Perianal Crohn disease encompasses perianal skin tags, fissures, enterocutaneous fistulas and abscesses. It affects up to one-third of the CD population and it may be the only manifestation of the disease [33, 34]. Numerous clinical trials have been devised, primarily in adult populations, to determine the optimal management of this challenging condition.

There are two disease activity measures currently used in clinical trials to follow perianal CD activity: the perianal disease activity index (PDAI) and the Fistula Drainage Assessment (Appendices 2-1 and 2-2). The PDAI was developed and validated by Irvine and colleagues [35]. It contains five items, each scored 0–4, with higher scores representing more severe disease. In the validation cohort, it had moderate correlation with both physician and patient assessment of perianal disease activity (r  =  0.72 and 0.66 respectively). As expected, it correlated poorly with the CDAI and HBI (r  =  0.23 and 0.21 respectively) implying that perianal disease can present in the absence of other CD symptoms. The PDAI scores were reproducible in patients with stable disease over 4–8 weeks, implying test–retest reliability. Although the PDAI changed in patients who improved or deteriorated on repeat visits, robust and accepted responsiveness statistics were not presented. The second instrument, the Fistula Drainage Assessment, was introduced as part of a clinical trial of infliximab therapy for perianal CD [36]. This instrument simply defines response as a 50% decrease in draining fistulas and complete response as fistula closure or the absence of any draining fistulas. In that study, the Fistula Drainage Assessment was significantly lower in the treatment versus the control group, similar to the PDAI calculated simultaneously. These results have been replicated in other clinical trials [37–39]. Reliability of this subjective index was never formally assessed; therefore, it is recommended that two independent physicians score the index in future clinical trials. Moreover, in view of the questionable validity and responsiveness of these indices, and since no other similar measure is available, both should be calculated in clinical trials, until more data is available.

There is no unique PDAI for children with perianal CD. In order for the PDAI to be considered in children, the sexual dysfunction component should be modified to reflect more age appropriate issues such as participation in school and other social activities. Obviously, this modification requires appropriate validation and assessment of reliability and responsiveness before being used in pediatric studies. At present, the use of the Fistula Drainage Assessment definitions appears to be the preferred perianal outcome in pediatric clinical trials, as an anchor for physician assessment of disease activity.

The earliest classification of UC disease activity was a qualitative scale published by Truelove and Witts in 1955 (Appendix 3-1) [40]. This index defined remission, mild and severe disease. A score between mild and severe was considered as moderate. Because of this simple gradation and poor definition of moderate severity, significant ambiguity exists in defining change in disease activity with this index. Arbitrary quantitative indices have since been introduced, including the Powell-Tuck Index [41], the Mayo-Clinic Score [42], Rachmilewitz Index [43] and Lichtiger Score [44]. The first three include an endoscopic evaluation of the rectosigmoid as part of the global assessment. Their validation has therefore been largely a side product of clinical trials in which they have been used and developed. Seo and colleagues recently developed and evaluated an UC disease activity index [45, 46], weighted against the Truelove and Witts classification. The initial development process was biased towards severe disease since the investigators used a retrospective cohort of hospitalized patients. It is not surprising, therefore, that the index is heavily weighted toward three (of five) lab items which may be more important in severe disease. This index has not gained wide use despite its multi-step development. Walmsley and colleagues developed a Simple Clinical Colitis Activity Index that removed all laboratory parameters [47]. It correlated highly with both the Powell-Tuck Index and Seo index. However, this index is not commonly used, perhaps because it was aimed only for initial assessment by non-specialist physicians as a guide for the need to seek further medical advice. Recently, the Endoscopic Clinical Correlation Index (ECCI) was developed prospectively in 137 adults with UC undergoing full colonoscopy [48]. Items were chosen based on their ability to predict endoscopic outcome. The ECCI highly correlated with the endoscopy colitis score (r  =  0.81) slightly higher than the Seo, Truelove and Witts, Powell-Tuck and Walmsley’s simple colitis index. However, these correlations were obtained from the derivation cohort only and not from a separate validation cohort. Moreover, the reliability and responsiveness of the index were not assessed. In a prospective head-to-head study in adults of all non-invasive UC disease activity indices, the Walmsley index and PUCAI (see below) were best in assessing disease activity [49] when compared to a number of parameters including the full Mayo score and PGA. At this point in time, however, there is no prominent disease activity index, such as the CDAI for studies in adults with Crohn disease, for clinical trials in adults with ulcerative colitis (Appendix 3-2).

