Personalizing Therapy for Inflammatory Bowel Diseases

Ashwin N Ananthakrishnan


Expert Rev Gastroenterol Hepatol. 2013;7(6):549-558. 

In This Article

Can Response to Therapy Be Predicted?

The advent of the anti-TNF biologics over the past decade has improved our ability to achieve disease remission over traditional immunomodulators, reduced the likelihood of hospitalizations and surgery and improved health-related quality of life. Yet, response to therapy remains difficult to predict. About 30–40% of patients experience no benefit while an additional 10–15% lose response annually.[16,17]

One consistent clinical parameter predicting response to anti-TNF biologics is duration of disease. Given the progressive nature of CD with incremental risk of developing fibrostenosing complications, it is intuitive that therapy early on in the course of disease is likely to achieve a greater benefit.[76] Indeed, response rates are higher in pediatric CD when compared with adults. In a subanalysis of the PRECiSE 2 randomized controlled trial of certolizumab pegol for CD, maintenance of response was achieved in nearly 90% of patients with disease duration <1 year at initiation of treatment compared with only 57% of those with at least 5 years of disease.[77] Similar data also exist in the postoperative setting where infliximab had substantially greater efficacy in preventing postoperative recurrence than when used to treat endoscopic or clinically recurrent disease.[78] Patients with Crohn's colitis may have a higher likelihood of response than those with small intestinal involvement.[79] Current smoking may reduce response to infliximab, although not all studies have consistently demonstrated such an association.[79]

Serologic and genetic predictors of response to therapy have proven mostly disappointing. Esters et al. found no relationship between pANCA or ASCA positivity and response to infliximab therapy.[80] However, a combination of positive pANCA and negative ASCA in patients with CD was associated with a reduced response rate. By contrast, another study by Taylor et al. showed that speckled ANCA positivity was associated with improved response to infliximab.[81] For UC, a pilot study associated pANCA positivity with treatment resistance in the pre-anti-TNF era in patients with left-sided disease.[56] The limited literature on IBD-risk alleles and response to anti-TNF biologic therapy has not yielded consistent associations between risk loci and treatment response. Initial studies suggested an association between polymorphisms in the TNFR-1 (TNFR-1 + 36) or TNF-α and reduced response to infliximab infusions though other studies failed to identify similar associations.[81–91] Infliximab and adalimumab exert part of their effect by inducing cellular apoptosis. Consequently, it is plausible that genetic variants influencing apoptotic function could influence response to therapy. Two studies examining this hypothesis emerged from the Leuven IBD group. Both individual polymorphisms in the Fas ligand and caspase gene loci, or a cumulative apoptotic index encompassing all three loci were useful in predicting response in luminal and fistulizing CD.[92,93] A higher apoptotic index was associated with response rates of 70–90%, while those with an apoptotic index of 0 had a response rate of only 50%. Interestingly, the effect of the low apoptotic index could be overcome by addition of immunomodulators such that people with combination therapy had response rates of 79%, comparable with those with a high API. Dubinsky et al. adopted an integrated approach including both genetic and serologic markers, in a cohort of 94 pediatric IBD patients (both CD and UC).[94] A model incorporating BRWDI locus, three pharmacogenetic genome-wide significant loci, pANCA positivity and diagnosis (UC vs CD) had a superior accuracy in predicting response than individual clinical, genetic or serologic parameters.[94]

Further work on the role of genetics or genomics in predicting treatment response emerged through examination of gene expression from colon tissue. Arijs et al. performed RNA microarrays on pretreatment colonic mucosal biopsies from 20 responders and 26 non-responders with UC who received a single infusion of infliximab.[95,96] They were able to identify five probe sets that were upregulated and 207 that were downregulated in responders. The top five gene signatures encoded were: osteoprotegerin, stanniocalcin-1, prostaglandin-endoperoxide synthase 2, IL-13 receptor alpha 2 and IL-11; together they were able to predict response with an accuracy of 89%.