Prednisolone or pentoxifylline for alcoholic hepatitisPrednisolone or Pentoxifylline for Alcoholic Hepatitis Mark R. Thursz, M.D., Paul Richardson, M.D., Michael Allison, Ph.D., Andrew Austin, M.D., Megan Bowers, M.Sc., Christopher P. Day, M.D., Ph.D., Nichola Downs, P.G. Cert., Dermot Gleeson, M.D., Alastair MacGilchrist, M.D., Allister Grant, Ph.D., Steven Hood, M.D., Steven Masson, M.A., Anne McCune, M.D., Jane Mellor, M.Sc., John O'Grady, M.D., David Patch, M.D., Ian Ratcliffe, M.Sc., Paul Roderick, Ph.D., Louise Stanton, M.Sc., Nikhil Vergis, M.B., B.S., Mark Wright, Ph.D., Stephen Ryder, D.M., and Ewan H. Forrest, M.D., for the STOPAH Trial* BACKGROUND
Alcoholic hepatitis is a clinical syndrome characterized by jaundice and liver im- The authors' affiliations are listed in the
Appendix. Address reprint requests to pairment that occurs in patients with a history of heavy and prolonged alcohol use. Dr. Thursz at the Hepatology Section, The short-term mortality among patients with severe disease exceeds 30%. Predniso- Imperial College, Norfolk Pl., London lone and pentoxifylline are both recommended for the treatment of severe alcoholic W2 1NY, United Kingdom, or at m . [email protected] imperial . ac . uk.
hepatitis, but uncertainty about their benefit persists.
*A complete list of the investigators in the Steroids or Pentoxifylline for Alco- We conducted a multicenter, double-blind, randomized trial with a 2-by-2 factorial holic Hepatitis (STOPAH) trial is pro- vided in the Supplementary Appendix, design to evaluate the effect of treatment with prednisolone or pentoxifylline. The available at NEJM.org.
primary end point was mortality at 28 days. Secondary end points included death N Engl J Med 2015;372:1619-28.
or liver transplantation at 90 days and at 1 year. Patients with a clinical diagnosis of DOI: 10.1056/NEJMoa1412278
alcoholic hepatitis and severe disease were randomly assigned to one of four groups: Copyright 2015 Massachusetts Medical Society. a group that received a pentoxifylline-matched placebo and a prednisolone-matched placebo, a group that received prednisolone and a pentoxifylline-matched placebo, a group that received pentoxifylline and a prednisolone-matched placebo, or a group that received both prednisolone and pentoxifylline.
A total of 1103 patients underwent randomization, and data from 1053 were available
for the primary end-point analysis. Mortality at 28 days was 17% (45 of 269 patients) in the placebo–placebo group, 14% (38 of 266 patients) in the prednisolone–placebo group, 19% (50 of 258 patients) in the pentoxifylline–placebo group, and 13% (35 of 260 patients) in the prednisolone–pentoxifylline group. The odds ratio for 28-day mortality with pentoxifylline was 1.07 (95% confidence interval [CI], 0.77 to 1.49; P = 0.69), and that with prednisolone was 0.72 (95% CI, 0.52 to 1.01; P = 0.06). At 90 days and at 1 year, there were no significant between-group differences. Serious infections occurred in 13% of the patients treated with prednisolone versus 7% of those who did not receive prednisolone (P = 0.002).
