A-OK for Accutane

Another article has reaffirmed that isotretinoin (Accutane) does not increase the risk of IBD (JAMA Dermatol 2013; 149: 216-20).  Thanks to Mike Hart for this reference. However, this data will not reverse the millions of dollars that have been lost in litigation (Isotretinoin – Wikipedia, the free encyclopedia).

Using a large U.S. health claims database (68 million patients), the authors examined women ages 18-46 years who had received at least one oral contraceptive prescription between 2001-2009.  For each patient with IBD, 20 controls were identified in a nested case-control study design.

In total, 2159 IBD cases (1056 UC, 1103 CD) were matched with 43,180 controls.  Only 10 patients with IBD were exposed to isotretinoin compared with 191 controls.  The adjusted relative risk (RR) for IBD was 0.99; for ulcerative colitis the RR was 1.1 (confidence intervals 0.44-2.7) and for Crohn’s disease the RR was 0.91 (confidence interval 0.91).  For the meta-analysis which was a secondary part of this study, the RR for IBD with 5 studies was 0.94.

Conclusion: The study results do not suggest an increase risk of IBD with isotretinoin use.

Why did previous studies suggest a link between IBD and isotretinoin? The authors note that this is the first study to adjust for two main confounders, mainly a diagnosis of acne and use of oral tetracycline antibiotics.  Oral antibiotics, including tetracyclines, have been associated with IBD previously.  In addition, the design limits the confounding of contraceptive usage.

Limitations of this study:

  • Only women were studied; however, there are no known biologic factors that would make isotretinoin more problematic for males.
  • Other risk factors were not examined: smoking, ethnicity, diet, IBD family history

Additional references:

  • -Am J Gastroenterol 2009; 104: 2774-78.  Population-based study in Winnipeg, <40yrs.  n=1960 cases and 19,419 controls.  No differences in the proportions of IBD cases taking isotretinoin vs controls.  1.2% of IBD cases received isotretinoin prior to IBD diagnosis (n=25) compared with 1.1% of controls (n=213).  Mean # of days prior to IBD dx was 1102.  Thus, isotretinoin unlikely to be causally-associated with idiopathic IBD.
  • -IBD 2009; 12: Supplement -abstract O -0002  Increased risk of UC after isotretinoin.  OR 4.36 for developing UC
  • -Am J Gastroenterol 2009; 104: 2387-93.  7 country study found no causal association between isotretinoin & colitis.

Mortality from IBD

A large Danish cohort provides useful information on the mortality effects of having inflammatory bowel disease (IBD) and how this has changed over the last three decades (Clin Gastroenterol Hepatol 2013; 11: 43-48).

Using a national cohort of all individuals living in Denmark between 1982-2010 (on average 5.3 million), the authors studied 36,080 patients with ulcerative colitis (UC) and 15,361 with Crohn’s disease (CD) in comparison to data from 2,858,096 matched controls (50 controls from each IBD patient) from the general population.  For UC, the median age was 45.2 years and for CD 36.3 years.

Findings:

  • With UC, the first year of diagnosis carried a higher risk of dying (HR 2.43) then rapidly declined to around HR 1.1 after 2 years.  Overall, long-term mortality was increased 10% among UC patients.  Mortality from UC decreased from 1982 to 2010 due to decreased mortalities from gastrointestinal disorders, including colorectal cancer.
  • Mortality was higher among patients diagnosed at younger ages.  Patients diagnosed with UC in childhood or adolescence had a 2.15-fold higher relative mortality than patients diagnosed with UC at 60-79 years.
  • Cause-Specific Mortality: during the first year after UC diagnosis, the HR ratio was increased markedly for gastrointestinal disease (HR 13.3) and infection (HR 9.17).  These areas were most prominent at 10+ years, but the HR ratios had decreased to 1.95 and 1.64 respectively at that time.
  • For CD, mortality was markedly increased in the first year with HR 3.69 and declined to HR 1.53 during years 2-4; HR was 1.49 at 10+ years following diagnosis.  Overall, long-term mortality was increased 50% among CD patients.  Unlike UC, no improvement in mortality rate occurred during the study.
  • Cause-Specific Mortality: during the first year after CD diagnosis, the HR ratio was increased markedly for gastrointestinal disease (HR 23.02) and infection (HR 10.19). These areas were most prominent at 10+ years, but the HR ratios had decreased to 3.67 and 2.70 respectively at that time.  Infections were not increased in the most recent decade, indicating that thiopurines and biologics have not increased the overall risk of fatal infections.
  • While the relative risk of cardiovascular disease was only slightly increased (HR 1.11 for both CD and UC), this is important given the overall frequency of cardiovascular disease in the population.  Presumably, systemic inflammation contributes to the formation of atherosclerosis.
  • Suicide was also increased, especially in the first year after diagnosis (HR 2.05 for UC and HR 1.37 for CD)

