HEPATOLOGY, Vol. 62, No. 5, 2015
References 1. Garcıa-Alvarez M, Pineda-Tenor D, Jimenez-Sousa MA, FernandezRodrıguez A, Guzman-Fulgencio M, Resino S. Relationship of vitamin D status with advanced liver fibrosis and response to hepatitis C virus therapy: a meta-analysis. HEPATOLOGY 2014;60:1541-1550. 2. Kitson MT, Sarrazin C, Toniutto P, Eslick GD, Roberts SK. Vitamin D level and sustained virologic response to interferon-based antiviral therapy in chronic hepatitis C: a systematic review and meta-analysis. J Hepatol 2014;61:1247-1252. 3. Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. HEPATOLOGY 2011;53:1118-1126. 4. Falleti E, Bitetto D, Fabris C, Fattovich G, Cussigh A, Cmet S, et al. Vitamin D binding protein gene polymorphisms and baseline vitamin D levels as predictors of antiviral response in chronic hepatitis C. HEPATOLOGY 2012; 56:1641-1650. 5. Bitetto D, Bortolotti N, Falleti E, Vescovo S, Fabris C, Fattovich G, et al. Vitamin A deficiency is associated with hepatitis C virus chronic infection and with unresponsiveness to interferon-based antiviral therapy. HEPATOLOGY 2013;57:925-933.
CORRESPONDENCE
1643
Nevertheless, we appreciate the suggestion, and we will take it into account for future analyses. Fourthly, we included a study5 with a reduced percentage of patients (35%) receiving vitamin D supplementation. When we excluded this report, the odds ratio value did not change significantly (odds ratio 5 0.58, 95% CI 0.32-0.99). Finally, it is noteworthy that the two meta-analyses have sought the same goal from different perspectives. Kitson et al.C ¸ s metaanalysis included seven articles and four conference abstracts, while our meta-analysis included 11 studies. Only two articles were analyzed in both meta-analyses. Both studies showed differences in the inclusion/exclusion criteria and in the analysis strategy; thus, Kitson et al. compared baseline levels of 25(OH)D, but we analyzed the data considering several thresholds frequently used to define 25(OH)D status in clinical practice. For these reasons, we consider that both studies are complementary and might be useful.
DANIEL PINEDA-TENOR, PH.D. MO´NICA GARCI´A-A´LVAREZ, PH.D. MARI´A A. JIME´NEZ-SOUSA, PH.D. AMANDA FERNA´NDEZ-RODRI´GUEZ, PH.D. SALVADOR RESINO, PH.D. Viral Infection and Immunity Unit Centro Nacional de Microbiologıa Instituto de Salud Carlos III Majadahonda, Madrid, Spain
C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27797 Potential conflict of interest: Nothing to report.
References Reply: We appreciate the interest of Kitson et al. in our recent article.1 We found an association between the cutoff of 20 ng/mL of 25-hydroxyvitamin D (25[OH]D) and sustained virologic response in patients on pegylated interferon-a/ribavirin therapy (odds ratio 5 0.53, 95% confidence interval 0.31-0.91), which is in disagreement with the meta-analysis published by Kitson et al. Thus, these authors suggested some issues that should be clarified. Firstly, Kitson et al. pointed out that our analysis included three different studies with the same Italian cohort.2-4 However, these studies were published by prestigious journals during consecutive years and provided independent and remarkable information, each of them including a different number of patients. Nevertheless, the significant association was lost if we reanalyzed it considering only one of these studies. Secondly, with regard to the studies that were not included in our meta-analysis, we carefully selected articles from PubMed, SCOPUS, LILACS, and the Cochrane Library. When data were unclear, incomplete, or not explicitly reported, we contacted the corresponding author up to three times in order to obtain additional information. Unfortunately, some authors did not reply or refused to provide the requested data. Thirdly, we only included studies in which vitamin D was assessed as a categorical variable (the thresholds for deficiency, suboptimal, and optimal 25[OH]D levels) because we considered that it was more relevant and useful than the continuous variable.
1. Garcıa-Alvarez M, Pineda-Tenor D, Jimenez-Sousa MA, FernandezRodrıguez A, Guzman-Fulgencio M, Resino S. Relationship of vitamin D status with advanced liver fibrosis and response to hepatitis C virus therapy: a meta-analysis. HEPATOLOGY 2014;60:1541-1550. 2. Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. HEPATOLOGY 2011;53:1118-1126. 3. Falleti E, Bitetto D, Fabris C, Fattovich G, Cussigh A, Cmet S, et al. Vitamin D binding protein gene polymorphisms and baseline vitamin D levels as predictors of antiviral response in chronic hepatitis C. HEPATOLOGY 2012;56:1641-1650. 4. Bitetto D, Bortolotti N, Falleti E, Vescovo S, Fabris C, Fattovich G, et al. Vitamin A deficiency is associated with hepatitis C virus chronic infection and with unresponsiveness to interferon-based antiviral therapy. HEPATOLOGY 2013;57:925-933. 5. Bitetto D, Fabris C, Fornasiere E, Pipan C, Fumolo E, Cussigh A, et al. Vitamin D supplementation improves response to antiviral treatment for recurrent hepatitis C. Transpl Int 2011;24:43-50.
