From August 2004 to March 2009, 117 consecutive type C cirrhotic patients underwent LS at our institute for the induction of PEG-IFN/RBV therapy, and DNA was available for genotyping from 37 of these. They were compared to 90 cirrhotic patients who did not undergo splenectomy before induction of PEG-IFN/RBV therapy in the same period at the Department of General Internal Medicine, Kyushu University. Patients in the control group were selected on account of their platelet count <10⁵/ml or by METAVIR fibrosis stage F4 13 . Therefore, 127 patients were enrolled in the current study, which was approved by the Ethics Committee of Kyushu University.

Genomic DNA was extracted from the patients’ spleen tissues obtained at operation, or from peripheral blood mononuclear cells in the control group. IL28B genetic polymorphism (rs8099917) and ITPA genetic polymorphism (rs1127354) were genotyped using StepOnePlus™ real-time PCR system (Applied Biosystems, Carlsbad, CA, USA).

VR was defined as a lack of HCV RNA in response to the treatment regimen, regardless of whether a relapse occurred when treatment was terminated. SVR was defined as undetectable HCV-RNA 24 weeks after the end of the therapy. Patients who had achieved VR but had showed relapse were defined as end of treatment response (ETR)-relapse. Hemoglobin (Hb) level or platelet count was assessed before splenectomy, at induction of therapy, and every 4 weeks during the course of PEG-IFN/RBV therapy.

Total RNA was extracted from exenterate spleen specimens using ISOGEN (Nippon Gene, Tokyo, Japan) and reverse transcription was performed using SuperScriptIII (Invitrogen, Carlsbad, CA, USA). Quantitative PCR was performed using SYBR® Green assay (Applied Biosystems) on the LightCycler®480 Real-Time PCR system (Roche Applied Sciences, Indianapolis, IN, USA). Specific primers for ISG15 were as follows: 5^′-agcgaa ctcatctttg-3^′ for sense primer 5^′-cagctctgacaccgacatgga-3^′ for antisense primer. Specific primers for protein kinase R (PKR) were 5^′-acgtgtgagtcccaaagcaac-3^′ for sense and 5^′-ctgagaccattcataagcaacg-3^′ for antisense. β-Actin expression was used for endogenous control with 5^′-ct ggcaccacac cttctacaatg-3^′ for sense primer and 5^′-ggcgtacagggatagcacagc-3^′ for antisense primer.

All data were analyzed using JMP® statistical software (SAS Institute, Cary, NC, USA). A χ² test was performed for qualitative variables and a Wilcoxon test was performed for quantitative variables.

Characteristics of patients who underwent splenectomy (Spx) and those who did not (non-Spx) are shown in Table  1. Although patients in the non-Spx group were selected on account of their platelet count <10⁵/μl, count before surgery in the Spx group was lower than that in the non-Spx group (6.1 × 10⁴/μl vs 8.7 × 10⁴/μl, respectively, P < 0.0001), as well as their Hb level (12.5 g/dl vs 13.2 g/dl, respectively, P = 0.03). Both alanine aminotransferase and γ-glutamyl transpeptidase levels were significantly higher in the non-Spx group (55 IU/l vs 91 IU/l, P = .001, 45 IU/l vs 60 IU/l, P = 0.02, respectively), but albumin level was higher in the non-Spx group (3.3 mg/dl vs 3.9 mg/dl, P < 0.0001). The progression of cirrhosis was considered to be more severe in the Spx group. Neither pretreatment viral load nor the frequency of HCV genotype differed among these groups. The allele frequency of both IL28B (rs8099917) and ITPA (rs1127354) genetic polymorphisms was similar in these groups.

HCV; hepatitis C virus, IL28B; interleukin-28B, ITPA; inosine triphosphate pyrophosphatase, n.s ; not significant.

Spx; splenectomy.

Fifty-four of 127 patients achieved a SVR (42.5%). Patients carrying IL28B the major genotype (TT allele at rs8099917, n = 88) showed a significantly higher SVR rate compared to those carrying the minor genotype (TG or GG, n = 39) (56.8% vs 10.3%, P < 0.0001, Figure  1A). To clarify the effect of splenectomy on the outcome of PEG-IFN/RBV therapy, SVR rate was compared between the Spx and non-Spx groups. Although the ETR-relapse rate in the Spx group seemed higher than in the non-Spx group (27.0% vs 18.9%), there was no significant difference in SVR rate between the two groups (43.2% vs 42.3%, P =0.91, Figure  1B).

SVR rate among each IL28B genotype in the Spx and non-Spx groups was compared. As shown in Figure  1C, IL28B genetic variants were significantly associated with SVR rate in the non-Spx group (58.9% vs 4.1%, P < 0.0001). SVR rate in the Spx group was not different between IL28B genotype (50% vs 27.3%, P=0.09), probably because of small number, but VR rate (SVR + ETR-relapse) of Spx group was higher in IL28B major group compared to those in minor group (80.8% vs 45.5%, P <0 .05). Interestingly, patients carrying the minor genotype in the Spx group showed a significantly higher SVR rate than those in the non-Spx group (27.3% vs 3.6%, P < 0.05), whereas there was no difference in SVR rate among patients carrying the major genotype between the SPx and non-Spx groups (50.0% vs 59.7%, P = 0.40).