Pentoxifylline did not improve survival in patients with alcoholic hepatitis. Prednisolone
was associated with a reduction in 28-day mortality that did not reach significance and with no improvement in outcomes at 90 days or 1 year. (Funded by the National Institute for Health Research Health Technology Assessment program; STOPAH EudraCT number, 2009-013897-42, and Current Controlled Trials number, ISRCTN88782125.) n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Alcoholic hepatitis is a distinct Among four randomized, controlled trials in manifestation of alcoholic liver disease which pentoxifylline was compared with placebo that is characterized by jaundice and liver for the treatment of alcoholic hepatitis, one showed failure. This condition develops in persons with a significant benefit.10 All 100 patients enrolled a history of prolonged and heavy alcohol use.1 had a discriminant function that was greater than A Quick Take The severity of alcoholic hepatitis is convention- 32. The mortality was 24.6% in the pentoxifylline
available at ally defined by Maddrey's discriminant function, group as compared with 46.1% in the placebo
NEJM.org which is calculated as 4.6 × (patient's prothrombin group (P = 0.04). The principal benefit related to
time in seconds − control's prothrombin time in pentoxifylline appeared to be a reduction in the seconds) + patient's serum bilirubin level in milli- number of deaths attributed to the hepatorenal grams per deciliter; a value of 32 or higher indi- syndrome. However, two meta-analyses have not cates severe alcoholic hepatitis that carries an shown any convincing benefit associated with adverse prognosis, with mortality of 20 to 30% pentoxifylline.11,12 within 1 month after presentation and 30 to Two small studies have compared glucocorti- 40% within 6 months after presentation.2 coids with pentoxifylline, but the results were in- A number of therapies have been evaluated for consistent.13,14 Two other studies have compared the treatment of alcoholic hepatitis, but only two the effect of glucocorticoid monotherapy with drugs have been incorporated into the treatment that of combined treatment with glucocorticoids guidelines published by the American Association and pentoxifylline but showed no benefit from for the Study of Liver Disease and the European the addition of pentoxifylline.15,16 The purpose of Association for the Study of the Liver 3,4. In a 2008 our trial — Steroids or Pentoxifylline for Alco- Cochrane meta-analysis of 15 randomized trials holic Hepatitis (STOPAH) — was to determine published since 1971 that compared glucocorti- whether prednisolone or pentoxifylline adminis- coids with placebo or no intervention, Rambaldi tered for a 28-day period reduced short-term and et al. investigated the role of glucocorticoid ther- medium-term mortality among patients admitted apy for this condition.5 Despite this apparent wealth to a hospital with severe alcoholic hepatitis.
of evidence, controversy persists. Advocates of the treatment cite significant reduction in short-term mortality, whereas detractors raise questions about the risks of sepsis and gastrointestinal hemor- Study Design and Oversight
rhage with glucocorticoid therapy. In the largest The design and rationale for the trial have been placebo-controlled study to date, investigators described previously.17 STOPAH was a multicenter, treated 90 patients with prednisolone and found randomized, double-blind trial with a 2-by-2 fac- no benefit of that therapy over placebo adminis- torial design that was intended to evaluate the tered in a similar group of patients.6 This study treatment effect of prednisolone and of pentoxi- was hampered by the inclusion of patients with fylline in the management of severe alcoholic moderate or severe alcoholic hepatitis and those hepatitis. A trial-management group designed the with alcohol-related cirrhosis. In the only study study (see the protocol, available with the full we found that required histologic confirmation text of this article at NEJM.org). The study was of alcoholic hepatitis in all patients, prednisolone approved by the Multicenter Research Ethics Com- was associated with a short-term reduction in mittee (reference number 09/MRE09/59), and clini- mortality, but this benefit was not apparent after cal trial authorization was received from the 2 years.7,8 The systematic review by the Cochrane Medicines and Healthcare Products Regulatory group revealed a trend toward a benefit with glu- Agency. Written informed consent was obtained cocorticoids that was not statistically significant.5 from each patient or from his or her legal repre- However a reanalysis of the five largest studies in- sentative until such time as the patient recovered dicated a significant benefit from glucocorticoids; mental capacity. The trial was conducted and re- in this meta-analysis, 28-day mortality among ported with fidelity to the protocol, the Medicines patients with discriminant function scores of 32 or for Human Use (Clinical Trials) Regulations 2004, higher was 20% among those who were treated as amended in 2006, the European Union Clinical with prednisolone, as compared with 34% among Trials Directive (Directive 2001/20/EC) guidelines, those who received placebo (P<0.001).9 the principles of the International Conference on n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Prednisolone or Pentoxifylline for Alcoholic Hepatitis Harmonisation Good Clinical Practice under the port were given to each patient. The clinician re- oversight of University Hospital Southampton NHS sponsible for each patient made the decision re- Foundation Trust, and the provisions of the Dec- garding other treatments, such as terlipressin for laration of Helsinki. An independent data moni- patients in whom hepatorenal failure was devel- toring and ethics committee, whose members were oping, acid suppression for prophylaxis against aware of the group assignments, was convened gastrointestinal hemorrhage, antibiotics, and vita- to review the conduct of the trial and to analyze min supplementation. Patients with renal failure primary end-point data, using prespecified stop- (defined as a creatinine level >500 μmol per liter ping guidelines, after the recruitment of 200, 400, [>5.7 mg per deciliter] or the requirement for re- and 800 patients, to avoid continued recruitment nal-replacement therapy), active gastrointestinal in the event that a definitive result had been bleeding, or untreated sepsis, and patients requir- ing inotropic support with epinephrine or nor- Data collected by site investigative teams were epinephrine, were excluded unless the condition submitted to the clinical trials unit and analyzed stabilized within the first 7 days after admission by study statisticians. The first author wrote the to the hospital.
first draft of the manuscript, with substantial contributions from the coauthors. All the authors Randomization
vouch for the accuracy and completeness of the A Web-based computer system (Tenalea, Forms- data and analyses.