Unlike Crohn disease, UC has a more homogenous presentation and it is believed that most patients in remission have a normal, or near normal, endoscopic appearance. A position paper on the selection of outcomes in adult UC clinical trials recommended that the primary endpoint for therapeutic trials should be induction of remission (and exacerbation in maintenance studies), confirmed by an endoscopic assessment of the intestinal mucosa [50]. However, endoscopic assessment of disease activity has several major limitations. First, endoscopic scoring depends on the subjective assessment of mucosal properties with very low inter-observer reliability [51]. Second, endoscopic appearance lags after clinical improvement, thereby underestimating response to treatment [52]. Third, endoscopic remission usually starts in the ascending colon and progresses distally where the rectosigmoid is the last part to improve. Thus, limited sigmoidoscopy may not reflect response to treatment especially in children in whom extensive disease is the most common phenotype. Finally, although mucosal healing has been convincingly shown to predict favourable clinical outcome in UC [53–55], it is yet to be proven that endoscopy is superior to clinical judgement for assessing mucosal healing in general. Indeed, in the combined ACT cohorts (466 adults with UC treated with either infliximab or placebo), endoscopic healing had no predictive value among the subset of patients with clinical remission after 8 weeks of therapy [53].

Endoscopy in children requires sedation or general anaesthesia. Therefore, the need for repeat endoscopic assessment can deter children and their caregivers from participating in clinical trials. As it does not seem justified to recommend endoscopic assessment solely for research purposes, the PUCAI was developed omitting endoscopy variables (Appendix 3-3) [56]. The PUCAI was validated on a prospective cohort of 48 children undergoing full colonoscopy, wherein it showed very good to excellent correlation with PGA (r  =  0.91), colonoscopic appearance (r  =  0.77) and the adult reference index, the Mayo Clinic Score (r  =  0.95). Correlations were higher than two non-invasive adult indices, the Seo and Lichtiger indices, calculated concurrently. Inter-observer and test–retest reliability were excellent (ICC  =  0.95). The PUCAI differentiated well the categories of disease activity of none, mild, moderate and severe judged by a physician’s global assessment (areas under the ROC curves of >0.97). Responsiveness was shown to be high at repeated visits of 74 children. Remission was defined as a PUCAI score of <10, and the minimal clinically important difference as a change of at least 20 points (area under the ROC curve of 0.97). The laboratory items did not improve the validity or responsiveness of the PUCAI and thus were removed, making it attractive for pediatric use.

Since its introduction, the good performance of the PUCAI has been replicated in multiple independent cohorts [49, 57–59]. The cut-off values for remission and the different categories of disease activity have been replicated in two different studies [27, 49]. The PUCAI has been proven feasible on 215 children with UC from a prospectively maintained Pediatric registry wherein its predictive validity has been demonstrated by showing that children requiring escalation of medical therapy had higher PUCAI scores than those who did not. The correlation of the PUCAI with PGA in this cohort was 0.9. In two independent cohorts of children requiring admission for intravenous treatment of corticosteroids for UC exacerbations, the PUCAI has shown predictive validity, accurately identifying children who will require treatment escalation to second line medical therapy or colectomy, both by discharge and up to one year post discharge [57, 58]. In this setup, the PUCAI has shown to have superior predictive validity to five faecal biomarkers, including calprotectin [60, 61].

The PUCAI has since been reported in several pediatric clinical studies including two randomized controlled trials. In the T72 trial evaluating the effectiveness of infliximab in pediatric UC, the PUCAI determined week 8 remission rate was 33%, identical to the rate of complete mucosal healing found by sigmoidoscopy [62]. Similarly, the week 12 remission rate in a clinical trial evaluating beclomethasone 17,21-dipropionate in children with UC was similar whether determined by sigmoidoscopy or the PUCAI [63].

Both adults and children diagnosed with IBD are at increased risk of emotional problems and decreased social functioning [78, 79]. Thus, quality of life (QOL) assessment has been increasingly recognized as an important and independent clinical outcome in IBD research. There are two types of QOL instruments, generic and disease specific. The choice of instrument depends on the clinical question. Generic QOL instruments are preferred when comparing QOL in different diseases. However, for assessing only IBD populations, a disease specific QOL instrument should be used. A thorough discussion of QOL instruments available for pediatric IBD research is found in chapter 45.

Various instruments are available to measure clinical outcomes in pediatric inflammatory bowel disease (Table 41.2). Valid pediatric clinical indices and QOL measures exist for both UC and CD. Endoscopic scores, however, have not yet been evaluated in children. The evaluation of health related indices is an ongoing process and therefore, the development of new indices and the re-evaluation of the performance of existing indices will continue to be explored in different clinical and research settings.