In clinical practice we conventionally assess patients’ symptoms, but often this is done without using activity indices to guide management, perhaps with the exception of Truelove and Witts’ criteria to define acute severe colitis [32]. Many indices have arbitrarily assigned quantitative scores for improvement and lack rigorous design or evaluation but have been used in clinical trials [33]. Indices tend to have been designed for particular disease severities or purpose (such as the MTWSI for hospitalised patients with severe colitis, in contrast to the Mayo Clinic score for outpatients with mild or moderately active disease). Therefore, if applied to patients with a different pattern of disease, activity may be over- or underestimated, which causes further confusion when interpreting outcome data.

Partial Mayo Score. Six indices evaluate symptoms independently of endoscopic scoring or biochemical markers (Table 32.1). The Partial Mayo score evolved from the need to evaluate patients in clinical trials during the interval between endoscopies [34]. Although readily criticised because it is an unvalidated derivation of an unvalidated index, it serves a purpose when examining the speed of symptom relief or trends in response. When compared with other noninvasive indices, it performed well for discriminating remission from active disease and responsiveness [33] but depends on a Physician’s Global Assessment (PGA).

Simple Clinical Colitis Activity Index. The Simple Clinical Colitis Activity Index [14] was based on the Powell-Tuck Index, modified to exclude sigmoidoscopic assessment but to include nocturnal bowel movements and urgency of defecation. Urgency is a symptom of vital importance to patients, but neglected by other indices. The general well-being score was based on the Harvey-Bradshaw Index for Crohn’s disease [35]. The index was derived from a study of 57 patients with variable disease extent and severity. It included hospitalised patients. Scores range from 0 to 19 points, with generally applied thresholds shown in Table 32.2. It has been compared prospectively with the Partial Mayo score, Lichtiger (MTWSI), PUCAI, Rachmilewitz (CAI) and Seo indices in 86 adult patients [33]. Along with the PUCAI, it performed best of all noninvasive indices for validity, reliability, responsiveness and feasibility. Since it does not include a PGA, it can readily be completed by patients.

Modified Truelove and Witts Severity Index. The MTWSI (Lichtiger Index) was introduced during a pilot study of cyclosporine for acute severe colitis in 1990 [15]. It is important to remember that like its progenitor (Truelove & Witts’ Index), the focus was on patients with severe colitis and it may be less responsive or reliable for patients with less severe disease. The MTWSI has a score of 0–21 and comprises eight descriptors: number of daily stools, nocturnal stools, visible blood in stools, faecal incontinence, abdominal pain/cramping, general well-being and need for antidiarrhoeal agents. The authors arbitrarily defined “response” as a 50 % decrease in baseline score, and remission was subsequently defined as a score ≤3 [36] although this does not necessarily mean absence of symptoms. Although neither the score, thresholds nor response has been validated, it quantitates activity which the original Truelove & Witts’ Index fails to do, so it may yet be the most suitable index for trials on hospitalised patients with acute severe UC [37].

Ulcerative Colitis Clinical Score. The Ulcerative Colitis Clinical Score was designed for a placebo-controlled trial of an α(alpha)4β(beta)7 integrin antagonist for UC [16]. The UCCS is a modification of the Mayo Clinic score excluding endoscopy, so it is very similar to the Partial Mayo score. It comprises four descriptors: stool frequency, rectal bleeding, patient’s functional status and PGA. By including patient’s functional status in the index, it differs from the Partial Mayo score (which comprises the three other descriptors, even though functional status is separately scored). Although remission (score 0 or 1, as long as the endoscopy score was 0–1 on a modified Baron grading defined in the paper) and response (improvement by ≥3 points) were described, neither the score nor these thresholds have been validated. Nevertheless, it disarticulated symptom scoring from endoscopy, even though it recognised interdependence in the PGA. The Partial Mayo score has largely superseded the UCCS for clinical trials, since the full Mayo Clinic score is currently (2012) the most widely used index in trial design.