Vision) was used to enroll eligible patients and randomly assign them to study groups. The ran- domization schedule was created with the use of The trial included patients with a clinical diagno- Stata software, version 11 (StataCorp). Random- sis of alcoholic hepatitis. Clinical diagnosis was ization was performed with a block size of four, chosen because the use of liver biopsy in this group with stratification according to geographic area of patients is uncommon, and the aim was to re- and risk category. The high-risk category con- cruit a large number of participants whose con- sisted of patients who had had an occurrence of dition would reflect as closely as possible the gastrointestinal bleeding, renal impairment, or condition of patients seen in clinical practice.18,19 sepsis before randomization. All other patients Patients were recruited from January 2011 through were assigned to the intermediate-risk category.
February 2014 at 65 hospitals across the United Patients were randomly assigned to one of four Kingdom. All patients admitted with suspected groups, with one group receiving a pentoxifylline- severe alcoholic hepatitis were evaluated for eli- matched placebo and a prednisolone-matched pla- gibility. A clinical diagnosis that was based on a cebo, the second group receiving 40 mg of history of recent excess alcohol consumption and prednisolone daily and a pentoxifylline-matched the absence of other causes of liver disease was placebo, the third group receiving 400 mg of pent- used for trial recruitment. Inclusion criteria were oxifylline three times daily and prednisolone- an age of 18 years or older, a clinical diagnosis of matched placebo, and the fourth group receiving alcoholic hepatitis, an average alcohol consump- 40 mg of prednisolone daily and 400 mg of pent- tion of more than 80 g per day for men and more oxifylline three times daily. All patients were pre- than 60 g per day for women, a serum bilirubin scribed treatment for 28 days.
level greater than 80 μmol per liter (4.7 mg per deciliter), and a discriminant function of 32 or End Points
higher. Key exclusion criteria were jaundice for The primary end point of the trial was mortality more than 3 months, cessation of alcohol con- at 28 days. Secondary end points included mortal- sumption for more than 2 months before ran- ity or liver transplantation at 90 days and at 1 year.
domization, the presence of other causes of liver disease, a serum aspartate aminotransferase level Statistical Analysis
greater than 500 IU per liter or serum alanine We estimated that a sample of 513 patients re- transaminase level greater than 300 IU per liter, ceiving each active agent and an equal number and previous entry into the study within the pre- not receiving each agent would be required to ceding 6 months.
detect a reduction in 28-day mortality from 30% Standard supportive care and nutritional sup- in the latter groups to 21% in the former groups. n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Thus, in total, our trial would require enrollment completed at least 28 days of follow-up. At the of 1026 patients. We allowed for a rate of with- time the trial was stopped, 33 patients who un- drawal or loss to follow-up of approximately 10% derwent randomization during the last 90 days of and therefore aimed to recruit 1200 patients to the trial could not be included in the 90-day or 12-month analyses. In addition, there were 159 The analysis was conducted on an intention- patients who underwent randomization within to-treat basis. The comparison of mortality at 28 90 days to 12 months before the end of trial who days between treated and untreated groups was could not be included in 12-month analyses. The performed with the use of logistic regression, four groups were well matched with regard to their with adjustments for risk category (high or inter- baseline characteristics, including laboratory val- mediate) and factorial design. Mortality and rates ues (Table 1). At 28 days, 16% of the patients had of liver transplantation at 90 days and 1 year were died, 1% had been lost to follow-up, and 2% had compared with the use of the same strategy. A withdrawn from the study. At 90 days, 29% of the test for treatment interaction was performed as patients (285 of 968 patients) had died, 5% had a secondary analysis. Cox proportional-hazards been lost to follow-up, 3% had withdrawn, and 4% regression was used to compare 1-year survival had not completed follow-up owing to cessation among the groups, and Kaplan–Meier curves for of the study. At 1 year, 56% of the patients (421 1-year survival were plotted. All P values were of 747 patients) had died or undergone liver trans- plantation (3 patients), 8% had been lost to follow- A univariate logistic-regression analysis was up, 4% had withdrawn, and 20% had not com- performed for the end points of 28-day mortality, pleted follow-up owing to cessation of the study.