Author names in bold designate shared co-first authorship. C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27796 Potential conflict of interest: Nothing to report.
Is Pentoxifylline Still an Option in Severe Alcoholic Hepatitis? To the Editor: We read with interest the Hepatology Elsewhere piece by Caballeria1 describing the study by Park et al. that investigated the rela-
tive efficacies of prednisolone and pentoxyfylline in acute alcoholic hepatitis in a noninferiority study.2 Caballeria states that the trial concluded that the efficacy of pentoxifylline is lower than that of prednisolone. However, the noninferiority endpoint of that study
1644
CORRESPONDENCE
was inconclusive, and whereas the authors concluded that the two treatments were nonequivalent, insufficient methodological information was presented for this to be inferred. The design and interpretation of noninferiority studies is complex and requires detailed analysis to prevent erroneous conclusions being drawn. Arbitrary decisions regarding the limit of the noninferiority margin are needed in the design of such studies, and the rationale for the limits used should be clearly stated. Park et al. set the limit at D15%. The reasoning behind this was not presented, but three studies were cited. The most relevant of these was the meta-analysis from Mathurin et al.3 Here, the difference between prednisolone and the other treatment arm was approximately 20%. This was the best available evidence when the trial was designed. Guidance for the design of noninferiority studies is for a noninferiority margin of 50% of the treatment benefit with the active control, in this case prednisolone,4 suggesting that the noninferiority limit set was too lenient, thus risking a false-positive result (indicating that pentoxifylline is erroneously noninferior to prednisolone). In addition, regarding the equivalence endpoint, the assumptions of the relative efficacies of prednisolone and pentoxifylline were not given for the power calculation. For nonequivalence to be inferred, the assumed efficacy of both agents would need to be identical and this was not stated. Whereas it may be that pentoxyfilline is not equivalent to prednisolone in the treatment of severe alcoholic hepatitis, the design and/or reporting of the study from Park et al. prevents clear conclusions from being drawn. A clearer estimate of the relative efficacy will be provided by a large, blinded, randomized, controlled trial comparing steroids and pentoxifylline with a placebo control that has recently been completed.5
HEPATOLOGY, November 2015
DOI 10.1002/hep.27782 Potential conflict of interest: Dr. Elsharkawy advises Gilead.
Reply: I would like to thank Townsend et al. for their comments. I agree with them that it should be more appropriate, according to the results of Park et al.,1 to conclude that the efficacy of pentoxifylline was noninferior, instead of lower than those of corticosteroids. I also agree that the noninferiority studies are difficult and the cutoff to evaluate significant differences is, in some way, arbitrary. The 15% limit established by Park et al. was perhaps too conservative, although it seems reasonable according to the results of the meta-analysis of Mathurin et al.2 Pentoxifylline is still considered the second therapeutic option in treatment of severe alcoholic hepatitis. However, the efficacy of pentoxifylline in this clinical situation has not been confirmed in different scenarios, as summarized in a recent systematic review3 and in an editorial.4 The final results of the STOPAH trial were presented in the past meeting of the American Association for the Study of Liver Diseases.5 This study, designed to evaluate the therapeutic benefit of both prednisolone and pentoxifylline, was performed in 1,053 patients and was unable to demonstrate an effect in disease progression or survival. Thus, Thursz et al. conclude that pentoxifylline should no longer be used for treatment of severe alcoholic hepatitis. This study reinforces the conclusion of my comment that we are still far from the ideal treatment of severe alcoholic hepatitis, and that it is urgent to identify new targets for more specific and effective therapies.