It is well-known that IFN response is correlated with HCV genotype. Therefore, we made the same analysis on 100 patients infected with only genotype 1 HCV. The similar results were obtained, but no statistical difference in SVR rate among IL28B minor allele carriers between Spx- and non-Spx-group, probably because of small number (20.0% vs 3.7%, data not shown).

Next, we checked the prognostic factors for achieving SVR among IL28B minor allele carriers. As shown in Table  2, splenectomy was the only significant factors for SVR with odds ratio of 10.1 (p<0.05).

CI; confidence interval, HCV; hepatitis C virus, IL28B; interleukin-28B, ITPA; inosine triphosphate pyrophosphatase, n.s ; not significant, Spx; splenectomy.

Hb level and platelet count at each time point were available for 30 patients in the non-Spx group. Although Hb decline at 4 weeks after the initiation of PEG-IFN/RBV tended to be greater in patients carrying the CC allele than CA/AA allele (1.34 g/dl vs 0.96 g/dl, P = 0.09, Figure  2A), there was no significant difference among them. However, patients carrying the CC allele in the non-Spx group were more susceptible to anemia than those carrying the CA/AA allele (Hb decline of 1.74 g/dl vs 0.60 g/dl, P < 0.05), whereas there was no difference in Hb decline between patients carrying each genotype in the Spx group (0.96 g/dl vs 1.28 g/dl, P = 0.89, Figure  2B).

In addition, progression of anemia among patients carrying the CC allele was compared in each group (Figure  2C). Before splenectomy, Hb level in patients in the Spx group was lower than in the non-Spx group (13.2 g/dl vs 12.3 g/dl, P < 0.05). In the Spx group, Hb level did not change after splenectomy (12.3 g/dl vs 12.5 g/dl, P = 0.50) and was still lower than that in the non-Spx group (13.2 g/dl vs 12.5 g/dl, P < 0.05). However, Hb levels at 4 and 8 weeks after the initiation of PEG-IFN/RBV were similar in each group (11.3 g/dl vs 11.7 g/dl, P = 0.55, 10.8 g/dl vs 10.8 g/dl, P = 0.89, respectively).

Progression of thrombocytopenia in patients carrying each ITPA genotype was compared in each group. There was no difference in platelet count before PEG-IFN/RBV between patients carrying the CC allele or CA/AA allele (86.2 × 10³/μl vs 86.9 × 10³/μl, P = 0.86). Progression of thrombocytopenia tended to be worse in patients carrying the CA/AA allele, as reported previously, but there was no significant difference in platelet count at 4, 8 and 12 weeks after initiation of PEG-IFN/RBV (69.9 × 10³/μl vs 63.2 × 10³/μl; P = 0.26, 71.5 × 10³/μl vs 61.3 × 10³/μl; P = 0.23, 69.5 × 10³/μl vs 66.2 × 10³/μl, P = 0.28, respectively, Figure  3A).

Compared with pre- and postoperative platelet count, although splenectomy improved thrombocytopenia in whole splenectomy patients (61.1 × 10³/μl vs 168.7 × 10³/μl, P < 0.01, data not shown), there was no difference in pre- or postoperative platelet count between patients carrying the CC and CA/AA allele (60.4 × 10³/μl vs 55.4 × 10³/μl; P = 0.68, 165.8 × 10³/μl vs 160.9 × 10³/μl, P = 0.86, respectively). As shown in Figure  3B, however, platelet counts in splenectomy patients carrying the CA/AA allele tended to be lower than in those carrying the CC allele at 4 and 8 weeks after initiation of PEG-IFN/RBV (118.8 × 10³/μl vs 91.6 × 10³/μl; P = 0.12, 117.4 × 10³/μl vs 89.3 × 10³/μl, P = 0.14, respectively) without any significant difference.

To clarify the efficacy of splenectomy in thrombocytopenia, platelet counts were compared between the Spx and non-Spx groups at each time point (Figure  3C). The patients in the non-Spx group were selected on account of their platelet count <10⁵/μl, but those whose platelet count was <50–60 × 10³/μl were seldom induced with PEG-IFN/RBV, which led to a significant difference in platelet count before splenectomy (86.7 × 10³/μl vs 58.9 × 10³/μl, P < 0.01). However, splenectomy increased the platelet count in the Spx group to overtake that in the non-Spx group (86.7 × 10³/μl vs 164.3 × 10³/μl, P < 0.01), and its dominance was maintained even at 4 and 8 weeks after initiation of PEG-IFN/RBV (67.7 × 10³/μl vs 112.2 × 10³/μl; P < .01, 68.1 × 10³/μl vs 111.3 × 10³/μl, P = 0.01, respectively).

The correlation of IL28B and ITPA genetic variants with the outcome of PEG-IFN/RBV therapy following splenectomy was demonstrated. IL28B genetic variant has also been reported to correlate with hepatic ISG expression among chronic hepatitis C patients 14 . Whether IL28B genetic variant also correlated with splenic ISG expression has yet to be determined. Surprisingly, as shown in Figure  4, both ISG15 and PKR expression in the spleen was significantly higher in patients carrying the TT allele than the TG/GG allele, which was opposite to liver tissue as described previously (0.03 vs 0.02; P < 0.05, 0.13 vs 0.10; P < 0.05, respectively).