Strengths of study: since Denmark has free access to health care and all citizens have a unique 10-digit personal identification, this enables capture of virtually all patients with IBD.  Previous studies have validated the database to be accurate and that the IBD diagnoses have been validated.

Related blog entries:

More frequent pediatric IBD

Using a national cohort of prospective and retrospective data on pediatric inflammatory bowel diagnosis, (PIBD) a recent report indicates a rising incidence of PIBD in Scotland (Inflamm Bowel Dis 2012; 18: 999-1005).

Key findings:

  • Between 2003-2008, 436 patients were diagnosed with PIBD; this equates to 7.82/100,000; Crohn’s was 4.75 and UC 2.06 per 100,000 respectively
  • Between 1990-1995, 260 patients were diagnosed with PIBD; this equates to 4.45/100,000; Crohn’s was 2.86 and UC 1.59 per 100,000 respectively
  • Mean age of diagnosis in more recent cohort was younger: 12.7 years compared with 11.9 years

Currently, a data-base coordinated in a single center along with collaboration with regional networks enables the capture of all new IBD diagnosis (since 1999).  Prior cohort was determined by “exhaustive examination” of previous IBD records.

How new therapies impact colectomy in UC patients

Has the need for colectomy changed with the increasing use of more potent medical therapies for ulcerative colitis (UC)?  One article gives some insight into this question (Inflamm Bowel Dis 2012; 18: 1641-46).

This French study followed 151 patients with newly diagnosed UC from 2000-2008; median followup was 58 months. During this time, 21 patients (14%) underwent colectomy.  1.3% required colectomy in the first year following diagnosis.

Looking closer at their study, 55% of patients had pancolitis.  Cyclosporin usage, typically given in refractory cases, was the only medication determined to be a predictive factor for surgery.

Medication usage during study period:

  • 68% oral mesalamine products
  • 72% systemic corticosteroids
  • 7% methotrexate
  • 49% azathioprine
  • 9% cyclosporin
  • 30% had received at least one anti-TNF agent

The authors concede several limitations, including the evolving nature of UC treatment.  Yet, they conclude that colectomy still is frequently needed and the use of IBD medications, including anti-TNF, “does not appear to reduce the long-term need for surgery in UC.”

I take issue with the last sentence.  Whether anti-TNF agents prove to be a disease-modifying treatment over the long-term is not known.  In this particular cohort, only 30% were even exposed to these agents.  My conclusion: we need a study designed to answer the question.  This would require larger numbers of patients followed prospectively for many years.

Related blog entries:

According to the study which you would never qualify for…

When I give patients advice on treatment, I rely on well-designed studies to help select the medications/treatments that are most likely to be effective.  It is interesting to ponder how applicable these studies are and how closely they reflect the patient population that I’m seeing.  A recent article indicates that in a tertiary referral center for inflammatory bowel disease, the majority of patients would be excluded from participation in these studies (Clin Gastroenterol Hepatol 2012; 10: 1002-07).

These authors performed a retrospective cohort study of patients presenting to the Mount Sinai Medical Center between January 2008 and June 2009.  For Crohn’s disease (CD) patients, the authors selected 7 published randomized control trials of biological therapy for patients with moderate-severe disease (ACCENT 1, CLASSIC 1, CHARM, PRECISE 1, ENCORE, ENACT, and SONIC).  For Ulcerative Colitis (UC), they selected ACT 1 and ACT 2 trials for moderate-severe UC.