SARAH A. TOWNSEND, M.R.C.P. IAN A. ROWE, M.R.C.P., PH.D. AHMED M. ELSHARKAWY, M.R.C.P., PH.D. Queen Elizabeth Hospital Liver and Hepatobiliary Unit Birmingham, United Kingdom
JUAN CABALLERIA, M.D., PH.D. Liver Unit Hospital Clınic IDIBAPS CIBERehd Barcelona, Spain
References
References
1. Caballeria J. Is pentoxifylline still an option in severe alcoholic hepatitis? HEPATOLOGY 2015;61:1425-1427. 2. Park SH, Kim DJ, Kim YS, Yim HJ, Tak WY, Lee HJ, et al. Pentoxiphylline vs. corticosteroids to treat severe alcoholic hepatitis: a randomised, non-inferiority open trial. J Hepatol 2014;61:792-798. 3. Mathurin P, Mendenhall CL, Carithers RL, Ramond MJ, Maddrey WC, Garstide P, et al. Corticosteroids improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol 2002;36:480-487. 4. Peggio C, Elbourne D, Altman D, Pocock S, Evans S; CONSORT Group. Reporting of non-inferiority and equivalence randomised trials: an extension of the CONSORT statement. JAMA 2006;295:1152-1160. 5. Forrest E, Mellor J, Stanton L, Bowers M, Ryder S, Austin A, et al. Steroids or Pentoxiphylline for Alcoholic Hepatitis (STOPAH): study protocol for a randomised controlled trial. Trials 2013;14:262.
1. Park SH, Kim DJ, Kim YS, Yim HJ, Tak WY, Lee HJ, et al. Pentoxifylline vs corticosterois to treat severe alcoholic hepatitis: a randomised, non-inferiority open trial. J Hepatol 2014;61:792-798. 2. Mathurin P, Mendenhall CL, Carithers RL, Ramond MJ, Maddrey WC, Garstide P, et al. Corticosteroid improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol 2002;36:480-487. 3. Parker R, Amstrong MJ, Corbett C, Rowe IA, Houlihan DD. Systematic review: pentoxifylline for the treatment of severe alcoholic hepatitis. Aliment Pharmacol Ther 2013;37:845-854. 4. Louvet A. Prednisolone vs pentoxifylline for severe alcoholic hepatitis. J Hepatol 2014;61:723-724. 5. Thursz MR, Richardson P, Allison ME, Auastin A, Bowers M, Day CP, et al. Steroids or pentoxifylline for alcoholic hepatitis: results of the STOPAH trial. HEPATOLOGY 2014;60(Suppl):1267A.
Author names in bold designate shared co-first authorship. C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com.
C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27781 Potential conflict of interest: Nothing to report.
References 1. Garcıa-Alvarez M, Pineda-Tenor D, Jimenez-Sousa MA, FernandezRodrıguez A, Guzman-Fulgencio M, Resino S. Relationship of vitamin D status with advanced liver fibrosis and response to hepatitis C virus therapy: a meta-analysis. HEPATOLOGY 2014;60:1541-1550. 2. Kitson MT, Sarrazin C, Toniutto P, Eslick GD, Roberts SK. Vitamin D level and sustained virologic response to interferon-based antiviral therapy in chronic hepatitis C: a systematic review and meta-analysis. J Hepatol 2014;61:1247-1252. 3. Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. HEPATOLOGY 2011;53:1118-1126. 4. Falleti E, Bitetto D, Fabris C, Fattovich G, Cussigh A, Cmet S, et al. Vitamin D binding protein gene polymorphisms and baseline vitamin D levels as predictors of antiviral response in chronic hepatitis C. HEPATOLOGY 2012; 56:1641-1650. 5. Bitetto D, Bortolotti N, Falleti E, Vescovo S, Fabris C, Fattovich G, et al. Vitamin A deficiency is associated with hepatitis C virus chronic infection and with unresponsiveness to interferon-based antiviral therapy. HEPATOLOGY 2013;57:925-933.
CORRESPONDENCE
1643
Nevertheless, we appreciate the suggestion, and we will take it into account for future analyses. Fourthly, we included a study5 with a reduced percentage of patients (35%) receiving vitamin D supplementation. When we excluded this report, the odds ratio value did not change significantly (odds ratio 5 0.58, 95% CI 0.32-0.99). Finally, it is noteworthy that the two meta-analyses have sought the same goal from different perspectives. Kitson et al.C ¸ s metaanalysis included seven articles and four conference abstracts, while our meta-analysis included 11 studies. Only two articles were analyzed in both meta-analyses. Both studies showed differences in the inclusion/exclusion criteria and in the analysis strategy; thus, Kitson et al. compared baseline levels of 25(OH)D, but we analyzed the data considering several thresholds frequently used to define 25(OH)D status in clinical practice. For these reasons, we consider that both studies are complementary and might be useful.
DANIEL PINEDA-TENOR, PH.D. MO´NICA GARCI´A-A´LVAREZ, PH.D. MARI´A A. JIME´NEZ-SOUSA, PH.D. AMANDA FERNA´NDEZ-RODRI´GUEZ, PH.D. SALVADOR RESINO, PH.D. Viral Infection and Immunity Unit Centro Nacional de Microbiologıa Instituto de Salud Carlos III Majadahonda, Madrid, Spain
C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27797 Potential conflict of interest: Nothing to report.
