The Gastro & Hep Report: Fall 2012

Gastroenterology & Hepatology  Fall 2012


Comprehensive Reports on the Latest Advances in Gastroenterology and Hepatology From:

• The 47th Annual Meeting of the European Association for the Study of the Liver
April 18–22, 2012 • Barcelona, Spain

• Digestive Disease Week 2012
May 19–22, 2012 • San Diego, California


Presentations in IBS

Oral Methylnaltrexone Relieves Opioid-Induced Constipation in Patients with Noncancer Pain

Opioid-induced constipation can influence patients’ quality of life and pain control. Currently, a subcutaneous, injectable form of methylnaltrexone bromide (Relistor, Salix Pharmaceuticals) is available to treat opioid-induced constipation in patients with advanced illness. To evaluate whether an oral formulation of this drug is also effective for treatment of opioid-induced constipation in patients with chronic noncancer pain, Richard L. Rauck and associates recently conducted a phase III, double-blind, placebo-controlled trial. The results of this trial were presented at the 2012 Digestive Disease Week (DDW) meeting, held May 19–22, 2012 in San Diego, California.

This study enrolled 804 patients who were taking at least 50 mg oral morphine equivalents per day for at least 1 month and had a history of opioid-induced constipation. Patients were randomized to receive oral methylnaltrexone at 1 of 3 doses (150 mg, 300 mg, or 450 mg) or placebo once daily for 4 weeks and then as needed for another 8 weeks.

Over the 12 weeks of the study, the proportion of patients who experienced rescue-free bowel movements (RFBMs) within 4 hours of dosing was significantly higher among patients treated with 300-mg and 450-mg doses of methylnaltrexone compared to placebo. There was also a significant difference in all treatment groups compared to placebo in terms of the proportion of patients who attained a RFBM within 24 hours after the first dose (150 mg, 34%; 300 mg, 41%; 450 mg, 42%; placebo, 23%). Among patients treated with 300-mg and 450-mg doses of methyl-naltrexone, a significantly greater proportion of patients achieved at least 3 RFBMs per week with an increase of at least 1 RFBM per week over baseline, compared to placebo (300 mg, 47.8%; 450 mg, 50.5%; placebo, 36.8%). Finally, there was a linear dose response for all 3 methyl-naltrexone treatment groups over the first 4 weeks of dosing (P<.0001). The investigators noted that the incidence of adverse events was similar among treatment groups and placebo; abdominal pain, nausea, flatulence, and diarrhea were the most common adverse events reported.

Linaclotide Significantly Improved Abdominal Pain in Patients with IBS Regardless of Baseline Pain Severity

Previously, 2 phase III trials showed that linaclotide (Linzess, Ironwood Pharmaceuticals/Forest Pharmaceuticals) can effectively improve symptoms in patients with irritable bowel syndrome (IBS) with constipation. In these trials, patients were randomized to receive oral linaclotide (290 µg once daily) or placebo. During the 2-week baseline period and the 12-week treatment period, patients used an interactive, voice-response system to rate their highest level of abdominal pain in each 24-hour period; pain was rated on a 10-point scale from 0 (none) to 10 (very severe). Once treatment began, patients rated their relief of abdominal pain over the past week compared to baseline using a balanced 7-point scale.

During the 2012 DDW meeting, Philip S. Schoenfeld and colleagues presented the results of a post–hoc analysis in which they pooled data from these phase III studies and assessed the efficacy of linaclotide when patients were stratified according to their mean baseline abdominal pain score (<5, ≥5 to <7, or ≥7). This analysis found a significant improvement in 12-week abdominal pain scores with linaclotide compared to placebo (P<.0001). Further, improvements in abdominal pain with linaclotide occurred regardless of baseline abdominal pain severity (decreases in pain scores of 29–36% with linaclotide vs 18–20% with placebo; P<.0001). There was a correlation between baseline abdominal pain scores and the absolute improvement in abdominal pain (magnitude of change from baseline; r=0.26; P<.0001), but baseline abdominal pain scores did not correlate with the percent improvement from baseline (r=0.00; P=.92). For patient-reported relief of abdominal pain, linaclotide achieved higher ratings than placebo (2.9 vs 3.5 overall; P<.0001); in addition, significant improvements occurred with linaclotide across the 3 baseline pain level subgroups. Overall, the authors of the study concluded that linaclotide resulted in significant improvement in abdominal pain.

26 Weeks of Linaclotide Treatment Is Effective for Patients with IBS and Constipation

In another analysis of linaclotide, a phase III trial investigated this drug for treatment of IBS and constipation over an extended treatment period. In this randomized, double-blind trial presented at the 2012 DDW meeting, William D. Chey and associates assessed adequate relief and symptom severity over 26 weeks of linaclotide treatment among patients with IBS and constipation.

Patients with IBS and constipation were randomized to receive oral linaclotide (290 µg once daily) or placebo for 26 weeks. On a daily basis, patients rated abdominal pain at its worst during the previous 24 hours (0=none to 10=very severe), and they provided data on spontaneous bowel movement frequency. On a weekly basis, patients reported adequate relief of IBS symptoms (yes/no) and IBS severity (1=none to 5=very severe). For this analysis, adequate relief responders were defined as patients with adequate relief of IBS symptoms for at least 13 of the
26 weeks of treatment.

The proportion of patients who were adequate relief responders was significantly higher in the linaclotide arm than in the placebo arm (49.1% vs 25.1%; P<.0001). Adequate relief responders had a mean decrease in abdominal pain of 52% compared to 18% for nonresponders (P<.0001), and they had an increased spontaneous bowel movement frequency of 3.0 per week compared to 0.8 per week for nonresponders (P<.0001). Both improvements in abdominal pain and spontaneous bowel movement frequency correlated with adequate relief (r=0.48 and r=0.53, respectively). The investigators concluded that patients with IBS and constipation who were treated with 26 weeks of linaclotide were more likely than placebo-treated patients to achieve adequate relief of symptoms and improvement in IBS severity.

Patients Respond Well to Lubiprostone for the Treatment of Moderate-to-Severe IBS with Constipation 

Lubiprostone (Amitiza, Sucampo Pharmaceuticals) is a selective CIC-2 chloride channel activator that is approved for the treatment of IBS with constipation in women. At the 2012 DDW meeting, Taryn R. Joswick and colleagues presented the results of a post–hoc analysis of 2 pivotal, phase III trials in which patients were randomized 2:1 to receive lubiprostone (8 µg twice daily) or placebo for 12 weeks.

In this analysis, the authors investigated the effect of lubiprostone for the treatment of moderate-to-severe IBS with constipation; specifically, this study included patients with at least moderate abdominal pain and fewer than 3 spontaneous bowel movements per week at baseline. In analysis 1, patients were classified as responders if they achieved a 30%-or-greater improvement in mean abdominal pain scores from baseline, at least 1 more spontaneous bowel movement per week than at baseline, and at least 3 spontaneous bowel movements per week for either 6 of 12 treatment weeks or 9 of 12 treatment weeks. In analysis 2, patients with severe or very severe abdominal pain at baseline were assessed for weekly improvements in abdominal pain.

Of the 318 patients in analysis 1, treatment with lubiprostone resulted in significantly greater proportions of 6-of-12 week responders compared to placebo (24.1% vs 9.2%; P=.0031); the proportion of 9-of-12
week responders was also significantly higher among lubiprostone-treated patients compared to placebo (12.6% vs 3.4%; P=.0109). In analysis 2, patients with severe or very severe abdominal pain at baseline (n=277) showed significant improvements in abdominal pain on a weekly basis with lubiprostone compared to placebo (P=.0002). In addition, 35.1% of patients with severe or very severe pain at baseline had overall improvements of 30% or greater with lubiprostone.

Overall, lubiprostone was well tolerated; nausea, headache, and diarrhea were the most commonly reported adverse events. The researchers concluded that patients with severe symptoms at baseline had significant improvement in symptoms when treated twice daily with 8 µg of lubiprostone.