Pediatric Ulcerative Colitis Activity Index. The PUCAI was devised by paediatric gastroenterologists to provide a noninvasive instrument to assess disease activity in children in whom repeated endoscopy is less acceptable to patients and parents [2]. It comprises six descriptors with different levels, creating a total score ranging from 0 to 85: abdominal pain, rectal bleeding, average stool consistency, number of stool in 24 h, nocturnal stools and activity level. The PUCAI was rigorously developed using descriptor generation by a group of 36 experts and descriptor weighting by stepwise multiple regression analysis of prospectively collected data from 157 paediatric UC patients. Validation was assessed on a separate prospective cohort of 48 children with UC undergoing colonoscopy. Responsiveness was evaluated at follow-up visits in 75 children. It has predictive value for children admitted with acute severe colitis, for whom it has become the standard of care for evaluating activity and decision-making (see below). The PUCAI has also been shown to be valid, reliable and responsive in adults [33]. This may permit less frequent endoscopic assessment for patients with UC both in clinical practice and clinical trials [33].

Beattie and colleagues developed a disease activity instrument for children in a study published in 1996 [17]. This index generated a numerical score from 0 to 10 with four descriptors: stool frequency, rectal bleeding, abdominal pain and rectal prolapse. It has been evaluated in adults but did not perform as well as the PUCAI (which supersedes it in children) or the SCCAI [33].

Several disease activity instruments combine clinical symptoms, endoscopy and quality of life descriptors into a composite index. This is superficially appealing, because the physician in clinical practice considers all aspects and makes a judgement. Nevertheless, the subjectivity of that judgement creates concern for consistency, and validating a composite index creates particular difficulty, where interaction between descriptors has to be evaluated. It is easier to separately validate the symptomatic, endoscopic, quality of life and (if appropriate) histological components. Indeed, since independent indices have been validated for all aspects bar symptoms, a composite index appears to swim against the tide. On the other hand, the archetypal composite index, the Mayo Clinic score, is the index most widely used in clinical trials. This has an inherent value, since it allows the efficacy of different trials to be compared, assuming the same endpoints and definitions of response [38, 39]. Nevertheless, not many gastroenterologists routinely use the Mayo Clinic score in clinical practice.

Mayo Clinic Score. The Disease Activity Index (DAI) was first described in 1987 for a placebo-controlled trial of mesalamine for active UC [18]. There are four descriptors: stool frequency, rectal bleeding, findings at proctosigmoidoscopy and PGA. The stool frequency score is not an absolute number, but relative to “normal” for that subject, which may itself introduce variation between observers that has yet to be quantified. Symptoms are assessed over 3 days: some clinical trialists take the average of symptom scores, others the worst score in the 3 days preceding the visit. Scores range from 0 to 12 points. The physician has access to the patient’s functional assessment as a measure of general well-being when determining the PGA, but the patient’s functional assessment is not used to calculate the score. Definitions of remission and improvement depend on descriptor subscores. Complete resolution is defined as a DAI of 0 (normal stool frequency, no rectal bleeding, normal proctosigmoidoscopy and a PGA of 0). Response has been defined as improvement in the PGA and at least one other item subscore and no worsening in any other descriptor subscore, although definitions of response vary [1]. A more liberal definition of remission (≤2) has been recognised by the FDA for registration trials of infliximab in the treatment of UC [34]. A trial endpoint is more easily reached with a lower threshold (i.e. a higher score) for remission. This is apparent from the ACT (Active Ulcerative Colitis Trials [ACT] I and II) of infliximab for patients with UC refractory to standard therapy. The definition of remission was a DAI ≤2, with no individual subscore >1 [34]. When this definition was applied to a population of patients without treatment-refractory disease in a retrospective analysis of two large trials of mesalazine [40, 41], the remission rate for 2.4 g mesalazine increased from 22 % (according to the original trial definition) to 50 % [42]. The Mayo Clinic score is the activity index for UC against which others have to be compared; at present (2012) the advantages of common usage outweigh its inherent disadvantages.

Sutherland Index. The UC Disease Activity Index (UCDAI) was originally used in a placebo-controlled trial of mesalamine enemas for the treatment of distal UC [19]. This is a notably different patient population to the MTWSI (above). It is a simplified composite index, quite similar to the Mayo Clinic score, incorporating four descriptors: stool frequency relative to normal, rectal bleeding, endoscopic mucosal appearance and the PGA. Scores range from 0 to 12 points. Subsequent studies defined remission as a DAI of 1, with a score of 0 for rectal bleeding and 0 for stool frequency and at least a 1 point reduction from baseline in sigmoidoscopy score with friability moved from a score of 1 to two within the sigmoidoscopy score making a more stringent definition of remission [1]. The Sutherland Index has not been formally validated. The relative simplicity of the index provides a means of reducing the impact of physician and patient subjectivity in disease scoring. The index has been adopted in large clinical studies [43, 44]. Of particular note, a score <2.5 points has been shown to correlate with patient-defined remission [45], indicating that sigmoidoscopy contributes little to the definition of remission in clinical practice.

Powell-Tuck Index. The PTI was originally developed at St Mark’s Hospital, London, when comparing dosing schedules of oral prednisolone for the treatment of active UC [20]. The index scores from 0 to 20 points and includes ten descriptors: general well-being, abdominal pain, bowel frequency, stool consistency, bleeding, nausea/vomiting, anorexia, abdominal tenderness, extra-intestinal manifestations and fever. Later studies used a variation of the PTI which added an extra two possible points by including a sigmoidoscopy assessment score. Remission was defined as a score of 0 and improvement as a decrease in the baseline score 2 or more points. Neither the PTI nor the definitions of remission or improvement have been validated, but the index was the basis for developing the SCCAI (above). The relative complexity reflects the lack of validation, which would have identified redundant descriptors. It is unlikely to have a role in future clinical trials and is impracticable for everyday practice.

Rachmilewitz Index. The Clinical Activity Index (CAI) was originally used in a controlled trial of coated mesalamine compared to sulfasalazine for the treatment of active UC. This index generates a score from 0 to 29 points based on seven descriptors: number of stools weekly, presence of blood, investigators global assessment of symptomatic state, abdominal pain/cramps, temperature, extra-intestinal manifestations and laboratory findings (ESR and haemoglobin). It continues to be used in some clinical trials [46]. Like many indices, the CAI considers an “investigator’s global assessment” to be an essential component. This, however, introduces a layer of subjectivity, depending on the amount and quality of time spent with the patient. Similarly, the endoscopic element of the CAI depends on the subjective assessment of mucosal properties, including friability. Its main weakness is that clinical remission has come to be defined as any score less than that used to define disease activity (CAI score >4). By this measure, a score ≤4 points includes a level of symptoms that cannot conceivably be used to define remission: this might mean, for instance, that a patient with 36–60 stools/week (score 2) and a little blood in the stools (score 2, total = 4) met the criteria for “remission”! It fails to recognise the simple fact that there is a “grey area” in scoring systems between the level used to define remission and the threshold used for defining disease activity. In the prospective comparison of different disease activity indices, a score ≤2 best reflected remission [33].

Seo Index. The Activity Index (a term best avoided, since it is so readily confused with CAI, DAI or UCDAI) was devised using multivariate regression analysis, similar to that used to develop the Crohn’s Disease Activity Index (CDAI) [22]. 18 clinical, laboratory and sigmoidoscopy variables were initially derived from prospective data collected from 72 patients during 85 clinical events. The Seo needs a calculator: 60 × blood in the stool + 13 × bowel movements + 0.5 × ESR − 4 × haemoglobin (in g/dL) − 15 × albumin + 200. When correlated against the MTWSI as a standard, subjects in remission had a mean Seo of 100 ± 11: 90 % those with mild disease had scores <150; 83 % those with moderate activity had scores between 150 and 220 and those with severe disease on average scored above 220. The definition of remission was therefore set at 120, but a prospective comparison has shown that a score of 108 is more appropriate [33]. This comparison, however, showed that the Seo failed one of the four fundamental psychometric criteria for indices: that of feasibility. It also performed less well than others with regard to discriminative ability, test-retest reliability and responsiveness. Nevertheless, a subsequent study in patients with moderate to severe UC suggested that the Seo Index might have predictive value: following 2 weeks treatment with intravenous corticosteroids, 65 % of subjects with a score >180 underwent colectomy [47], and following infliximab treatment, the index predicted response to therapy or need for colectomy [48]. Indices such as the Seo use multiple biomarkers and may yet identify patients who meet the “regulatory definition” of remission, without the need for endoscopy (i.e. patients with no more than grade I or II on a modified Baron endoscopic score), and the absence of visible blood are identified using a cut-off score of <120 [45]. The associations are relatively weak (around 60 %) and the index is too complex to use in practice.

Truelove and Witts. The first instrument to assess disease activity in UC was devised in 1955 by Truelove and Witts in the first clinical trial in gastroenterology, evaluating cortisone treatment for UC [23]. This index has five descriptors: bloody stool frequency, temperature, heart rate, haemoglobin and erythrocyte sedimentation rate (ESR). The original instrument provided definitions for mild and severe UC, with all cases in-between classified as moderate. This apparent lack of precision is off-set by the objective criteria for defining (acute) severe colitis, which are widely used to define a course of action (hospital admission for intravenous therapy, in 27/32 trials of steroids for UC) [49] and which predict outcome [50]. Its principal disadvantage is that it does not generate a quantitative activity score, which makes it unsuitable for assessing outcomes in clinical trials. Instead, ambiguous terms for evaluating response were originally proposed (“improved,” “no change” or “worse”), which have no place in clinical trials today [1]. On the other hand, the T&W Index is the most amenable for daily clinical practice by defining acute severe colitis and continues to be used as an entry or exclusion criteria for clinical trials (e.g. CONSTRUCT ISRCTN22663589) and is almost universally recommended by national or international guidelines for the management of acute severe colitis [51–53].

Montréal Classification. The Montréal Classification was developed by an international working group for the World Congress of Gastoenterology in 2006 [54]. It incorporates both extent of disease, divided into proctitis (distal to the rectosigmoid junction = E1), left sided (distal to the splenic flexure = E2) and pancolitis (proximal to splenic flexure = E2) and severity. The severity scores are based on the Truelove and Witts Index, with S0 = remission [54]. Mild activity (S1) was defined as 4 or fewer bloody stools/day without signs of systemic toxicity and with a normal ESR. Moderate activity (S2) was defined as more than 4 bloody stools/day, with minimal systemic toxicity. Severe colitis was defined in the classical T&W description, with 6 or more bloody stools/day with a pulse >90 bpm, temperature >37.8 °C, haemoglobin <10.5 g/dL or ESR >30 mm/h. The index is pragmatic and was meant to be applied in clinical practice, especially with regard to large clinical studies (such as those involving genetics or disease databases), but although responsiveness has never formally been tested, it appears too insensitive for use in clinical trials—other than as a threshold for defining acute severe colitis.

Endoscopic indices evolved from the Baron score, initially developed for rigid proctoscopy in ambulatory patients with mild to moderate disease, which rated mucosal bleeding and friability [25]. Subsequent endoscopic indices were of increasing complexity and incorporated the presence of ulcers, mucopus, granularity or light scattering in addition to bleeding and friability [18, 19, 54, 55]. Such modifications were intended to improve the capture of disease activity, but they invariably increased the subjectivity of the scoring system.

In clinical practice, endoscopic assessment of disease activity is used to confirm diagnosis and assess disease activity. Endoscopic confirmation of disease improvement is uncommon in clinical practice if the patient’s symptoms have resolved. In the context of clinical trials, endoscopic assessment has a key role for measuring outcome, because it is intended to be independent of symptom score. The newly introduced term of “mucosal healing” has shown that where treatment achieves this within 8 weeks, this correlates with a lower colectomy rate over the succeeding 12 months (p = 0.0004) and steroid-free remission (p < 0.0001) [56]. In composite indices its relative weighting varies. The FDA currently uses two measures for defining remission: endoscopic mucosal healing and rectal bleeding. Although considerable efforts have been made to derive a patient symptom score that negates the need for endoscopy [45] and the endoscopic component of some, symptom scores (such as the UCDAI) have been calculated to contribute only 2.5 % of the total score [45]. There has been renewed interest in endoscopy with the advent of a validated endoscopic scoring system [3]. Efforts have been made to compare indices, which have resulted in another activity index that shows good interobserver agreement ((kappa)κ = 0.65–0.79 between 4 expert endoscopists) but has not followed the criteria for index development (Table 32.3) [57].