Only 31.1% of 206 patients with IBD would have been eligible to participate in any of the selected RCTs.  The average age of their patients was about 35 and did not differ significantly among patients who would qualify compared to those who would not.

Among the CD patients, the inflammatory phenotype was most likely to qualify; 37 patients with inflammatory disease behavior would have been eligible.  While only 54% (37 of 68) of the inflammatory phenotype would have qualified, this accounted for 86% of the total population who would have qualified (only 34% of the entire CD cohort would have been eligible for study participation).

Among the UC patients, there were no disease characteristics or demographics associated with eligibility; that is, there was a similar distribution of disease extents, Mayo scores, disease duration, gender, and age among the patients who qualified and those who did not.

Reasons for exclusion: stricturing or penetrating disease, steroid dosing, comorbities, concomitant therapies, or prior exposure to biologics.

The authors note that the results from this study mirror that from similar studies with other chronic diseases including rheumatoid arthritis, chronic obstructive pulmonary diseases, and congestive heart failure.

When the authors looked at the trial-ineligible patients, 60% had a response to biological therapy. The response rates for ineligible patients (compared to eligible patients) were lower for CD patients but not for UC patients (Figure 2 in study).

While stringent criteria for eligibility limit the external validity of the study results, these criteria are in place for multiple reasons.  Less rigid selection criteria would require larger study populations, longer duration studies, and greater costs.

The issue of study applicability is even a bigger problem in pediatrics.  This is primarily due to fewer high quality studies in children. As such, many clinical judgements require extrapolation from adult studies.

Related blog post:

Interpreting industry-funded studies

What is the risk of colon cancer in IBD?

Some recent studies have shown that colorectal cancer complicating IBD may not be as common as previously thought or may be decreasing in incidence (Gastroenterology 2012; 143: 375-81 & Gastroenterology 2012; 143: 382-89 ).

The first study used a nationwide cohort of 47,374 Danish patients with IBD over a 30-year period.  During a 178 million person-years of follow-up evaluation, 268 patients with ulcerative colitis (UC) and 70 patients with Crohn’s disease (CD) developed colorectal cancer (CRC).  The risk was comparable to the general population (RR 1.07)!  Furthermore, the relative risk for CRC decreased over sequential time periods: 1.34  (1979-88) to 0.57 (1999-2008).

Increased risk:

  • UC diagnosed in childhood or adolescence
  • longer disease duration
  • patients with primary sclerosing cholangitis

Their conclusion, ‘the overall risk of CRC in patients with UC has decreased markedly over time and no longer exceeds that of the general population, at least in the first decade after diagnosis.”  And, based on their data, patients with Crohn’s disease are not at increased risk for CRC.

The second study from California used a large Kaiser Permanente database from 1998-June 2010.  29 cancers were identified in CD (n=5603) and 53 in UC (n=10,895) patients.  The incidence per 100,000 for CRC was 75 for CD, 76 for UC, and 47 for general population.

These authors conclude that there is an increased risk of CRC in a community-based IBD population between 1998-2010 despite advances in medical treatment.

To understand the discrepancy of these reports, the same issue provides an editorial (page 288).  My take is that the current incidence of CRC is lower than in previous reports but that the risk factors identified in the Danish cohort (see above) likely remain important.

Additional references:

  • -Gastroenterol 2009; 136: 718. 22% of cancers occurred before recommended surveillance in adults (only 9% if exclusion of patients who had IBD and cancers diagnosed at same time).
  • Colon Cancer Survival Calculator http://www.mayoclinic.com/calcs-Gastro 2010; 138: 207-2177 (entire issue)
  • -JAMA 2009; 302: 649. Aspirin use likely increases survival after dx of colon cancer. Commentary-Gastro 2010;138: 2012.
  • -NEJM 2009; 361: 2449. molecular basis of colorectal cancer
  • -Gut 2008; 57: 1246. IBD and colon cancer
  • -IBD 2007; 4: 367. 5ASA Rx did not reduce rate of cancer in large study of UC/CD -review of 18,000 colorectal cancer cases. IBD increased risk of CRC 6-7-fold..
  • -Clin Gastro & Hep 2006; 4: 1346. aminosalicylates reduce CRC.
  • -Gastroenterol 2006; 130: 1030, 1039, 1350. 600 patients followed for 35 yrs: CRC by colitis duration: 2.5% @ 20yrs, 7.6% @ 30yrs, 10.8% @ 40yrs. 5 year survival was 73% among those with cancer. 2nd study showed main CRC risk in UC pts is among those with extensive colitis.
  • -Clin Gastro & Hepatol 2004; 2: 1088. Lower cancer risk in this cohort, n=1460. CRC 0.4 @10yrs, 1.1% @ 20yrs, 2.1% @ 30yrs. Lower CRC may be due to more surgery in Rx failures & use of 5-ASA.
  • -Clinical perspectives in Gastro 1999; 2: 9 & 25. (review) surveillance/ overview take 4 bx q10cm. start p 8yrs in pancolitis & 15yrs for left-sided dz. Cancer risk: ~5% at 20yrs + 1%/yr p 20yrs in pancolitis.
  • -Gastroenterol 2003; 125: 1311. Advancement of dysplasia to cancer in UC.

Vedolizumab -another new IBD treatment

Occasionally a parent will express concern that their child may run out of effective treatments for inflammatory bowel.  At this time, there are a limited number of effective therapies. If a patient develops reactions or antibodies to medications and/or does not have a clinical response, then the options become quite limited.

Against this backdrop, it is encouraging that new medical treatments are being tested, including Vedolizumab (Inflamm Bowel Dis 2012; 18: 1470-79).

The cited study reports the experience of using Vedolizumab in a randomized controlled phase 2 dose-ranging study in 46 patients (9 placebo, 37 vedolizumab).

Vedolizumab is “a gut-selective monoclonal antibody that antagonizes α4β7 integrin on select subsets of leukocytes binding to MAdCAM-1 on gut vascular endothelium.”  Many have considered vedolizumab to be similar to natalizumab, except it is considered  gut-specific and therefore should not increase the risk of progressive multifocal leukoencephalopathy (PML).  By altering the immune surveillance/lymphocyte tracking of the GI tract and not the brain, theoretically there should not be an increased risk of PML.

Another previous limitation of vedolizumab in small studies was the development of human antihuman antibodies (HAHA) –no joke!  HAHA occurred in up to 44% of patients in previous studies and were associated with reduced efficacy.  As a consequence, a new formulation of vedolizumab has been developed with presumably reduced  immunogenicity.

In the cited study, patients between 18-70 were recruited to receive one to three doses of vedolizumab (2, 6, or 10 mg/kg) or placebo.

Results:

  • Over 50% of vedolizumab-treated patients maintained a clinical response between days 29-253 whereas placebo response ranged between 22-33%.
  • This study was not powered for efficacy; however, the treated patients did have reduced fecal calprotectin.  At baseline, the calprotectin in the treatment groups averaged 405 μg/g and by day 113 had reduced to 54 μg/g.  In contrast, placebo group started with calprotectin of 310 μg/g and dropped to 192 μg/g during same time period.
  • HAHA were detected in 4 (11%)

According to a presentation at DDW (GEMINI I trial), more than 2500 patients have been treated in trials with vedolizumab.  None has developed PML. Encouraging results for efficacy (dosing 300mg IV q4weeks) were also noted; 1 year clinical remission noted in 45% of treated patients compared with 14% of placebo group. Digestive Disease Week 2012 – Vedolizumab Scores on Safety …)

Results of vedolizumab for Crohn’s disease are expected soon as well.

Related blog entries:

Quantifying Risk of PML with Natalizumab

Tofacitinib –a JAK Inhibitor for UC

Tofacitinib –a JAK Inhibitor for UC

There are definitely a lot of new therapies on the horizon for inflammatory bowel disease.  One of these agents is likely to be tofacitinib which has shown efficacy for active ulcerative colitis (NEJM 2012; 367: 616-24).