Alterations of Intestinal Microbiota After Probiotic Treatment Are Associated with Symptomatic Improvement in IBS

In a double-blind, placebo-controlled study presented at the 2012 DDW meeting, Kang Nyeong Lee and associates investigated whether changes in intestinal microbiota following treatment with probiotics correlated with changes in IBS symptoms. Patients diagnosed with IBS using the Rome III criteria were randomized to receive a mixture of probiotics (n=25 patients)—including Bifidobacterium longum, Bifido-bacterium breve, Bifidobacterium lactis, Lactobacillus acidophilus, Lactobacillus rhamnosus, and Streptococcus thermophilus—or placebo (n=24) twice daily for 4 weeks.

Patients in both the probiotic treatment group and the placebo group showed similar improvements in diarrhea and constipation. However, patients treated with probiotics had significantly greater improvements in abdominal pain severity, frequency of pain, and frequency of defecation compared to patients treated with placebo (P<.05). In addition, the probiotic treatment group had greater improvements at 4 weeks in abdominal pain, satisfaction with bowel habits, and daily life activities compared to the placebo group (P<.05).

In addition to assessing patient symptoms, a quantitative real-time polymerase chain reaction assay was used to analyze DNA in feces both before and after treatment in a subset of patients (18 patients in the probiotic treatment group and 15 patients in the placebo group). There was no difference in fecal microbiota at baseline. Four weeks of treatment with probiotics resulted in significant increases in total numbers of B. lactis, L. rhamnosus, and S. thermophilus; this change was not seen in patients who received placebo. No changes were seen in the numbers of B. longum, Bifidobacterium bifidum, L. acidophilus, Clostridium perfringens, or Escherichia coli in either the probiotic group or the placebo group after 4 weeks of treatment. Finally, improvement rates for frequency of abdominal pain were significantly greater among those patients who showed a 10-fold increase in B. lactis compared to patients with no change in B. lactis numbers (91.9% vs 54.5%; P=.03).


Presentations in GERD

Proton Pump Inhibitor Treatment May Benefit Some Patients with Suspected Reflux–Chronic Cough 

At the 2012 DDW meeting, Peter J. Kahrilas and colleagues presented the results of a study that sought to determine whether chronic cough responds to proton pump inhibitor (PPI) therapy in patients with evidence of gastroesophageal reflux disease (GERD). To address this question, the authors used systematic searches in PubMed and Embase from 1966 to August 2011 to identify clinical trials reporting PPI therapy and cough response in patients with GERD or laryngopharyngeal reflux (LPR). The search identified 52 potentially relevant clinical trials published in English. Of these trials, 10 were randomized, placebo-controlled studies that included an assessment of whether cough responds to either PPI or histamine receptor antagonist (H2RA) therapy. Six studies included data on PPI efficacy in patients with confirmed GERD or confirmed LPR, 1 study included data on H2RA efficacy in patients with confirmed GERD, and 4 studies included data from patients with unconfirmed GERD or unconfirmed LPR.

Five of the 7 studies with sufficient data on patients with confirmed GERD showed a positive therapeutic gain for treatment with PPIs or H2RAs compared to placebo (range, 12.5–35.8%). Among all 7 studies of patients with confirmed GERD, 2 showed significantly improved cough scores with therapy, 1 study revealed a nonsignificant trend toward improved cough scores, 1 study showed no significant difference in the prevalence of cough response, and 3 studies found no significant improvement in cough scores.

Two studies had sufficient data to calculate therapeutic gain in patients with unconfirmed GERD or LPR (range, 0.0–8.6%); only 1 of these studies showed a positive therapeutic gain for PPI therapy versus placebo. Only 1 study had separate data for patients with unconfirmed GERD, and it showed no significant improvement in cough scores with treatment. Kahrilas and coauthors concluded that they could not rule out the possibility that PPI or H2RA acid suppression therapy might yield a therapeutic benefit for chronic cough.

Weight Loss Reduces GERD Symptoms

High body mass index (BMI) is known to be associated with GERD symptoms, but whether weight loss can provide symptom relief remains unclear. At the 2012 DDW meeting, Elvind Ness-Jensen and associates presented the results of a study in which they assessed the effect of weight loss on GERD symptoms. This analysis used data from the HUNT Study, a population-based, cohort study conducted in the Norwegian county of Nord-Trøndelag. A total of 29,610 patients were prospectively followed from baseline (HUNT 2, 1995–1997) to follow-up (HUNT 3, 2006–2009). Data on GERD symptoms were collected by questionnaires. A logistic regression analysis was then performed, and the results were stratified according to antireflux medication. The results were adjusted for sex, age, smoking status, alcohol consumption, education, and frequency of physical exercise.

In HUNT 2, a total of 9,299 persons reported GERD symptoms; 2,398 (26%) of these individuals reported no GERD symptoms in HUNT 3. Among study participants who had a BMI reduction of at least 3.5 kg/m2,
the odds ratio (OR) for loss of GERD symptoms was 1.98 (95% confidence interval [CI], 1.45–2.72; P<.001) if the patients were taking antireflux medications less than once weekly and 3.95 (95% CI, 2.03–7.65; P<.001) if they were taking antireflux medication at least weekly.

Severe GERD symptoms were reported in 1,553 persons in HUNT 2, with 284 individuals (18%) reporting no GERD symptoms in HUNT 3. For those persons who had a BMI reduction of at least 3.5 kg/m2, the OR for loss of GERD symptoms was 0.90 (95% CI, 0.32–2.55; P=.189) for patients who were taking antireflux medications less than once weekly and 3.11 (95% CI, 1.13–8.58; P=.047) if they were taking antireflux medication at least weekly. The investigators concluded that weight loss was associated with loss of GERD symptoms, and reductions in BMI appeared to increase the efficacy of antireflux medications.

Underreporting of Symptoms Has Consequences for Interpretation of Ambulatory Reflux Monitoring

Ambulatory pH and impedance monitoring can be used to determine an association between symptoms and reflux events, but timely documentation of symptoms by patients is essential. At the 2012 DDW meeting, Michael F. Vaezi and associates presented the results of a study that determined the temporal accuracy of patient-reported acid reflux events. The investigators employed a novel, ambulatory, acoustic monitoring system that detects coughs through tracheal and chest wall sounds. This system was used in conjunction with ambulatory pH/impedance monitoring that records acid reflux events and nonacid reflux events. Patient-recorded symptoms were compared to the time of the cough event as detected by the acoustic recording device. The events were considered to be concordant if they occurred within 1, 2, or 5 minutes of each other. The study enrolled patients with chronic cough who had not received PPI therapy in the previous 10 days.

The audio device detected significantly more cough events than the number of cough events reported by patients (median of 216 events [range, 90–275] vs median of 34 events [range, 22–60], respectively; P<.001). According to the 1-minute, 2-minute, and 5-minute concordance windows, patients failed to report 91%, 82%, and 71% of audible coughs, respectively. The investigators noted that there was a higher degree of concordance among the 6 audio-recording listeners than between the audio recording and the patient-reported symptoms (P<.001). The pattern of cough frequency detected by the audio-recording listeners was similar to the pattern of coughs reported by the patient, but the listeners reported more cough events. The authors of the study concluded that most cough events were not reported by patients; thus, they cautioned against making clinical decisions based solely on ambulatory reflux monitoring.

Utility of Biopsies for Eosinophilic Esophagitis in Patients with Dysphagia

If patients with dysphagia undergo an endoscopy evaluation that does not reveal an obvious cause for their complaint, then tissue biopsies may be taken and evaluated for eosinophilic esophagitis (EoE). To determine whether this approach is beneficial, Vu Le and colleagues assessed the diagnostic yield of EoE in patients with normal and abnormal endoscopy findings. The results of this study were presented at the 2012 DDW meeting.