90-day mortality, and 1-year mortality. Separate models were fitted for conventional prognostic End Points
scores (discriminant function, Model for End-Stage At 28 days, 45 of 269 patients (17%) in the pla- Liver Disease score, Glasgow alcoholic hepatitis cebo–placebo group had died, 38 of 266 patients score, and Lille score) and for clinical and labo- (14%) in the prednisolone–placebo group had died, ratory variables, to investigate whether they were 50 of 258 patients (19%) in the pentoxifylline– significant predictors of mortality. As part of a placebo group had died, and 35 of 260 patients prespecified analysis, individual and treatment (13%) in the prednisolone–pentoxifylline group variables that were found to be significant in uni- had died. There was no significant treatment in- variate analyses were used in multivariate logistic- teraction between prednisolone and pentoxifylline regression analysis, and backward elimination (P = 0.41). In the prespecified analysis of the pri- (which was not prespecified) was applied at a 5% mary outcome (a logistic-regression analysis that significance level.
was adjusted for the risk category [high or inter- mediate] used in the randomization and for the factorial design), the odds ratio for 28-day mor- tality among patients who received pentoxifylline (those in the pentoxifylline–placebo group or those Over a 3-year period, 5234 patients were screened, in the prednisolone–pentoxifylline group), as com- and after the application of eligibility criteria, 1103 pared with patients who did not receive pentoxi- patients were randomly assigned to one of the four fylline was 1.07 (95% confidence interval [CI], treatment groups: 276 to the placebo–placebo 0.77 to 1.49; P = 0.69), and the odds ratio among group, 277 to the prednisolone–placebo group, patients who received prednisolone (those in the 276 to the pentoxifylline–placebo group, and 274 prednisolone–placebo group or those in the pred- to the prednisolone–pentoxifylline group (see nisolone–pentoxifylline group), as compared with Fig. S1 in the Supplementary Appendix, available patients who did not receive prednisolone, was at NEJM.org). All the patients were followed for 0.72 (95% CI, 0.52 to 1.01; P = 0.06) (Table 2). 12 months or until the time of their death, with Neither prednisolone nor pentoxifylline was found the exception of patients enrolled at the end of to influence mortality or the need for liver trans- the trial. Owing to limitations on funding, the plantation at 90 days or 1 year (Table 2). Kaplan– trial was stopped after all enrolled patients had Meier curves for survival in each treatment group n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Prednisolone or Pentoxifylline for Alcoholic Hepatitis Table 1. Baseline Characteristics of the Patients.*
(N = 1092)
Male sex — no. (%) Alcohol consumption — g/day Time from admission to initiation of treatment — days Encephalopathy — no. (%)† Sepsis on admission — Renal failure on admission — Gastrointestinal bleeding on admission — no. (%) Laboratory results Bilirubin — mg/dl Albumin — g/liter Aspartate aminotransferase — Alkaline phosphatase — Creatinine — mg/dl White-cell count — per mm3 Neutrophils — per mm3 Prothrombin time — sec Prognostic scores Discriminant function§ * Plus–minus values are means ±SD. To convert the values for bilirubin to micromoles per liter, multiply by 17.1. To convert the values for creatinine to micromoles per liter, multiply by 88.4.
† In patients with encephalopathy, grade 1 indicates mild confusion and impaired attention; grade 2 lethargy, personality change, and inap- propriate behavior; grade 3 comatose behavior with responsiveness to verbal or noxious stimuli; and grade 4 coma without responsiveness to verbal or noxious stimuli.
‡ Renal failure was defined as a serum creatinine level greater than 5.7 mg per deciliter or the requirement for renal-support therapy.
§ Discriminant function is calculated as 4.6 × (patient's prothrombin time in seconds − matched control's prothrombin time in seconds) + pa- tient's serum bilirubin level in milligrams per deciliter; a value of 32 or higher indicates severe alcoholic hepatitis that carries an adverse prognosis, with mortality of 20 to 30% within 1 month after presentation and 30 to 40% within 6 months after presentation.
¶ In the Model for End-Stage Liver Disease (MELD), scores range from 6 to 40, with higher scores indicating worse prognosis.
‖ The Glasgow alcoholic hepatitis score (GAHS) ranges from 5 to 12, with higher scores indicating worse prognosis.
n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. and for survival with prednisolone versus no pred- nisolone and with pentoxifylline versus no pent- oxifylline are provided in Figure 1.
The baseline variables that influenced 28-day mortality in univariate analyses included age, en- cephalopathy, white-cell count, prothrombin ratio, and serum levels of bilirubin, creatinine, and urea (Table 3). In multivariate analyses, age, encepha- lopathy, white-cell count, prothrombin ratio, and serum levels of bilirubin, creatinine, and urea re- mained significant. In a secondary analysis, in which a multivariate logistic-regression model was used that adjusted for these prognostic variables, we found that the odds ratio for 28-day mortal- ity among the patients who received predniso- lone, as compared with those who did not, was 0.61 (95% CI, 0.41 to 0.91; P = 0.02) (Table 3). However, the effect of prednisolone on mortality at 90 days (odds ratio, 1.00; 95% CI, 0.73 to 1.36; P = 0.98) and at 1 year (odds ratio, 1.01; 95% CI, 0.74 to 1.39; P = 0.94) was nonsignificant.
Adverse Events Including Death
Serious adverse events were reported in 42% of the patients, with an equal distribution in each of the treatment groups, and 20% of all serious ad- verse events resulted in death. Infection occurred in 71 of the 547 patients (13%) who received pred- nisolone as compared with 38 of the 545 patients (7%) who did not receive prednisolone (P = 0.002). Acute kidney injury occurred in 9 of the 546 pa- tients (2%) who received pentoxifylline as com- pared with 14 of the 546 patients (3%) who did not receive pentoxifylline (Table 4, and Table S2 in the Supplementary Appendix).
There were 418 deaths during the trial; 168 (40%) of these occurred before day 29, 28% oc- curred between day 28 and day 90, and 32% occurred between day 91 and 1 year. Investiga- tors attributed 95% of the deaths to liver-related causes. Among these causes, infection account- ed for 24% of the deaths, with similar numbers reported for the groups receiving prednisolone and those not receiving prednisolone. The occur- rence of gastrointestinal bleeding, sepsis, or renal failure before randomization did not affect mor- tality during the trial.
mortality transplantation transplantation — Controversy over the use of glucocorticoids in se- * The interaction between interventions was investigated as a secondary analysis.
vere alcoholic hepatitis has persisted for many n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Prednisolone or Pentoxifylline for Alcoholic Hepatitis Figure 1. Kaplan–Meier Curves Showing Overall
A Prednisolone vs. No Prednisolone
Survival According to Study Group.
Kaplan–Meier survival curves show a nonsignificant survival advantage during the first 28 days among pa- tients who received prednisolone as compared with those who did not receive prednisolone (odds ratio, 0.72; 95% confidence interval [CI], 0.52 to 1.01; P = 0.06) (Panel A). No significant survival advantage was seen for patients who received pentoxifylline as compared with those who did not receive pentoxifyl- line (odds ratio, 1.07; 95% CI, 0.77 to 1.49; P = 0.69) (Panel B). Survival curves for all four study groups up to 1 year are also shown (Panel C).
Days since Start of Treatment
No. at Risk
years despite the results of meta-analyses of se- lected trials.5,9 In our study, the reduction in 28-day mortality observed among patients treated with B Pentoxifylline vs. No Pentoxifylline
prednisolone did not reach the conventional No pentoxifylline threshold of statistical significance, and no sig- nificant differences were observed in 90-day or 12-month outcomes. However, in a secondary analysis that included adjustments for baseline determinants of prognosis, a significant advantage with respect to 28-day mortality was seen with prednisolone. The survival differences may have been a chance finding or may represent a benefit of prednisolone for short-term mortality that did not translate to longer-term benefit.
Although we used the same threshold of dis- Days since Start of Treatment
ease severity that has been used in most other tri- No. at Risk
als of alcoholic hepatitis (a score of 32 or higher No pentoxifylline for discriminant function), the 28-day mortality overall in STOPAH was appreciably lower than C One-Year Survival in All Groups
the 28-day mortality in the trials included in the meta-analysis by Mathurin et al.9 However, many of the studies included in the meta-analysis were performed more than 30 years ago, and the mor- tality reported in two more recently published tri- als was similar to that reported here.13,15,20 There was a lower incidence of infection and acute kidney injury in STOPAH than in some previous trials, a factor that may have contributed to the lower mortality.21 A comparison of the baseline char- acteristics of the patients in our study with those of patients in other trials conducted in the past Days since Start of Treatment
4 years shows that in our trial, the mean age was No. at Risk
slightly younger and the rate of encephalopathy lower; both characteristics have been consistently shown to influence mortality.15,20 However, the bilirubin, creatinine, and albumin levels observed n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Table 3. Analysis of Factors Associated with Mortality at 28 Days.*
Prednisolone vs. no prednisolone 0.61 (0.41–0.91) Pentoxifylline vs. no pentoxifylline 1.10 (0.74–1.64) Discriminant function 1.02 (1.02–1.03) 2.17 (1.86–2.53) 1.15 (1.12–1.18) 1.03 (1.02–1.03) Prothrombin ratio or INR 1.38 (1.10–1.73) 1.38 (1.13–1.69) 1.05 (1.03–1.07) 1.05 (1.03–1.07) 1.06 (1.03–1.08) 1.03 (1.00–1.06) 1.14 (1.10–1.18) 1.06 (1.01–1.12) 3.07 (2.32–4.08) 1.56 (1.05–2.33) 0.66 (0.37–1.16) 1.20 (0.79–1.83) 1.09 (0.71–1.65) 1.00 (1.00 –1.00) 0.99 (0.96–1.02) Alkaline phosphate 1.00 (1.00–1.00) 1.07 (1.05–1.09) 1.03 (1.01–1.06) Hepatic encephalopathy 3.70 (2.59–5.29) 3.07 (2.05–4.60) * Values reported for all variables were obtained at baseline, and the value for bilirubin was also obtained on day 7 to cal- culate the Lille score. GAHS denotes Glasgow alcoholic hepatitis score, which ranges from 5 to 12, with higher scores indicating worse prognosis, INR international normalized ratio, and MELD the Model for End-Stage Liver Disease, in which scores range from 6 to 40, with higher scores indicating worse prognosis.
† A separate univariate model was fitted for all baseline prognostic factors.
‡ Significant variables in the univariate analysis were added to a multivariate model, and backward elimination was ap- plied at the 5% significance level. Prognostic scores (calculated on the basis of discriminant function, MELD, GAHS and Lille score) were excluded from the multivariate analysis owing to multicolinearity with individual component vari- § The Lille score defines the risk of death according to the number of adverse prognostic factors, with higher scores indi- cating a worse prognosis. We calculated Lille scores on the basis of patient age, albumin level, bilirubin levels (at base- line and on day 7), the prothrombin ratio, and prothrombin time; the formula we used is shown in the protocol.
in patients in STOPAH were similar to those seen accuracy.22 Nevertheless, incorrect diagnostic la- in patients in other studies.
beling could have reduced the power of the study The use of liver biopsy to provide histologic to detect a therapeutic effect.
confirmation of alcoholic steatohepatitis remains A recognized drawback of glucocorticoid use controversial.4,18,19 Although many investigators in patients with alcoholic hepatitis is increased have used histologic confirmation as an entry susceptibility to infection.21 The higher rate of in- criterion, it is rarely applied in clinical practice fection among the patients treated with predniso- apart from those cases in which the diagnosis is lone was therefore expected, but mortality attrib- uncertain.19 Although it is possible that the diag- uted to infection was similar across the groups, nosis of alcoholic hepatitis was incorrect in a regardless of whether prednisolone was admin- small number of the patients in our study, when istered. Investigators attributed deaths to infection strict clinical criteria are used, it is possible to in 103 of 416 cases (24.8%), but infection also diagnose the condition with a high degree of probably played a role in deaths attributed to other n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Prednisolone or Pentoxifylline for Alcoholic Hepatitis Table 4. Serious Adverse Events.*
number of patients (percent) Serious adverse event
Any that led to death Serious adverse event occurring in ≥5% of patients
Upper gastrointestinal hemorrhage Hepatobiliary disorder Infection or infestation† Lung infection‡ Renal or urinary disorder * Adverse events were categorized according to the Common Toxicity Criteria for Adverse Events.
† P = 0.002 for the comparison of serious adverse events of infection or infestation, which occurred in 71 of 547 (13%) of the patients who re- ceived prednisolone versus 38 of 545 (7%) of the patients who did not receive prednisolone.