Background: Tofacitinib is a selective oral inhibitor of Janus kinase (JAK) which mediates activity for multiple cytokines, including interleukins 2, 4, 7, 9, 15, and 21.  Blockage of a common signaling molecule by these cytokines “should result in suppression of both T and B cells while maintaining regulatory T-cell function.  It has shown efficacy for organ allograft rejection, rheumatoid arthritis, and psoriasis.”  A previous small study by these investigators did not demonstrate efficacy in Crohn’s disease (Gastroenterology 2011; 140: Suppl: S124 Abstract).

Design: In this study which began as a double-blind, placebo-controlled, phase 2 trial, tofacitinib or placebo was given to 194 adult patients (from 51 centers in 17 countries) with moderate-to-severe active ulcerative colitis. Dosing for tofacitinib included groups receiving 0.5 mg, 3 mg, 10 mg, or 15 mg (all BID).  “The primary outcome was a clinical response at 8 weeks, defined as an absolute decrease from baseline in the score on the Mayo scoring system.”  Most of these patients had failed conventional therapy, including mesalamine, corticosteroids, immunosuppressants, and anti-TNF agents.

Results: At 8 weeks, the primary outcome with the highest doses (10 mg & 15 mg) of tofacitinib had clinical response rates of 61% (p=0.1) and 78% (p<0.001) respectively compared with a 42% placebo response.  Clinical remission (Mayo score ≤ 2) occurred in 48% and 41% respectively (both p<0.001) compared with 10% in placebo group.  Endoscopic remission was noted in 30% and 27% respectively (both p<0.001) compared with 2% of placebo group.

In addition, tofacitinib administration improved CRP values and fecal calprotectin concentrations (Figure 2 of article).

Potential adverse effects included the following

  • neutropenia (ANC 1000-1500) observed in three treated patients
  • two tofacitinib patients (10 mg group) developed abscesses
  • mild increases in LDL and HDL were noted and dose-related (these changes have been seen in rheumatoid arthritis patients as well)

Additional reference:

  • N Engl J Med 2012; 367:495-507 | August 9, 2012.  Tofacitinib for rheumatoid arthritis

Serology in IBD

Serological antibodies against a number of antigens have shown some utility in differentiating inflammatory bowel disease (IBD) from non-IBD and in distinguishing Crohn’s disease (CD) from ulcerative colitis (UC).  A recent article evaluated 204 articles in a systematic review of these serological markers (Inflamm Bowel Dis 2012; 18: 1340-55).

The study has several useful tables and a long list of references.  In its Table 1, 10 serologies are listed with a range for prevalence in CD, UC, alternative GI conditions, and in healthy population.  Table 2 summarizes the data in terms of sensitivity, specificity, positive predictive value, negative predictive value for these antibodies in determining IBD from non-IBD.

With regard to specific antibodies, the review highlights 10 antibodies:

1. Anti-neutrophil cytoplasmic antibodies (ANCA).  Autoantibody directed against a constituent of neutrophil granules.  With IBD (especially UC), an atypical perinuclear (pANCA) staining pattern with indirect immunofluorescence and DNase-sensitive make this pattern different from ANCA due to vasculitis.

2-7. Anti-glycan antibodies –directed against cell wall microbes and reflect interaction between the immune system and glycosylated cell wall components of microbiota.

2. Anti-Saccharomyces cerevisiae (ASCA IgA and IgG) –antibodies directed against yeast cell wall.  While ASCA antibodies are commonly found in CD patients, 20-25% (or higher in some studies) of healthy relatives will test positive for these antibodies as well.  Approximately 6% of relatives of UC patients will be ASCA-positive.