The investigators retrospectively identified patients with dysphagia who had endoscopic biopsies taken for pathologic evaluation at the University of Oklahoma Health Sciences Center. The analysis included 547 upper endoscopies in 527 patients (age, 18–90 years). Among the 547 endoscopies, 52% (n=284) were normal, 16% (n=85) showed findings suggestive of EoE (furrowing or longitudinal rings), 13% (n=73) showed signs of esophagitis, 8% (n=45) revealed lower esophageal strictures or Schatzki rings, and 11% (n=60) showed other findings, such as erythema or erosions. The pathology reports found that 55% of biopsies (n=303) showed signs of esophagitis, 38% (n=207) showed normal mucosa, and 7% (n=37) showed signs of EoE. Of the 37 patients with EoE that was identified by pathology, only 23 cases were identified as EoE via endoscopy. The remainder of the pathology-diagnosed EoE cases had normal endoscopies (n=7) or other findings (n=7).

The study authors concluded that biopsies for pathologic evaluation of EoE often correlate with observed endoscopic features that are suggestive of EoE—such as furrowing or longitudinal rings. However, some patients had EoE identified by pathology despite normal endoscopic findings. More studies are needed to investigate the diagnostic yield for EoE in patients with normal endoscopic findings.

Smoking Is an Independent Risk Factor for Barrett Esophagus

Smoking has been associated with esophageal adeno-carcinoma, but data are lacking to answer the question of whether smoking is a risk factor for Barrett esophagus. To explore this possibility, Gokulakrishnan Balasubramanian and associates conducted a prospective study that assessed the association between smoking and the risk of Barrett esophagus in patients with GERD. Results of this study were presented at the 2012 DDW meeting.

A total of 1,111 consecutive patients undergoing endoscopy for evaluation of GERD were asked to complete a validated GERD questionnaire. The mean age of the patients was 57.5 years, 82.8% of patients were white, 92.8% were male, and patients’ mean BMI was 29.5 kg/m2. The study included 378 (34%) current smokers and 733 (66%) nonsmokers. A univariate analysis found that the smokers were younger (52.3 years vs 60.3 years; P<.0001), had a lower BMI (28.8 kg/m2 vs 29.9 kg/m2; P=.001), and were less likely to be taking aspirin (36.6% vs 44.1%; P=.018). No significant differences in race, gender, family history, GERD symptoms, or treatments (PPIs, H2 blockers, or statins) were observed between smokers and nonsmokers.

Barrett esophagus was diagnosed in 153 patients (13.8%). Among smokers, the calculated adjusted OR for the presence of confirmed Barrett esophagus (intestinal metaplasia on histology) was 1.56 (95% CI, 1.02–2.38). Thus, smokers had a 50–60% higher risk for Barrett esophagus than nonsmokers. The investigators concluded that smoking is an independent risk factor for Barrett esophagus.

Risk Factors for Helicobacter pylori Infection in Latin America

At the 2012 DDW meeting, Carolina Porras and associates presented the results of a study that investigated the prevalence of Helicobacter pylori infection and the risk factors associated with this infection among adults in Latin America. For this analysis, the authors used data collected during the initial screening visit of a randomized clinical trial for H. pylori eradication. This initial population included 1,852 adults who were screened for H. pylori infection by urea breath testing; study participants were from Santiago, Chile; Túquerres, Colombia; Guanacaste, Costa Rica; Copán, Honduras; Obregón, México; Tapachula, México; and León, Nicaragua. The patients were interviewed and completed the Rome III gastrointestinal symptom history questionnaire. Multivariate analyses were adjusted for sex, age, and study center.

The overall prevalence of H. pylori infection was 79.4% (95% CI, 77.5–81.2) regardless of sex, age, or study site. (One exception was Tapachula, Mexico, where the prevalence was only 70.1%). Several childhood demographic and socioeconomic conditions were associated with increasing odds of H. pylori infection, including increasing number of siblings (P<.0001) and factors such as earth flooring (OR, 1.8), more than 2 persons per bedroom (OR, 1.4), and lack of indoor plumbing (OR, 1.3). Current demographic and socioeconomic conditions that were associated with an increased prevalence of H. pylori infection included more than 3 children per household (OR, 1.7) and crowding (OR, 1.8). In contrast, the prevalence of H. pylori infection was lower among persons with more than 12 years of schooling (OR, 0.5) and those employed outside the home (OR, 0.7). Smoking, alcohol use history, and chronic dyspeptic symptoms were not associated with H. pylori infection. The researchers concluded that there is a high prevalence of H. pylori infection in these 6 Latin American countries and that infection was associated with poor socioeconomic conditions.

Effect of Amiloride on Acid-Induced Heartburn in Patients with Nonerosive Esophageal Reflux Disease

Acid-sensing ion channels (ASICs) are esophageal nocicepters that have been proposed to mediate heartburn. To determine if ASICs are involved in heartburn in patients with nonerosive esophageal reflux disease, William J. Bulsiewicz and colleagues performed a randomized, double-blind, crossover study in which patients underwent esophageal perfusion with either amiloride (a diuretic agent known to inhibit ASICs) or placebo prior to acid-induced heartburn. The results of this study were presented at the 2012 DDW meeting.

Inclusion criteria included heartburn symptoms for at least 6 months, moderate heartburn for 3 of the last 7 days, partial or complete response to PPI therapy, absence of erosive/EoE, and a positive modified Berstein test. Twenty-three patients were screened for this study, and 14 patients met the inclusion criteria (7 complete PPI responders and 7 partial PPI responders); 13 of the 14 patients were negative for H. pylori infection. Patients were randomized to receive intraesophageal perfusion with amiloride (10 mg total) or placebo for 5 minutes. Patients then underwent perfusion with hydrochloric acid (100 mM concentration) for
15 minutes or until the onset of heartburn. Once heartburn had resolved, patients were perfused with the other treatment (either amiloride or placebo) for 5 minutes, after which they again underwent acid perfusion.

As measured by an increase from baseline on a 10-point visual analog scale, heartburn severity was only slightly lower when patients were pretreated with amiloride versus placebo (2.5±0.33 points vs
2.64±0.45 points, respectively). However, the time to heartburn onset, as measured from the start of acid perfusion, was longer with amiloride than with placebo (2.93±0.3 min vs 2.36±0.29 min, respectively; P>.05). The results were comparable regardless of which agent was perfused first. No differences were observed between partial PPI responders compared to complete PPI responders. While amiloride prolonged the time until onset of heartburn, the difference compared to placebo was not statistically significant. The investigators suggested that this finding might be due to suboptimal absorption or contact time of the drug.


Presentations in Hepatology

High SVR Rates in Prior Treatment Failures After Re-Treatment with Boceprevir-Based Therapy

The PROVIDE study enrolled patients from the control arms of phase II and phase III clinical trials of boceprevir (Victrelis, Merck); these patients had all received at least 12 weeks of peginterferon and ribavirin but failed to achieve sustained virologic response (SVR) due to futility, virologic breakthrough, or relapse. During the PROVIDE study, patients received boceprevir (800 mg 3 times daily), peginterferon (1.5 µg/kg/week), and weight-based ribavirin (600–1,400 mg/day) for up to 44 weeks. Interim efficacy and safety data from this study were presented by Jean-Pierre Bronowicki at the 47th Annual Meeting of the European Association for the Study of the Liver (EASL), held April 18–22, 2012 in Barcelona, Spain.

The study enrolled 168 patients; 51% of patients were partial responders, 31% were null responders (<2 log10 decline in hepatitis C virus [HCV] RNA level at Treatment Week 12), and 15% were relapsers. A total of 138 patients were included in this interim analysis.

After the lead-in phase, 78% of prior null responders and 24% of prior partial responders/relapsers had a less-than-1 log10 decline in HCV RNA level; among these patients, SVR rates were 36% in prior null responders and 64% in prior partial responders. Among patients with at least a 1 log10 decline in HCV RNA level, the overall SVR rate was 68%: 55% in prior null responders, 72% in prior partial responders, and 56% in relapsers. The overall SVR rate at the end of the follow-up period was 59% (81 of 138 patients): 40% in prior null responders, 68% in prior partial responders, and 56% in relapsers. A multivariate analysis found that prior nonresponder status, baseline platelet levels, gender, and high viral load were independent predictive factors for SVR.

Serious adverse events were observed in 10% of patients. The most common adverse event was anemia, which occurred in 48% of patients; severe anemia (hemoglobin level <8.5 g/dL) occurred in 11% of patients. Other common adverse events included fatigue (47%), dysgeusia (34%), nausea (30%), and neutropenia (22%).

Ribavirin Dose Reduction Versus Erythropoietin for Anemia Management

The addition of an HCV protease inhibitor to peginterferon and ribavirin therapy can increase the risk for anemia. At the 2012 EASL Annual Meeting, Fred Poordad and colleagues presented the results of a study that sought to determine the efficacy, safety, and tolerability of erythropoietin administration versus ribavirin dose reduction for the treatment of anemia in patients receiving boceprevir, peginterferon, and ribavirin.

Treatment-naïve patients with chronic, genotype 1 HCV infection were enrolled in this study. Participants were at least 18 years of age, showed no evidence of hepatocellular carcinoma or co-infection, and had normal baseline hemoglobin levels (12–15 g/dL for female patients; 13–15 g/dL for male patients). Patients received 4 weeks of lead-in therapy with peginterferon
(1.5 µg/kg/week) and ribavirin (600–1,400 mg/day) followed by boceprevir (800 mg 3 times daily) plus peginterferon and ribavirin for a total treatment duration of either 28 or 48 weeks.

Hemoglobin levels at or below 10 g/dL after the 4-week lead-in period occurred in 73% (500/687) of patients; these patients were randomly assigned to either a 200–400-mg/day reduction in ribavirin dose (n=249) or 40,000 IU/week erythropoietin (n=251). If hemoglobin levels dropped to 8.5 g/dL or lower, secondary intervention was allowed. If hemoglobin levels dropped to 7.5 g/dL or lower, the patient was discontinued from the study.

Comparison of the ribavirin dose-reduction arm and the erythropoietin arm revealed no differences in end-of-treatment virologic response rates (82% in both groups), relapse rates (10% in both groups), or SVR rates (71% in both groups). Statistical analyses revealed that the probability of achieving SVR was similar for patients managed with ribavirin dose reduction and those given erythropoietin (P=.769), and SVR was not associated with the degree of hemoglobin decline among patients who developed anemia. Finally, the rates of serious adverse events and study discontinuations were similar for patients who were managed with ribavirin dose reduction and those who received erythropoietin.

100% SVR in IL-28B CC Patients Treated with 12 Weeks of Telaprevir, Peginterferon, and Ribavirin

In the ADVANCE trial, 90% of patients with genotype 1 HCV infection and the interleukin (IL)-28B CC genotype achieved SVR when treated with telaprevir (Incivek, Vertex), peginterferon, and ribavirin. To investigate SVR rates in patients with IL-28B genotype CC versus non-CC, Jean-Pierre Bronowicki and associates performed a retrospective analysis of data from the PROVE2 study; their results were presented at the 2012 EASL Annual Meeting.

Samples from treatment-naïve patients with genotype 1 HCV infection were analyzed for the presence of the IL-28B CC genotype at polymorphic site rs12979860. Of the 156 patients who consented to genetic testing, data were available for 141 patients. In terms of IL-28B genotype, 43 patients (30%) had genotype CC, 83 patients (59%) had genotype CT, and 15 patients (11%) had genotype TT.

All of the patients with IL-28B genotype CC who received telaprevir, peginterferon α-2a, and ribavirin for 12 weeks achieved SVR (n=12). SVR was also achieved in the majority of patients with IL-28B genotype CC who received other treatment regimens: 94% (15/16) of patients who were treated with telaprevir, peginterferon α-2a, and ribavirin for 12 weeks followed by peginterferon α-2a and ribavirin for an additional 12 weeks; 75% (3/4) of patients treated with telaprevir and peginterferon α-2a for 12 weeks; and 64% (7/11) of patients in the control group.

Burden of Illness in Treatment-Naïve HCV-Infected Patients in the United States

At the 2012 EASL Annual Meeting, Antoine El Khoury and colleagues presented the results of a study that examined work productivity, daily activities, healthcare resource use, economic costs, and health-related quality of life among treatment-naïve, HCV-infected patients in the United States. The analysis was restricted to patients who reported physician-diagnosed HCV infection, no HIV/AIDS or hepatitis B virus (HBV) co-infection, and no prior or current treatment for HCV infection (n=306). The HCV-infected group was compared to an unmatched control group (n=73,586) and to a matched control group (n=306).

The Work Productivity and Activity Impairment questionnaire was used to assess impairment in work and nonwork activities. This questionnaire revealed that activity impairment was significantly greater in untreated HCV-infected patients (42.2%) than in matched controls (27.3%; P<.001). Impairment at work was assessed among employed HCV-infected patients (n=121) and matched controls (n=141). There was no significant difference in absenteeism (the percentage of work time missed due to the patient’s health in the past 7 days) between untreated HCV-infected patients and matched controls (5.0% vs 2.8%; P=.089); however, untreated HCV-infected patients had increased rates of presenteeism (the percentage of impairment at work due to the patient’s health in the past 7 days; 23.2% vs 13.1%; P<.001) and overall work impairment (combination of absenteeism and presenteeism; 26.2% vs 14.9%; P<.001).

In terms of healthcare resource utilization, untreated HCV patients had significantly more physician visits annually than matched controls (12.2 vs 8.2; P<.001), as well as more emergency room visits (0.76 vs 0.54; P=.023); however, there was no significant difference in hospitalizations (0.42 vs 0.25; P=.071). Associated direct costs were all higher among untreated HCV-infected patients. Additionally, health-related quality of life in untreated HCV-infected patients was poorer than in matched controls as shown by a lower mean Mental Component Summary score (43.7 vs 48.6; P<.001), a lower mean Physical Component Summary score (40.2 vs 44.9; P<.001), and a lower Health Utility score (0.65 vs 0.73; P<.001).

Rifaximin Salvage Therapy Is the Most Cost-Effective Strategy for Management of Chronic Hepatic Encephalopathy

Navin Paul and associates conducted a decision-analysis study to determine which treatment strategy is most cost-effective for treatment of hepatic encephalopathy (HE): rifaximin (Xifaxan, Salix Pharmaceuticals), lactulose, or a hybrid strategy involving both drugs. Results of this analysis were presented at the 2012 DDW meeting.

The investigators employed Markov modeling to test the cost-effectiveness of rifaximin monotherapy, lactulose monotherapy, combination therapy with both rifaximin and lactulose, and rifaximin salvage therapy (initiation of therapy with lactulose followed by crossover to rifaximin in cases of inadequate response or intolerance). This model analyzed the effects of these therapies for a cohort of 50-year-old patients with overt HE.

A systematic literature review identified randomized, controlled trials of patients with HE; all studies had at least 4 weeks of follow-up. Data from these studies were used to calculate probability estimates, which were then varied over a wide range in a sensitivity analysis. The model utilized a third-party payer’s perspective and included cost estimates from Medicare and Red Book for a patient with cirrhosis and HE. The primary outcome was discounted cost per life-year (LY) gained.

Rifaximin salvage therapy was the overall dominant strategy in a limited 6-month model. During the first year, 1.3 hospitalizations were projected for the lactulose arm, and 0.8 hospitalizations were projected for the rifaximin arm. When the analysis was extended to a lifetime horizon, rifaximin salvage therapy was the most effective strategy (2.5 LY), and lactulose monotherapy was the least effective strategy (2.1 discounted LY).

In terms of cost, rifaximin monotherapy was the most expensive strategy ($65,800). Lactulose monotherapy was the least expensive strategy ($61,300), but this lower long-term cost reflected a higher mortality rate in this treatment group. When balancing both costs and effectiveness, rifaximin monotherapy and rifaximin salvage therapy were found to be the most relevant options. Of these 2 options, the investigators concluded that rifaximin salvage therapy should be the dominant and preferred approach for treating patients with chronic HE, as it is both more effective and less expensive.

Efficacy of 5 Years of Tenofovir Disoproxil Fumarate in Chronic HBV-Infected Patients with High Viral Loads

Successful treatment of HBV infection in patients with high baseline viral loads remains a significant clinical challenge. At the 2012 DDW meeting, Stuart C. Gordon and coauthors presented the results of a phase III, randomized study that assessed the long-term efficacy of tenofovir disoproxil fumarate (TDF) in this population. All patients were initially randomized to receive 300 mg TDF or 10 mg adefovir dipivoxil (ADV; Hepsera, Gilead). At Week 48, eligible patients initiated open-label TDF for up to an additional 7 years.

Patients were classified according to baseline viral load: 129 patients had high viral loads (HBV DNA ≥9 log10copies/mL) and 512 patients had non–high viral loads (HBV DNA <9 log10copies/mL). Most of the patients in both groups were male (74%), but patients with high baseline viral loads were younger than patients with non–high viral loads (31 years vs 43 years, respectively). Among patients with high baseline viral loads, 34.9% had alanine aminotransferase (ALT) levels more than twice the upper limit of normal (ULN), 91.5% were hepatitis Be antigen (HBeAg)-positive at baseline, and only 10% had antibodies to HBeAg at baseline. In contrast, 53% of non–high viral load patients had ALT levels more than twice the ULN, 29% were HBeAg-positive at baseline, and 73% had antibodies to HBeAg at baseline.

At Week 240 on treatment, 69.5% of high viral load patients and 83.5% of non–high viral load patients had achieved ALT normalization. Among patients who were HBeAg-positive at baseline, hepatitis B surface antigen (HBsAg) loss occurred in 19.3% and 4.3% of high viral load patients and non–high viral load patients, respectively (P<.001); HBsAg seroconversion occurred in 13.6% and 4.3%, respectively (P=.011).

Regression of histologic cirrhosis was observed in both the high viral load and non–high viral load groups. Neither group had persistent viremia at Week 240. The majority of patients (96%) achieved HBV DNA levels below 400 copies/mL by Week 240 regardless of their baseline viral load; however, patients with non–high baseline viral loads achieved more rapid viral decline. Among high viral load patients who experienced virologic breakthrough, 2 patients had a conserved site change that did not reduce sensitivity to TDF, 3 patients had a polymorphic site change, and 6 patients had no change.

Outcomes of Oral Antiviral Treatment for Chronic HBV Infection in Routine Clinical Practice

The approved oral antiviral agents that are commonly prescribed for the treatment of chronic HBV infection include lamivudine (LAM), ADV, entecavir (ETV; Baraclude, Bristol-Myers Squibb), and TDF. At the 2012 DDW meeting, Mindie H. Nguyen and associates presented the results of a study in which they examined outcomes when these agents were used in a routine clinical care setting. The study included 957 consecutive patients treated with LAM, ADV, ETV, or TDF for at least 6 months. The endpoints of the study were complete viral suppression (HBV DNA level <40–60 IU/mL) and ALT normalization (ALT level ≤40 U/L) at 6 and 12 months of treatment.

Among treatment-naïve patients, complete viral suppression at 6 months was achieved in 38% of patients treated with LAM, 44% of patients treated with ADV, 63% of patients treated with ETV, and 63% of patients treated with TDF. At 12 months, complete viral suppression rates were 37% with LAM, 50% with ADV, 73% with ETV, and 79% with TDF. The ETV and TDF treatment groups also had similar rates of ALT normalization at 6 months (91% vs 88%, respectively) and 12 months (94% vs 93%, respectively). Baseline clinical characteristics were similar among treatment-naïve patients who received ETV (n=373) and those who received TDF (n=107).

Among treatment-experienced patients who were switched to ETV (n=165) or TDF (n=67), those who were switched to ETV had a higher mean HBV DNA level at the time of the switch (3.82 log10 IU/mL vs 2.33 log10 IU/mL, respectively; P<.001); patients switched to ETV also had a higher median ALT level at the time of the switch (35 U/L vs 29 U/L, respectively; P=.07). Among treatment-experienced patients, the ETV and TDF treatment groups had similar rates of complete viral suppression at both 6 months (75% vs 79%, respectively) and 12 months (82% vs 89%, respectively); they also had similar rates of ALT normalization at both 6 months (94% vs 90%, respectively) and 12 months (91% vs 94%, respectively). Patients treated with ETV or TDF showed no evidence of viral breakthrough or viral resistance. Overall, the authors concluded that ETV and TDF were associated with more favorable patient outcomes (compared to LAM and ADV) in both treatment-naïve and treatment-experienced patients.


Presentations in Endoscopy

Should Colonoscopy Be Repeated When Bleeding Recurs?

In a retrospective review presented at the 2012 DDW meeting, Parit Mekaroonkamol and colleagues sought to assess the usefulness of repeat colonoscopies performed for the same indication (other than colorectal cancer [CRC] screening or polyp surveillance). This study included patients who had undergone more than 1 colonoscopy for the same indication within 3 years. Patients were excluded if they had repeat colonoscopies due to poor preparation or suspected complications from the first colonoscopy, or if the colonoscopies were performed for CRC screening and/or surveillance.

Of the 19,772 colonoscopies performed between 2000 and 2010 at Albert Einstein Medical Center, 139 pairs of colonoscopies met the inclusion criteria. The mean time between procedures was 261 days. Reasons for repeating the colonoscopy included lower gastrointestinal bleeding (88.4%), a change in bowel habits (6.4%), and abdominal pain (5%). A change in management occurred after 27 of the 123 repeat colonoscopies performed for lower gastrointestinal bleeding and after 2 of the 7 repeat colonoscopies performed for abdominal pain (20.86% overall). Among the cases of recurrent lower gastrointestinal bleeding, the repeat colonoscopies identified 8 new hemorrhoid lesions, 7 actively bleeding lesions that required intervention, 7 previously undetected polyps, 3 cases of radiation colitis, 1 rectal ulcer, and 1 previously undetected cancer.

Of all the clinical parameters that were evaluated, only the length of time between colonoscopies was associated with a decreased likelihood that the repeat colonoscopy would lead to a change in clinical management. The OR for a change in management for procedures performed 365–630 days apart was 0.09 (95% CI, 0.01–0.74; P=.025), and the OR for a change in management for procedures performed 630–1,095 days apart was 0.26 (95% CI, 0.09–0.72; P=.01).

Does Body Mass Index or Procedure Difficulty Affect the Force Applied During Colonoscopy?

Several factors have been shown to affect the application of force during colonoscopy, including the endoscopist, anesthesia, and patient gender. In an observational study presented at the 2012 DDW meeting, Louis Y. Korman and colleagues assessed whether the patient’s BMI or the difficulty of the colonoscopy affected axial and longitudinal force patterns. To measure these forces, the investigators attached a handheld Colonoscopy Force Monitor (CFM) to the colonoscope insertion tube; the CFM measured axial and applied forces used by the endoscopist during insertion and withdrawal.

This study included data from 114 colonoscopies; patients included 62 men and 52 women (mean age, 55.6 years). The study included 37 normal-weight patients (BMI ≤24.9 kg/m2), 50 overweight patients
(BMI 25–29.9 kg/m2), and 24 obese patients (BMI ≥30 kg/m2). Analysis of variance found that increased patient BMI did not affect force parameters during colonoscopy. Specifically, BMI did not significantly affect either the average peak push force (21.6 N in normal-weight patients vs 25.1 N in obese patients), nor did BMI significantly affect examination time (18.6 minutes in normal-weight patients vs 20.4 minutes in obese patients).

The procedure difficulty was rated in 77 of the 114 patients: 23 procedures were rated as not difficult, 38 were rated as moderately difficult, and 16 were rated as difficult. While procedure difficulty was found to increase procedure time, it did not affect the force applied during the procedure.

While the endoscopist, patient gender, and anesthesia do affect the amount of force applied during colonoscopy, the authors of this study concluded that patient BMI and procedure difficulty do not have an effect on the force applied during colonoscopy.

A Review of SpyGlass and Non-SpyGlass Techniques in the Management of Pancreaticobiliary Disease

In a retrospective review presented at the 2012 DDW meeting, Gregory Lutzak and associates compared the single-operator SpyGlass Direct Visualization System (Boston Scientific) to other endoscopic techniques for visualization of the pancreaticobiliary tract. The investigators searched billing codes for all cholangioscopy and pancreatoscopy procedures performed in the past 8 years at Virginia Mason Hospital. A total of 205 patients were identified: 163 underwent cholangioscopy, 35 underwent pancreatoscopy, and 7 underwent pancreatic cystoscopy. The mean age of the patients was 63.3 years, and 54% of the patients were female. The primary endpoint of this study was a change in diagnosis or management following endoscopic retrograde pancreatography in combination with cholangioscopy or pancreatoscopy.

Of the 205 procedures included in this analysis, 143 used the SpyGlass system. Among patients who underwent endoscopic retrograde pancratography in combination with cholangioscopy, there were 140 diagnostic procedures (104 using the SpyGlass system), 20 therapeutic procedures (15 using the SpyGlass system), and 3 procedures for percutaneous transhepatic biliary tract drainage (3 using the SpyGlass system). Among the diagnostic cholangioscopy procedures, malignancy was the most common diagnosis (n=48 patients). Changes in patient management following diagnostic cholangioscopy occurred in 75% of procedures that used the SpyGlass system and in 92% of the procedures that used other techniques. For therapeutic cholangioscopy procedures, fragmentation of stones occurred in 80% (12/15) of procedures that used the SpyGlass system versus 75% (3/4) of procedures that used other techniques; complete eradication of stones occurred in 27% (4/15) and 25% (1/4), respectively.

In the pancreatoscopy group, intraductal papillary mucinous neoplasm was the most common finding (n=18). Changes in patient management or diagnosis following pancreatoscopy occurred in 67% (16/24) of cases that used the SpyGlass system and in 55% (6/11) of procedures that used other techniques. No changes in patient management occurred in the pancreatic cystoscopy group (0/7). Overall, the investigators concluded that the SpyGlass system compared positively to other endoscopic tools for visualization of the pancreaticobiliary system.

Advanced Polypectomy and EMR Can Yield a Low Recurrence Rate for Previously Attempted Large Polyps

As presented by Niket Sonpal and associates at the 2012 DDW meeting, advanced polypectomy and endoscopic mucosal resection (EMR) techniques may be a viable alternative to surgery in patients with large or difficult-to-remove colorectal lesions. In addition to allowing patients to avoid the risks of surgery, recurrence rates following such procedures are fairly low (10.5–20.4%).

This study included patients with large (≥2 cm) colorectal polyps in whom primary removal had failed or was not attempted due to the polyp’s size and/or location. The EMR technique used in these patients included submucosal injection, multiple snare sizes, avulsion techniques with specialized forceps, and ablation with argon plasma. For polyps in difficult locations, caps and retroflexion were used.

Of the 262 patients included in this analysis, 67% were male, and the mean age was 74 years (range,
52–92 years). The success rate of the procedure was high, with only 5.7% of patients (15/262) experiencing a recurrence. The majority of recurrences (86%) were in the right colon. Of the 15 cases of polyp recurrence, histology revealed that 13 were tubular adenomas (87%), 2 were high-grade dysplasia (13%), and 1 was a serrated adenoma (7%). Diverticulosis was more common in patients who experienced a polyp recurrence (60%) than in patients who did not experience a recurrence (30%).

The investigators concluded that the EMR technique evaluated in this study was safe and effective for patients with large, sessile, or difficult-to-remove colorectal polyps. The vast majority of these patients did not require surgery (90%), and the overall success rate for the procedure was 94%, with a recurrence rate (5.4%) much lower than that previously reported in the literature. Sonpal and colleagues noted that this EMR approach should be developed further to improve patient care, reduce the need for surgery, and reduce costs.

Safety and Efficacy of EMR Prior to Radiofrequency Ablation for Patients with Dysplastic Barrett Esophagus

The outcomes of EMR followed by radiofrequency ablation (RFA) for the treatment of nodular dysplastic Barrett esophagus have not previously been investigated in a large series of patients. At the 2012 DDW meeting, however, William J. Bulsiewicz and colleagues reported results of an analysis in which they used data from the US RFA Registry to assess the safety and efficacy of this approach. This prospective registry includes patients with Barrett esophagus who were treated with RFA at 113 community-based and 35 academically affiliated institutions. Among patients with high-grade dysplastic Barrett esophagus or intramucosal carcinoma, treatment with EMR/RFA for nodular dysplastic Barrett esophagus was compared to treatment with RFA alone for non-nodular dysplastic Barrett esophagus. This study assessed both safety (rates of stricture, bleeding, and hospitalization) and efficacy (complete eradication of intestinal metaplasia [CEIM], complete eradication of dysplasia [CED], and number of treatment sessions to CEIM).

A total of 1,248 patients were treated with RFA for high-grade dysplastic Barrett esophagus or intramucosal carcinoma. Among the 418 patients (33%) who had undergone 1 or more prior EMR procedures, strictures developed in 3.6%, bleeding occurred in 0.7%, hospitalization was necessary in 1.4%, and 0% had perforations. There were no significant differences in the rates of stricture, bleeding, or hospitalization for patients receiving EMR/RFA compared to those receiving RFA alone.

In terms of efficacy, biopsy data were available for 44% of patients (554/1,248). Among patients treated with EMR/RFA, 65% achieved CEIM, and 81% achieved CED. Among patients with high-grade dysplastic Barrett esophagus and intramucosal carcinoma, those treated with EMR/RFA did not differ significantly from those who received RFA alone in terms of the rate of CEIM (61–72% vs 61–75%, respectively) or the rate of CED (77–88% vs 84–75%, respectively). Overall, the results of this large, multicenter registry found that treatment with EMR/RFA for nodular dysplastic Barrett esophagus yielded similar rates of CEIM, CED, and complications compared to treatment with RFA alone for non-nodular dysplastic Barrett esophagus.

Impact of Periampullary Diverticula on Success and Complication Rates of ERCP 

Periampullary diverticula (PAD) have been detected in 5–25% of patients undergoing endoscopic retrograde cholangiopancreatography (ERCP), but the effect of PAD on ERCP success and complication rates is unclear. At the 2012 DDW meeting, Olga Barkay and associates presented the results of a retrospective analysis that addressed this question using data from a prospectively maintained database at the ERCP Unit of the Indiana University Hospital. Between 1994 and 2009, a total of 31,635 ERCP procedures were performed in 19,197 patients; 1,315 of these patients had PAD (6.85%). After excluding those patients who had undergone a previous sphincterotomy, 780 patients remained in the PAD group (mean age, 63.5 years; 64.7% female). The age-matched control group for this study included 1,566 patients with naïve papilla who did not have PAD (58.2% female). Barkay and colleagues’ study compared these 2 groups in terms of the following outcomes: rate of prior ERCP failure, rate of successful cannulation, rate of difficult cannulation, use of precut sphincterotomy, and complication rate.

ERCP was performed for a variety of indications: choledocholithiasis (30% in the PAD group vs 13.2% in the control group), suspected sphincter of Oddi dysfunction (15.8% vs 16%), pancreatitis (22.9% vs 23.5%), obstructive jaundice (11.7% vs 23.9%), biliary dilation (11% vs 13.5%), and other indications (8.2% vs 9.9%). The proportion of patients who had previously failed ERCP in a community hospital was greater in the PAD group (21%) compared to the control group (13.8%; P<.001). There was no significant difference in the rate of successful cannulation (95.2% in the PAD group vs 94.9% in the control group; P=.7); however, cannulation was deemed to be difficult in a higher proportion of patients in the PAD group compared to the control group (18.2% vs 6.3%; P<.001). Other differences between the PAD group and the control group included more frequent use of precut sphincterotomy (10.8% vs 5.8%; P<.001), more frequent bleeding (1.28% vs 0.38%; P=.01), and more frequent perforation (1% vs 0.13%; P=.002). There was no difference in the rate of post-ERCP pancreatitis (2.95% in the PAD group vs 2.94% in the control group; P=.99). The authors concluded that ERCP can be successfully performed in patients with PAD, although these patients frequently require use of more aggressive techniques, such as precut sphincterotomy.

Early and More Aggressive Fluid Resuscitation Is Associated with Less Severe Post-ERCP Pancreatitis

Early fluid resuscitation can reduce the severity of acute pancreatitis, but data are lacking regarding the impact of volume resuscitation on the severity of post-ERCP pancreatitis. During the 2012 DDW meeting,

Sashidhar Sagi and colleagues presented the results of a retrospective cohort study that compared intravenous volume resuscitation in patients with mild or moderate/severe post-ERCP pancreatitis (PEP). All patients included in this analysis were admitted with pancreatitis (new or worsening abdominal pain with elevation in amylase or lipase level >3 times ULN) within 24 hours after ERCP. PEP was categorized as mild (hospitalization for ≤3 days), moderate (hospitalization for 4–10 days), or severe (hospitalization for >10 days). Exclusion criteria included absence of data on intravenous volume resuscitation, acute pancreatitis within 7 days of ERCP, chronic kidney disease, and congestive heart failure.

Of the 113 cases of PEP that were identified, 70 met the eligibility requirements. Forty patients had mild PEP, 27 patients had moderate PEP, and 3 patients had severe PEP. Patients with mild or moderate/severe PEP had comparable demographic and procedural risk factors for PEP; however, patients with moderate/severe PEP were older than those with mild PEP (median age, 50 years vs
36 years; P=.05).

There was no difference between patients with mild PEP versus those with moderate/severe PEP in terms of the median intravenous volume infused before and during ERCP (600 mL for both groups). However, a greater intravenous volume was infused during the first 24 hours after ERCP for patients with mild PEP (2,892 mL) compared to those with moderate/severe PEP (2,147 mL; P=.03). The proportion of patients who were discharged and then readmitted was also significantly lower in the group with mild PEP versus those with moderate/severe PEP (15% vs 40%; P<.01).


Presentations in IBD

Safety and Efficacy of Subcutaneous Golimumab Induction Therapy in Patients with Moderately to Severely Active Ulcerative Colitis

Numerous agents that inhibit tumor necrosis factor-α (TNF-α) have been tested for the treatment of ulcerative colitis (UC). One anti–TNF-α agent, golimumab (Simponi, Janssen Biotech)—currently approved for treatment of rheumatoid arthritis, psoriatic arthritis, and ankylosing spondylitis—is being evaluated for treatment of UC in the PURSUIT-SC study, the results of which were presented by William Sandborn at the 2012 DDW meeting.

PURSUIT-SC was a randomized, placebo-controlled, double-blind, phase II/III trial that enrolled UC patients with moderately to severely active disease who were naïve to anti–TNF-α therapy. The PURSUIT-SC trial began as a phase II dose-ranging study, after which patients were integrated into the confirmatory phase III portion of the study. During the dose-ranging portion of the study, patients were randomized to 1 of 4 arms: placebo, 100/50 mg golimumab (100 mg at Week 0 and 50 mg at Week 2), 200/100 mg golimumab (200 mg at Week 0 and 100 mg at Week 2), or 400/200 mg golimumab (400 mg at Week 0 and 200 mg at Week 2). During the phase III portion of the study, only the 200/100 mg and 400/200 mg doses of golimumab were used. Golimumab was administered subcutaneously in all groups.

The primary endpoint of the study was clinical response at Week 6, which was defined as a decrease in the Mayo Clinic score of at least 30% and at least 3 points from baseline, with either a decrease in the rectal bleeding subscore of at least 1 point from baseline or a rectal bleeding subscore of 0 or 1. Secondary endpoints included clinical remission (defined as a Mayo Clinic score ≤2 with no individual subscore >1), mucosal healing (defined as a Mayo Clinic endoscopy subscore of 0 or 1), and change from baseline in Inflammatory Bowel Disease Questionnaire (IBDQ) score, all assessed at Week 6.

A significantly higher proportion of patients in the golimumab treatment groups attained clinical response at Week 6 (51.8% and 55.0% in the 200/100 mg and 400/200 mg golimumab arms, respectively, vs 29.7% in the placebo arm; P<.0001 for both comparisons vs placebo). A highly significant difference also emerged in terms of the proportion of patients who achieved clinical remission at Week 6 (6.3%, 18.7%, and 17.8% in the placebo, 200/100 mg golimumab, and 400/200 mg golimumab groups, respectively; P<.0001 for both comparisons vs placebo) and mucosal healing at Week 6 (28.5% in the placebo group vs 43.2% in the 200/100 mg golimumab group and 45.3% in the 400/200 mg golimumab group; P=.0005 and P<.0001, respectively). The mean change from baseline in IBDQ scores at Week 6 was 14.6 points in the control group versus 27.4 points in the 200/100 mg golimumab group and 27.0 points in the 400/200 mg golimumab group (P<.0001 for both comparisons vs placebo).

The PURSUIT-SC study also evaluated the overall phase II/III trial population through Week 6 to assess the safety profile of golimumab; this analysis included a total of 1,065 patients. The total proportion of patients who experienced an adverse event was 38.2% in the placebo group versus 39.1% for the combined golimumab group. The number of patients who experienced a serious adverse event was also relatively similar in both groups (6.1% in the placebo group vs 3.0% in the combined golimumab group). One patient in the 400/200 mg golimumab arm died, and demyelination (a well-described toxicity of anti–TNF-α therapy) was reported in 1 patient in the 400/200 mg golimumab arm.

Vedolizumab Induction Therapy for Ulcerative Colitis

A potential new therapy for UC, vedolizumab, is a novel, gut-selective, monoclonal antibody directed against the α4β7 integrin that induces selective inhibition of lymphocytic trafficking in the gut. At the 2012 DDW meeting, Brian Feagan presented the results of a randomized, placebo-controlled, double-blind, multicenter, phase III trial designed to determine the long-term efficacy and safety of vedolizumab when given as induction therapy for UC. The intent-to-treat population for the induction phase of this study consisted of 374 patients with active UC. Patients were randomized 3:2 to treatment with vedolizumab or placebo; vedolizumab was administered as a 300-mg intravenous dose on Days 1 and 15.

The primary study endpoint for the induction phase of the study was clinical response at Week 6; clinical response was defined as a reduction in the total Mayo Clinic score of at least 3 points and a decrease from baseline of at least 30% plus a decrease in the rectal bleeding subscore of at least 1 point or an absolute rectal bleeding subscore no greater than 1. Secondary endpoints for the induction phase of the study included mucosal healing and clinical remission. Clinical remission was defined as a total Mayo Clinic score no greater than 2 points with no individual subscore greater than 1; mucosal healing was defined as a Mayo Clinic endoscopy subscore no greater than 1.

The rate of clinical response at Week 6 was significantly higher in the vedolizumab arm compared to the placebo arm (47.1% vs 25.5%; P<.0001). Vedolizumab also showed significantly higher rates of clinical remission at Week 6 (16.9% for vedolizumab vs 5.4% for placebo; P=.0010) and mucosal healing at Week 6 (40.9% for vedolizumab vs 24.8% for placebo; P=.0013). Finally, preliminary analyses of the safety data through Week 6 showed similar rates of adverse events, serious adverse events, and serious infections for the vedolizumab group and the placebo group.