‡ P = 0.007 for the comparison of lung infections, which occurred in 37 of 547 (7%) of the patients who received prednisolone versus 17 of 545 (3%) of the patients who did not receive prednisolone.
causes, such as multiorgan failure. Since infec- In summary, in the STOPAH trial, pentoxifyl- tion plays such an important role in the outcome line did not improve outcomes in patients with of alcoholic hepatitis, it is worth noting that in a alcoholic hepatitis. The findings suggest that the trial published in 2011, the addition of N-acetyl- administration of 40 mg of prednisolone daily for cysteine to prednisolone was associated with a 1 month may have a beneficial effect on short- reduced rate of infection.20 term mortality but not on the medium-term or The results of this trial showed that after 28 long-term outcome of alcoholic hepatitis.
days, neither prednisolone nor pentoxifylline in- fluenced mortality. Furthermore, the cumulative Supported by a grant (08 14 44) from the National Institute mortality at 90 days and at 1 year in this group for Health Research (NIHR) Health Technology Assessment of patients is alarming. Self-reported alcohol con- program.
Dr. Thursz reports receiving lecture fees and consulting fees sumption revealed complete abstinence in 37% of from Gilead, Bristol-Myers Squibb, AbbVie, and Abbott; and Dr. the patients at 1 year of follow-up. However, data Allison, receiving consulting fees from Norgine. No other po- on alcohol consumption are difficult to collect, tential conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with and this fact is reflected in the high proportion the full text of this article at NEJM.org.
of missing data. No matter what the exact figures We thank the NIHR Clinical Research Network for providing are, more clearly needs to be done to prevent re- support for research nurses and the Imperial College Biomedical Research Centre, Southampton Clinical Trials Unit, for manage- cidivism in this group of patients.
ment of the trial.
From Imperial College (M.R.T., N.V.), King's College Hospital (J.O.), and the Royal Free Hospital (D.P.), London, Royal Liverpool Hospital (P. Richardson) and Aintree Hospital (S.H.), Liverpool, Addenbrookes Hospital, Cambridge (M.A.), Derby Royal Hospital, Derby (A.A.), Southampton Clinical Trials Unit, University of Southampton (M.B., N.D., J.M., I.R., P. Roderick, L.S.), and University Hospital Southampton NHS Foundation Trust (M.W.), Southampton, Faculty of Medical Sciences, Newcastle University (C.P.D.), and n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Prednisolone or Pentoxifylline for Alcoholic Hepatitis Newcastle upon Tyne Hospitals NHS Foundation Trust (S.M.), Newcastle upon Tyne, Sheffield Teaching Hospitals Foundation Trust, Sheffield (D.G.), Edinburgh Royal Infirmary, Edinburgh (A. MacGilchrist), Leicester Royal Infirmary, Leicester (A.G.), Bristol Royal Infirmary, Bristol (A. McCune), Nottingham University Hospitals NHS Trust and National Institute for Health Research Biomedical Research Unit, Queens Medical Centre, Nottingham (S.R.), and the Glasgow Royal Infirmary, Glasgow (E.H.F.) — all in the United References
1. Lucey MR, Mathurin P, Morgan TR. treated with prednisolone. Gastroenterol-
alcoholic hepatitis: a randomized clinical Alcoholic hepatitis. N Engl J Med 2009; ogy 1996; 110: 1847-53.
trial. JAMA 2013; 310: 1033-41.
9. Mathurin P, O'Grady J, Carithers RL, 16. Sidhu SS, Goyal O, Singla P, et al.
2. Maddrey WC, Boitnott JK, Bedine MS, et al. Corticosteroids improve short-term Corticosteroid plus pentoxifylline is not
Weber FL Jr, Mezey E, White RI Jr. Corti- survival in patients with severe alcoholic better than corticosteroid alone for im- costeroid therapy of alcoholic hepatitis. hepatitis: meta-analysis of individual pa- proving survival in severe alcoholic hep- Gastroenterology 1978; 75: 193-9.
tient data. Gut 2011; 60: 255-60.
atitis (COPE trial). Dig Dis Sci 2012; 57: 3. European Association for the Study of 10. Akriviadis E, Botla R, Briggs W, Han 1664-71.
Liver. EASL clinical practical guidelines: S, Reynolds T, Shakil O. Pentoxifylline 17. Forrest E, Mellor J, Stanton L, et al.
management of alcoholic liver disease. improves short-term survival in severe Steroids or Pentoxifylline for Alcoholic J Hepatol 2012; 57: 399-420.
acute alcoholic hepatitis: a double-blind, Hepatitis (STOPAH): study protocol for a 4. O'Shea RS, Dasarathy S, McCullough placebo-controlled trial. Gastroenterolo-
randomised controlled trial. Trials 2013; AJ. Alcoholic liver disease. Hepatology gy 2000; 119: 1637-48.
2010; 51: 307-28.
11. Whitfield K, Rambaldi A, Wetterslev 18. Potts JR, Goubet S, Heneghan MA,
5. Rambaldi A, Saconato HH, Chris-
J, Gluud C. Pentoxifylline for alcoholic Verma S. Determinants of long-term out- tensen E, Thorlund K, Wetterslev J, hepatitis. Cochrane Database Syst Rev come in severe alcoholic hepatitis. Ali- Gluud C. Systematic review: glucocorti- 2009; 4: CD007339.
ment Pharmacol Ther 2013; 38: 584-95.
costeroids for alcoholic hepatitis — a 12. Parker R, Armstrong MJ, Corbett C, 19. Singal AK, Salameh H, Singal A, et al.
Cochrane Hepato-Biliary Group system- Rowe IA, Houlihan DD. Systematic re- Management practices of hepatitis C virus atic review with meta-analyses and trial view: pentoxifylline for the treatment of infected alcoholic hepatitis patients: sequential analyses of randomized clini- severe alcoholic hepatitis. Aliment Phar- A survey of physicians. World J Gastroin- cal trials. Aliment Pharmacol Ther 2008; macol Ther 2013; 37: 845-54.
test Pharmacol Ther 2013; 4: 16-22.
13. Park SH, Kim DJ, Kim YS, et al. Pent-
20. Nguyen-Khac E, Thevenot T, Piquet
6. Mendenhall CL, Anderson S, Garcia-
oxifylline vs. corticosteroid to treat severe MA, et al. Glucocorticoids plus N-acetyl- Pont P, et al. Short-term and long-term alcoholic hepatitis: a randomised, non- cysteine in severe alcoholic hepatitis. survival in patients with alcoholic hepati- inferiority, open trial. J Hepatol 2014; 61: N Engl J Med 2011; 365: 1781-9.
tis treated with oxandrolone and pred- 792-8.
21. Louvet A, Wartel F, Castel H, et al. In-
nisolone. N Engl J Med 1984; 311: 1464-70. 14. De BK, Gangopadhyay S, Dutta D, fection in patients with severe alcoholic
7. Ramond MJ, Poynard T, Rueff B, et al. Baksi SD, Pani A, Ghosh P. Pentoxifylline hepatitis treated with steroids: early re-
A randomized trial of prednisolone in pa- versus prednisolone for severe alcoholic sponse to therapy is the key factor. Gas- tients with severe alcoholic hepatitis. hepatitis: a randomized controlled trial. troenterology 2009; 137: 541-8.
N Engl J Med 1992; 326: 507-12 World J Gastroenterol 2009; 15: 1613-9.
22. Forrest EH, Gleeson D. Is a liver bi-
8. Mathurin P, Duchatelle V, Ramond 15. Mathurin P, Louvet A, Duhamel A, et opsy necessary in alcoholic hepatitis?
MJ, et al. Survival and prognostic factors al. Prednisolone with vs without pentoxi- J Hepatol 2012; 56: 1427-8.
in patients with severe alcoholic hepatitis fylline and survival of patients with severe Copyright 2015 Massachusetts Medical Society. n engl j med 372;17 nejm.org April 23, 2015 The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF NEW MEXICO on April 23, 2015. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved.
PRODUCT MONOGRAPH Modified-Release Capsule, 40 mg (as doxycycline monohydrate) GALDERMA CANADA INC. Date of Preparation: 105 Commerce Valley Dr. W., Suite 300 November 09, 2011 Thornhill, Ontario L3T 7W3 Date of Revision: Submission Control No: 156056 August 30, 2012 Page 1 of 27 Table of Contents August 30, 2012 Page 2 of 27
Helicobacter pylori eradication with a capsule containing bismuth subcitrate potassium, metronidazole, and tetracycline given with omeprazole versus clarithromycin-based triple therapy: a randomised, open-label, non-inferiority, phase 3 trial
Helicobacter pylori eradication with a capsule containing bismuth subcitrate potassium, metronidazole, and tetracycline given with omeprazole versus clarithromycin-based triple therapy: a randomised, open-label, non-inferiority, phase 3 trial Peter Malfertheiner, Franco Bazzoli, Jean-Charles Delchier, Krysztof Celiñski, Monique Giguère, Marc Rivière, Francis Mégraud, for the Pylera Study Group