria for this study are described elsewhere [15,16]. Briefly, patients were included if they were treated with PegIFN and RBV between 2000 and 2009. Patients treated with conventional interferon and patients co-infected with human immunodeficiency virus or the hepatitis B virus were excluded. In order to prevent confounding by ethnic origin, only Caucasian patients were included in the present study.
The study was conducted in accordance with the guidelines of the Declaration of Helsinki and the principles of Good Clinical Practice. The ethical review board of the Erasmus Medical Center, Rotterdam, The Netherlands approved this study as it was considered to be a low-risk study using retrospective and anonymized patient data. Written informed consent was obtained from each patient for storage of serum samples.

Data acquisition
We obtained baseline data on gender, age, race, body mass index, METAVIR score, HCV genotype, previous interferon-based treatment, platelet count, absolute neutrophil counts, Hb, bilirubin and albumin concentration, glucose levels, presence of hemophilia and use of anticoagulants and antiplatelet therapy, diabetes mellitus (DM), history of heroin use and/or smoking.
During therapy all patients visited the outpatient clinic in one to six weeks intervals. At every visit blood tests were performed and patients were assessed for dose reductions and discontinuation of antiviral therapy. Among patients who were treated within a standard of care protocol, PegIFN and RBV dose reductions were made at the discretion of the treating physician. Patient characteristics such as age, physical condition, virological response, comorbidities and side effects of antiviral therapy were taken into account when considering a dose reduction. All study protocols of the clinical trials stated that dose reductions should be made according to product labels. However, these guidelines were not applied in some patients, due to the expected decrease of antiviral efficacy. Therefore these patients were treated at the discretion of the treating physician as well. The use of blood transfusion and EPO were also registered.

Endpoints
The primary endpoint was the decline in Hb (mmol/L) and platelet count ( Ã 10 9 /L) which was assessed at week 4 (+/-7 days). This time point was chosen in order to limit the influence of dose reductions on these hematological outcomes [7]. A clinically significant decline in Hb was defined as a decrease of at least 3.0 g/dL (1.86 mmol/L) or an absolute value lower than 10 g/dL (6.21 mmol/L). These thresholds were also used in other studies on ITPase deficiency [13,14,17]. Anemia was defined according to the thresholds used by the World Health Organization; for women a Hb concentration below 7.45 mmol/L and for men below 8.1 mmol/L were used as cut-off. Thrombocytopenia was defined as a platelet count below 150 Ã 10 9 /L. A clinical relevant thrombocytopenia was defined as a platelet count below 50 Ã 10 9 /L, since current guidelines advise to reduce the dose of PegIFN when platelet counts fall below 50 Ã 10 9 /L [18].
As a secondary endpoint, the decline in Hb and platelet counts were also assessed at week 8 and 12 (+/-7 days) of antiviral therapy, as well as the nadir values of these parameters.
Sustained virological response (SVR) was defined as HCV RNA negativity in blood six months after cessation of antiviral therapy. Dose reductions of RBV and PegIFN, as well as the administration of EPO and blood transfusion, were considered as clinical endpoints.

Genotyping methods
Serum samples stored at -20°or -80°Celsius were used for DNA extraction and genotyping procedures, which were carried out centrally at LGC genomics. Purified genomic DNA of !5 ng was used for genotyping. Genotypes were assigned using all of the data from the study simultaneously. Genotype sequences were derived from NCBI. Genetic analyses were performed at the polymorphic sites rs12979860 (19:39248147, near IL28B, also known as interferon-λ3), rs1127354 (20:3213196, ITPA-1) and rs7270101 (20:3213247, ITPA-2). The IL28B SNP rs12979860 was chosen, since it best describes the association with sustained SVR for all genotypes [19][20][21]. Linkage disequilibrium and Hardy-Weinberg equilibrium (HWE) were tested for these SNPs using SNAP and OEGE [22].

Predicted ITPase activity
As freshly acquired erythrocytes were lacking in order to directly measure ITPase activity, the predicted ITPase activity was based on genotypes of both ITPA-1 and ITPA-2 as is determined by previous analyses [23,24]. Patients with normal ITPase activity (i.e. 100%) were defined as patients with the combined presence of CC-genotype and AA-genotype for rs1127354 and rs7270101 respectively (S1 Table). Patients with less than 100% ITPase activity were defined as patients with ITPase deficiency (non-CC-genotype and non-AA-genotype for rs1127354 and rs7270101 respectively, S1 Table).

Statistical Analysis
Continuous variables were summarised as median (interquartile range [IQR]) and categorical variables as frequencies (percentages). Comparisons between groups were performed using X 2 test for categorical variables or the Mann-Whitney U test for comparing medians. The genetic association analyses for ITPA and IL28B polymorphisms consisted of a dominant genetic model (CC-and AA-genotype vs non-CC-and non-AA-genotype for ITPA and CC-genotype vs non-CC genotype for IL28B) [14,25]. Linear regression analysis determined which variables were associated with the absolute decline in Hb and platelet count at week 4. Logistic regression was performed to determine which variables were associated with a clinically significant Hb decline at week 4, SVR and virological relapse. For decline in Hb and platelet count, a sensitivity analysis was performed which excluded patients whom were treated with a high PegIFN dose induction regimen. Age, sex and variables with a p-value of 0.2 in univariable analyses were included in multivariable analyses. All final models were created by using a backward stepwise method, in order to select the variables that were significantly and independently associated. Potential confounding was checked. All statistical tests were two-tailed, and p<0.05 was considered to be statistically significant. The significance level of interactions was set at 0.01 in order to correct for multiple testing. SPSS version 21.0 (SPSS, Chicago, IL) was used.

Patients
In total, 321 consecutive patients with chronic HCV infection were treated with PegIFN and RBV between 2000 and 2009, of which 256 were Caucasian. Two hundred thirteen (83%) of these patients, who had a sample available for genetic testing and could be genotyped for both ITPA polymorphisms, were included in the current analyses (S1

ITPase deficiency and on-treatment hemoglobin concentration
In total, 182 (85%) patients experienced at least one episode of anemia during antiviral treatment and in 157 (74%) patients a clinically significant decline in Hb was reported. At baseline, median Hb concentration was 9.2 mmol/L (IQR 8.7-9.9) for patients with normal ITPase activity and 9.3 mmol/L (IQR 8.7-9.9) for patients with ITPase deficiency (p = 0.793). At treatment weeks 4, 8 and 12, median Hb concentration was significantly lower among patients with normal ITPase activity (p<0.001 for all time points, Fig 1A). The nadir median Hb concentration was lower for patients with normal ITPase activity compared to patients with ITPase deficiency, respectively 6.6 mmol/L (IQR 5.8-7.2) and 7.2 mmol/L (IQR 6.6-8.1, p<0.001). Furthermore, patients with ITPase deficiency had a lower occurrence of a clinically significant decline in Hb as compared to patients without ITPase deficiency at week 4, 8 and 12 (p<0.001 for all timepoints, S2 Fig).
Multivariable linear regression analysis showed that baseline platelet count (Beta -0.02, 95% CI -0.04--0.00, p = 0.022), baseline Hb concentration (Beta 0.44, 95% CI 0.29-0.59, p<0.001), RBV dose per kilogram body weight (Beta 0.09, 95% CI 0.04-0.13, p<0.001) and normal ITPase activity (Beta 0.89, 95% CI 0.64-1.14, p<0.001) were associated with the absolute decline in Hb concentration at week 4 of treatment ( Table 2). The interaction terms between the variables in the final model were not statistically significant. Also in a sensitivity analysis for which patients with a PegIFN induction regimen were excluded the presence of normal ITPase activity was associated with the occurrence of a significant decline in Hb concentration at week 4 (Beta 0.91, 95% CI 0.65-1.18, p<0.001).

ITPase deficiency, dose reductions, EPO and blood transfusions
In total, 20 (9%) patients underwent at least one dose reduction of RBV and 44 (21%) patients at least one dose reduction of PegIFN. At least one blood transfusion was given to 27 (13%) patients and 12 (6%) patients received at least one dose of EPO. Nineteen (13%) patients with normal ITPase activity underwent at least one dose reduction of RBV, whereas one (2%) patient with ITPase deficiency underwent at least one dose reduction (p = 0.014, Fig 2). The dose of PegIFN was reduced among 36 (24%) patients with normal ITPase activity and among eight (13%) patients with ITPase deficiency (p = 0.085). Blood transfusion and EPO were administered to 23 (15%) and twelve (8%) patients with normal ITPase; and to 4 (7%) and none (0%) of the patients with ITPase deficiency (p = 0.089 and p = 0.024, respectively).

Discussion
This large cohort study in a real world setting found that genetic polymorphisms in the ITPA gene, resulting in ITPase deficiency, were associated with less Hb decline during pegIFN and RBV therapy among Caucasian patients with chronic HCV infection. Patients with normal ITPase activity were at higher risk to undergo dose reductions of both PegIFN and RBV, receive blood transfusions and be administered EPO. These patients may benefit from early administration of supportive hematopoietic growth factors in order to improve treatment adherence. Furthermore it could be a helpful tool in the decision of adding RBV, which has shown to be of additional value for some IFN-free regimens [1,2]. Patients with ITPase deficiency (27% of our cohort) experienced lower declines in Hb concentration compared to patients with normal ITPase activity during the first four weeks of antiviral treatment. As expected, besides the presence of normal ITPase activity, RBV dose was also associated with a significant decline in Hb at week 4. The results of the present study confirm observations described in previous studies [9,13,14]. More important, our study revealed that patients with normal ITPase activity underwent more dose adjustments for RBV (13% vs 2%), and more often EPO was administered (8% vs 0%). Two previous studies among patients with HCV genotype 2 and 3 found no association between the ITPase activity and the need for RBV dose reductions [9,14]. In the first study, including 238 patients with HCV genotype 2 and 3, RBV dose was reduced in 35 patients (15%) [14]. The presence of ITPase deficiency was not statistically associated with a lower hazard ratio (HR) for RBV dose reductions (HR 0.80, 95% CI 0.35-1.71, p = 0.57). This study used weight-based RBV and the dose was not reduced until Hb concentration fell below 5.9 mmol/L (i.e. 9.5 g/dL). The second study included 349 patients with HCV genotype 2 and 3, which showed that 6% of the patients did not receive the full planned dose of RBV. This low rate can be attributed to the use of low-dose RBV (i.e. 800 mg/ day) and the shorter treatment duration among this patients. In contrast, and in line with our data, a study among solely HCV genotype 1 infected patients did describe that the presence of ITPase deficiency resulted in fewer RBV dose reductions [13]. A high rate of RBV dose reductions was found in this study (47%), attributed to the fact that they included missed doses in this rate. Only four patients in this trial received EPO. In general, these data were based on clinical trial cohorts, solely including selected patients with specific HCV genotypes, different doses of RBV as well as limited use of blood transfusion and EPO. Furthermore these studies used strict dosing rules, instead of dose adjustments according to the treating physician, which is more representative for the clinical setting. Recently, another real-world study, also including all HCV genotypes, found that patients with ITPase deficiency required less RBV dose reductions and less EPO [17]. In 18% of the patients with ITPA deficiency a RBV dose reductions was required, whereas 41% of the patients with normal ITPase activity underwent at least one dose reductions of RBV. However, this study used strict dosing rules as well, which is not an optimal reflection of daily practice. Moreover, they did not report on the ethnic background of patients, which is essential in genetic studies [25].
The presence of ITPase deficiency was also associated with a deeper decline in platelet count at week 4 of treatment, which could be explained by the absence of thrombocytosis in reaction to the hemolytic anemia. This is in line with previous studies which showed that reduced ITPase activity, which protects against RBV-induced anemia, is associated with the occurrence of treatment-induced thrombocytopenia [9,10,26]. ITPase deficiency was not related to the occurrence of a platelet count below 50 x 10 9 /L, but this rarely happens among patients with chronic HCV infection who are treated with pegIFN and RBV. Indeed, also in the current study, only 10% of patients had a platelet count below 50 x 109/L during their treatment course. Nevertheless, it could be a predictive tool among patients with cirrhosis, who are more prone to develop severe thrombocytopenia [27].
We did not find ITPase deficiency to be associated with virological response to PegIFN and RBV, perhaps because of limited power in our current study. Data on the influence of ITPase activity on the virological response is inconclusive, probably due to the inclusion of various HCV genotypes, heterogeneity in treatment regimens and the various ways in which the association was analysed. Our results are in line with the largest study to date, among patients with HCV genotype 1, which did not show a relation between ITPase deficiency and SVR either [13]. Nevertheless, it could be hypothesised that the higher frequency of dose reductions among patients with normal ITPase activity would compromise treatment efficacy, as was reported previously [5,6]. However, the mechanism by which anemia and ITPase deficiency influence virological response is still not fully unravelled.
The clinical importance of ITPA polymorphisms in the era of regimes with DAAs could be debated. Although limited data is available, previous studies among patients treated with triple therapy including telaprevir also showed that a ITPA polymorphism (rs1127354) was associated with the development of on-treatment anemia [28][29][30]. Unfortunately, these studies are solely among patients of Asian ancestry, and these patients are monoallellic for ITPA polymorphism rs7270101. In contrast to these results, the limited data available for Caucasian patients showed contrasting results [31]. Among patients with advanced hepatic fibrosis, ITPA polymorphisms were associated with the severity of Hb decline at week 4, but not at week 12 of therapy. This effect was attributed to the increased plasma levels of RBV after the first 4 weeks of triple therapy [32]. Recently, it was shown that ITPA polymorphisms were also associated with anemia during IFN-free therapy [33]. Studies on the impact of RBV dose reductions among patients treated with triple therapy, including PegIFN, RBV and boceprevir, demonstrated that a lower dose of RBV did not affect SVR rates [34,35]. Moreover, RBV dose reductions among patients with HCV genotype 2 and 3, treated with IFN-free regimens, also suggested no effect on SVR rates [36,37]. Nevertheless, as RBV will still be a component of IFN-free treatment regimens, ITPA polymorphisms can select patients who are prone to develop RBV-induced hemolytic anemia in order to apply more conservative and/ or earlier dose reductions or early administration of supportive agents. Secondly, it may be used as an additive tool to select a specific IFN-free regimen for the individual patient as not all regimens require addition of RBV for optimal virological efficacy. However, more data is needed in order to conclude on the clinical utility of these SNPs.
A limitation of this study is that RBV concentrations were not available. A previous study showed higher RBV concentrations at week 4 among patients with normal ITPase activity compared to patients with ITPase deficiency [9]. This could be explained by the reduced hemolytic anemia that was present in the patients with ITPase deficiency, generating a larger distribution volume for the intracellular forms of RBV, leading to lower extracellular concentrations of RBV. On the other hand, lower plasma concentrations of RBV could have led to the protection against anemia. In contrast, another study among 546 patients showed no association between ITPase deficiency and RBV levels (p = 0.11) [13]. Finally, due to the retrospective character of the study, we were not able to measure ITPase activity directly, as this requires freshly acquired erythrocytes. Nevertheless, classification of the predicted ITPase activity by combining the two genotypes has been suggested to be reliable [23,24].
In conclusion, this real-world study showed that ITPase deficiency is associated with the protection against hemolytic anemia among Caucasian patients with chronic HCV infection who are treated with PegIFN and RBV. This led to less dose reductions of RBV and PegIFN as well as less administration of blood transfusions and EPO. Since treatment efficacy is hampered by dose reductions, patients with normal ITPase activity may benefit from early strategies in order to improve treatment adherence.