References Reply: We appreciate the interest of Kitson et al. in our recent article.1 We found an association between the cutoff of 20 ng/mL of 25-hydroxyvitamin D (25[OH]D) and sustained virologic response in patients on pegylated interferon-a/ribavirin therapy (odds ratio 5 0.53, 95% confidence interval 0.31-0.91), which is in disagreement with the meta-analysis published by Kitson et al. Thus, these authors suggested some issues that should be clarified. Firstly, Kitson et al. pointed out that our analysis included three different studies with the same Italian cohort.2-4 However, these studies were published by prestigious journals during consecutive years and provided independent and remarkable information, each of them including a different number of patients. Nevertheless, the significant association was lost if we reanalyzed it considering only one of these studies. Secondly, with regard to the studies that were not included in our meta-analysis, we carefully selected articles from PubMed, SCOPUS, LILACS, and the Cochrane Library. When data were unclear, incomplete, or not explicitly reported, we contacted the corresponding author up to three times in order to obtain additional information. Unfortunately, some authors did not reply or refused to provide the requested data. Thirdly, we only included studies in which vitamin D was assessed as a categorical variable (the thresholds for deficiency, suboptimal, and optimal 25[OH]D levels) because we considered that it was more relevant and useful than the continuous variable.
1. Garcıa-Alvarez M, Pineda-Tenor D, Jimenez-Sousa MA, FernandezRodrıguez A, Guzman-Fulgencio M, Resino S. Relationship of vitamin D status with advanced liver fibrosis and response to hepatitis C virus therapy: a meta-analysis. HEPATOLOGY 2014;60:1541-1550. 2. Bitetto D, Fattovich G, Fabris C, Ceriani E, Falleti E, Fornasiere E, et al. Complementary role of vitamin D deficiency and the interleukin28B rs12979860 C/T polymorphism in predicting antiviral response in chronic hepatitis C. HEPATOLOGY 2011;53:1118-1126. 3. Falleti E, Bitetto D, Fabris C, Fattovich G, Cussigh A, Cmet S, et al. Vitamin D binding protein gene polymorphisms and baseline vitamin D levels as predictors of antiviral response in chronic hepatitis C. HEPATOLOGY 2012;56:1641-1650. 4. Bitetto D, Bortolotti N, Falleti E, Vescovo S, Fabris C, Fattovich G, et al. Vitamin A deficiency is associated with hepatitis C virus chronic infection and with unresponsiveness to interferon-based antiviral therapy. HEPATOLOGY 2013;57:925-933. 5. Bitetto D, Fabris C, Fornasiere E, Pipan C, Fumolo E, Cussigh A, et al. Vitamin D supplementation improves response to antiviral treatment for recurrent hepatitis C. Transpl Int 2011;24:43-50.
Author names in bold designate shared co-first authorship. C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27796 Potential conflict of interest: Nothing to report.
Is Pentoxifylline Still an Option in Severe Alcoholic Hepatitis? To the Editor: We read with interest the Hepatology Elsewhere piece by Caballeria1 describing the study by Park et al. that investigated the rela-
tive efficacies of prednisolone and pentoxyfylline in acute alcoholic hepatitis in a noninferiority study.2 Caballeria states that the trial concluded that the efficacy of pentoxifylline is lower than that of prednisolone. However, the noninferiority endpoint of that study
1644
CORRESPONDENCE
was inconclusive, and whereas the authors concluded that the two treatments were nonequivalent, insufficient methodological information was presented for this to be inferred. The design and interpretation of noninferiority studies is complex and requires detailed analysis to prevent erroneous conclusions being drawn. Arbitrary decisions regarding the limit of the noninferiority margin are needed in the design of such studies, and the rationale for the limits used should be clearly stated. Park et al. set the limit at D15%. The reasoning behind this was not presented, but three studies were cited. The most relevant of these was the meta-analysis from Mathurin et al.3 Here, the difference between prednisolone and the other treatment arm was approximately 20%. This was the best available evidence when the trial was designed. Guidance for the design of noninferiority studies is for a noninferiority margin of 50% of the treatment benefit with the active control, in this case prednisolone,4 suggesting that the noninferiority limit set was too lenient, thus risking a false-positive result (indicating that pentoxifylline is erroneously noninferior to prednisolone). In addition, regarding the equivalence endpoint, the assumptions of the relative efficacies of prednisolone and pentoxifylline were not given for the power calculation. For nonequivalence to be inferred, the assumed efficacy of both agents would need to be identical and this was not stated. Whereas it may be that pentoxyfilline is not equivalent to prednisolone in the treatment of severe alcoholic hepatitis, the design and/or reporting of the study from Park et al. prevents clear conclusions from being drawn. A clearer estimate of the relative efficacy will be provided by a large, blinded, randomized, controlled trial comparing steroids and pentoxifylline with a placebo control that has recently been completed.5
HEPATOLOGY, November 2015
DOI 10.1002/hep.27782 Potential conflict of interest: Dr. Elsharkawy advises Gilead.
Reply: I would like to thank Townsend et al. for their comments. I agree with them that it should be more appropriate, according to the results of Park et al.,1 to conclude that the efficacy of pentoxifylline was noninferior, instead of lower than those of corticosteroids. I also agree that the noninferiority studies are difficult and the cutoff to evaluate significant differences is, in some way, arbitrary. The 15% limit established by Park et al. was perhaps too conservative, although it seems reasonable according to the results of the meta-analysis of Mathurin et al.2 Pentoxifylline is still considered the second therapeutic option in treatment of severe alcoholic hepatitis. However, the efficacy of pentoxifylline in this clinical situation has not been confirmed in different scenarios, as summarized in a recent systematic review3 and in an editorial.4 The final results of the STOPAH trial were presented in the past meeting of the American Association for the Study of Liver Diseases.5 This study, designed to evaluate the therapeutic benefit of both prednisolone and pentoxifylline, was performed in 1,053 patients and was unable to demonstrate an effect in disease progression or survival. Thus, Thursz et al. conclude that pentoxifylline should no longer be used for treatment of severe alcoholic hepatitis. This study reinforces the conclusion of my comment that we are still far from the ideal treatment of severe alcoholic hepatitis, and that it is urgent to identify new targets for more specific and effective therapies.
SARAH A. TOWNSEND, M.R.C.P. IAN A. ROWE, M.R.C.P., PH.D. AHMED M. ELSHARKAWY, M.R.C.P., PH.D. Queen Elizabeth Hospital Liver and Hepatobiliary Unit Birmingham, United Kingdom
JUAN CABALLERIA, M.D., PH.D. Liver Unit Hospital Clınic IDIBAPS CIBERehd Barcelona, Spain
References
References
1. Caballeria J. Is pentoxifylline still an option in severe alcoholic hepatitis? HEPATOLOGY 2015;61:1425-1427. 2. Park SH, Kim DJ, Kim YS, Yim HJ, Tak WY, Lee HJ, et al. Pentoxiphylline vs. corticosteroids to treat severe alcoholic hepatitis: a randomised, non-inferiority open trial. J Hepatol 2014;61:792-798. 3. Mathurin P, Mendenhall CL, Carithers RL, Ramond MJ, Maddrey WC, Garstide P, et al. Corticosteroids improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol 2002;36:480-487. 4. Peggio C, Elbourne D, Altman D, Pocock S, Evans S; CONSORT Group. Reporting of non-inferiority and equivalence randomised trials: an extension of the CONSORT statement. JAMA 2006;295:1152-1160. 5. Forrest E, Mellor J, Stanton L, Bowers M, Ryder S, Austin A, et al. Steroids or Pentoxiphylline for Alcoholic Hepatitis (STOPAH): study protocol for a randomised controlled trial. Trials 2013;14:262.
1. Park SH, Kim DJ, Kim YS, Yim HJ, Tak WY, Lee HJ, et al. Pentoxifylline vs corticosterois to treat severe alcoholic hepatitis: a randomised, non-inferiority open trial. J Hepatol 2014;61:792-798. 2. Mathurin P, Mendenhall CL, Carithers RL, Ramond MJ, Maddrey WC, Garstide P, et al. Corticosteroid improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH. J Hepatol 2002;36:480-487. 3. Parker R, Amstrong MJ, Corbett C, Rowe IA, Houlihan DD. Systematic review: pentoxifylline for the treatment of severe alcoholic hepatitis. Aliment Pharmacol Ther 2013;37:845-854. 4. Louvet A. Prednisolone vs pentoxifylline for severe alcoholic hepatitis. J Hepatol 2014;61:723-724. 5. Thursz MR, Richardson P, Allison ME, Auastin A, Bowers M, Day CP, et al. Steroids or pentoxifylline for alcoholic hepatitis: results of the STOPAH trial. HEPATOLOGY 2014;60(Suppl):1267A.
Author names in bold designate shared co-first authorship. C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com.
C 2015 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.27781 Potential conflict of interest: Nothing to report.