Analysis of a Prospective Registry of Pregnancy Outcomes in Women with IBD Exposed to Immunomodulators and Biologic Therapy

At the 2012 DDW meeting, Uma Mahadevan summarized results from a study designed to determine the safety of immunomodulator and biologic therapy during pregnancy. This study enrolled a large prospective cohort of pregnant women with inflammatory bowel disease (IBD; N=1,115); 57% of these patients had Crohn’s disease (CD), and 40% had UC. Patients were divided into 4 categories based on their drug exposure between conception and delivery: (1) unexposed patients (n=306) who did not receive immunomodulator or biologic therapy during the study; (2) immunomodulator-treated patients (n=204) who received either azathioprine or 6-mercaptopurine; (3) anti–TNF-α–treated patients (n=291) who received either infliximab (Remicade, Janssen Biotech), adalimumab (Humira, Abbott), certolizumab pegol (Cimzia, UCB), or natalizumab  (Tysabri, Elan/Biogen Idec) during the study; or (4) combination-treated patients (n=75) who received both immunomodulators and anti–TNF-α agents during the study.

After adjusting for the effects of the underlying disease, most adverse incidents—including spontaneous abortions or congenital anomalies—did not occur at a significantly increased rate among women enrolled in this study compared to community-based or national rates. However, there were a few exceptions: Women in this study had a higher rate of Cesarean sections, and their babies had a higher rate of neonatal intensive care unit stay. Also, there were higher rates of spontaneous abortions and Cesarean sections in the anti–TNF-α group, and there was a higher rate of preterm births among women in the combination therapy group.

While babies of mothers with CD showed no increase in any complications or adverse effects, a nearly 5-fold higher rate of spontaneous abortion was observed among mothers with UC who were treated with anti–TNF-α agents. Further, UC mothers in the combination therapy group had an increased risk of any complication—including preterm birth, low birth weight, and neonatal intensive care unit stay—after the analysis adjusted for disease activity.

There were no significant differences in the growth characteristics of the babies throughout their first year, including height, weight, and developmental measurements at 4, 9, and 12 months of age (adjusted for maternal age and disease activity). In addition, no association was found between congenital anomalies and drug exposure. Finally, the rate of infections among the infants was not significantly affected by drug exposure. Taken together, these results suggest that the mother’s disease may confer a higher risk to the fetus than the risks associated with the use of immunomodulator or biologic therapy.

One Third of Patients Treated with Adalimumab or Infliximab Permanently Dose Escalate Due to Loss of Response

Despite the significant efficacy of infliximab and adalimumab for the treatment of IBD, many patients lose response to these drugs. In a poster presented at the 2012 DDW meeting, Darryl Fedorak and colleagues reported the findings of a retrospective chart review in which they sought to determine the incidence of loss of response among patients treated with either of these 2 agents.

The investigators identified 363 patients who met the inclusion criteria for this study. All enrolled patients had an initial response to induction dosing with either infliximab (5 mg/kg administered at Weeks 0, 2, and 6) or adalimumab (160 mg and 80 mg administered at Weeks 0 and 2, respectively). Patients also had to have advanced to scheduled maintenance therapy (every 8 weeks with infliximab or every 2 weeks with adalimumab), achieved a stable corticosteroid-free clinical benefit that was durable for a minimum of 6 months, and exhibited a loss of response to their anti–TNF-α therapy.

At the time of the analysis, 65% of infliximab-treated patients remained in remission while on infliximab maintenance therapy (5 mg/kg every 8 weeks). Similarly, 72% of adalimumab-treated patients were in remission on maintenance therapy (40 mg every other week). Thirty-five percent of patients who received infliximab required dose escalation (to 5 mg/kg every 4 weeks). Twenty-eight percent of adalimumab-treated patients required dose escalation (to 40 mg weekly). There was no significant difference in the rates of dose escalation between UC and CD patients. Further, a Kaplan-Meier plot found no significant difference in the time to treatment failure between infliximab and adalimumab (log-rank test, P=.56). Only 7 infliximab-treated patients underwent dose de-escalation to the original doses; none of the adalimumab-treated patients underwent dose de-escalation.

New Assay Can Detect Infliximab Levels and Anti-Infliximab Antibodies From a Single Serum Sample 

The most widespread method for detection of antibodies to infliximab (ATIs) is a double-antigen enzyme-linked immunosorbent assay, which uses infliximab as both the ligand and the detection antibody. However, this assay is limited by its inability to accurately determine ATI levels in the presence of serum infliximab concentrations. In a poster presented at the 2012 DDW meeting, Gabor Veres and colleagues reported on the development of a novel homogeneous mobility shift assay and demonstrated that it could detect both infliximab and ATIs in the same serum sample.

This novel homogeneous mobility shift assay was used to measure serum infliximab concentrations and ATI levels in 230 serum samples from 71 pediatric IBD patients. A subset of these children (n=31) also had
6 serial trough infliximab measurements, each taken prior to an infusion. A 5 mg/kg–induction dose of infliximab was administered at Weeks 0, 2, and 6, followed by maintenance dosing every 8 weeks.

ATIs were detected in 20.4% of the serum samples (range, 0.28–800+ U/mL) and in 29.6% of the 71 children. Of the 47 ATI-positive serum samples, 8 also demonstrated measurable infliximab serum concentrations (range, 0.77–19.27 mg/mL). In the subset of children with serial trough level measurements, 8 had ATI-positive serum samples. Among ATI-positive samples, the median infliximab serum concentration was 0 mg/mL; in contrast, the median infliximab serum concentration among ATI-negative samples was 2.55 mg/mL (P<.0001). None of the ATI-positive samples exhibited infliximab serum concentrations of 3 mg/mL or higher, while 45% of the ATI-negative samples had infliximab levels of 3 mg/mL or higher. Finally, ATI-positive patients also had C-reactive protein (CRP) levels that were approximately 1.5-fold higher than CRP levels in ATI-negative patients. A linear regression model found that a majority (88%) of children in the subset of patients with serum infliximab concentrations of
3 mg/mL or higher showed a decrease in CRP levels.

Association of Serum Infliximab and Antibodies to Infliximab to Long-Term Clinical Outcome in Acute Ulcerative Colitis

In a poster presented at the 2012 DDW meeting, Sanjay Murthy and colleagues used a newly developed homogeneous mobility shift assay to assess the relationships among trough infliximab levels, ATI levels, and long-term clinical outcomes in patients with acute UC. A total of 134 patients with corticosteroid-refractory acute UC were included in this analysis; 103 patients had pancolitis, and 31 patients had disease limited to the splenic fixture. All patients had received 5 mg/kg infliximab induction therapy on Weeks 0, 2, and 6, followed by scheduled maintenance therapy.

After a median follow-up period of 19.9 months (interquartile range [IQR], 7.6–47.4 months), 43.3% of patients were in corticosteroid-free remission, and 39.6% had undergone colectomy. The median time to colectomy was 6.5 months (IQR, 2.3–13.4 months). Among 125 patients with evaluable serum samples, 54.4% (n=68) had detectable trough levels of serum infliximab. Of these 68 patients, 6 patients (8.8%) also had detectable levels of ATIs. Of the 57 patients (45.6%) who had undetectable trough serum infliximab levels, 45 patients (78.9%) were ATI-positive, and 12 patients (21.1%) were ATI-negative.

Importantly, the investigators showed that a trough infliximab level above 2 mg/mL was associated with a higher rate of corticosteroid-free remission compared to a trough infliximab level of 2 mg/mL or lower (69% vs 16%; P<.001). This relationship was sustained throughout the follow-up period. In contrast, a trough infliximab level below 2 mg/mL was significantly associated with an increased risk for colectomy compared to a trough infliximab level above 2 mg/mL (64% vs 13%; P<.001).

Millennium Medical Publishing, Inc