3. Anti-laminaribioside carbohydrate IgG antibodies (ALCA) –antibodies directed against laminaribioside

4. Anti-chitobioside carbohydrate IgA antibodies (ACCA) –antibodies directed against chitobioside

5. Anti-mannobioside carbohydrate IgG antibodies (AMCA) –antibodies directed against mannobioside

6. Anti-L –antibodies directed against laminarin (large polysaccharide)

7. Anti-C –antibodies directed against chitin (large polysaccharide)

8. Anti-OmpC.  OmpC is a transport protein of E coli

9. Anti-I2.  I2 is a Pseudomonas-associated antigen

10. Anti-CBir1.  CBir1 is a bacterial flagellin antigen

Conclusions:

  • Serology has only limited value for the initial diagnosis of IBD.
  • Serology has ‘better value’ in differentiating CD from UC, though there is substantial variability in serologic responses in both diseases.  Probably, serology is most useful in unclassified IBD (IBD-U) in preoperative setting; serology may help predict risk of developing complications among patients undergoing pouch surgery.
  • Serology is useful in predicting a complicated disease course. The presence and magnitude of these antibodies are strong predictors of disease progression.

Additional references:

  • Pediatrics. 2010 Jun ;125 (6):1230-6.  Shortcomings of the inflammatory bowel disease Serology 7 panel. 
  • -Clin Gastro & Hep 2008; 6: 1105. Increased immune reactivity/markers associated with aggressive disease.
  • -IBD 2006; 12:1122. Expression of I2 antibodies (against a bacterial antigen of psedomonas fluorescens) was highly associated with clinical response to diversion. 15/16 with I2-pos had clinical response; 2/11 I2-neg had clinical response.
  • -IBD 2008; 14: S4 abstract 0010. Practical experience with IBD serology (n=90) much less accurate than reported by Prometheus: overall accuracy of 63% (vs 92%), 66% sensitivity (vs 93%), 59% specificity (vs 95%), 75% PPV (vs 96%), and 49% NPV (vs 90%). In this population, 34% of known IBD were incorrectly predicted. Of 32 who did not have any evidence of IBD after clinical investigation, 40% (13) were seropositive.
  • -Clin Gastro & Hep 2008; 6: 1105. Increased immune reactivity/markers associated with aggressive disease.
  • -IBD 2008; 14; 129. Serologic markers not very useful clinically.
  • -Pediatrics 2007; 119: e193. IBD serology performed poorly in comparison to combination of Hgb/ESR with regard to sensitivity (60% vs. 83%), specificity (92% vs. 96%), positive predictive value (60% vs. 79%) for IBD in children, n=227. Also one third of all positive serology in patients w/o IBD. The positive predictive value in patients w/o rectal bleeding was 35% vs 60% for routine tests.
  • -Gastroenterol 2006; 131: 366. antibodies against laminaribioside, chitobioside, and mannan have predictive value in detecting Crohn’s disease.
  • -Gastroenterol 2006; 130: 1078. Unaffected relatives positive for either OmpC or ASCA in 20% in large cohor (n=619 unaffected relatives. OmpC present in up to 44% of CD pts, up to 24% of UC pts, and 6% of controls.

Colonic disease and PSC

While primary sclerosing cholangitis (PSC) has been associated with inflammatory bowel disease, and ulcerative colitis (UC) in particular, the pathogenesis of this relationship has not been established.  A fascinating observation on this relationship is that an inflamed colon is important in PSC development (Clin Gastroenterol Hepatol 2012; 10: 439-41).

In this study which reviewed 2754 Irish patients with IBD, 59 (2.2%) had PSC.  PSC incidence correlated with increasing colonic involvement.  Among the 13 patients with Crohn’s disease, none had isolated small bowel disease.  The second part of the study involved a review of 82 separate PSC patients attending the Irish National liver transplant unit.  The majority of ulcerative colitis patients had a pancolitis; all 10 PSC patients with Crohn’s disease had colonic involvement.

Since PSC occurs without IBD, colonic inflammation is not necessary for PSC development.  However, in patients with IBD, colonic inflammation is very important.  In fact, in a previous study of 53 PSC-IBD patients, no UC patient status post a colectomy had recurrent liver disease following liver transplantation whereas seven patients with intact colons had recurrent disease following liver transplantation.

The authors speculate that bacterial translocation along with subsequent portal bacteremia may be an important step in pathogenesis among these patients.

Additional references/previous posts: