How to use it
The patient keeps a 7-day diary: each day’s number of liquid or very soft stools, abdominal pain (0 none, 1 mild, 2 moderate, 3 severe) and general well-being (0 generally well to 4 terrible). Enter it day by day, or switch to totals or averages if the diary is already summarised. Then add complications, antidiarrhoeal use, abdominal mass, hematocrit and weight.
Hematocrit is entered in L/L (as Canadian labs report it) or %. A value that doesn’t fit the chosen unit (for example 38 in L/L) is flagged rather than converted. Missing diary days: with 4–6 days entered, the tool averages them and multiplies by 7, the rule used in trial protocols. Fewer than 4 days is treated as incomplete.
Reading the result
| CDAI | Usual label | Source |
|---|---|---|
| Below 150 | Remission | Best 1976; ACG 2018 |
| 150–219 | Mild to moderate | ACG 2018 (150–220) |
| 220–450 | Moderate to severe | ACG 2018; usual trial entry range |
| Above 450 | Severe to fulminant | Best 1976; ACG 2018 |
- ACG gives the mild-to-moderate band as 150–220 and the moderate-to-severe band as 220–450. Trials enrol from 220, so the tool puts 220 in the moderate-to-severe band.
- Response is a fall of 100 points or more (CR-100, used in UNITI and the upadacitinib trials) or, in older trials, 70 or more (CR-70, e.g. CHARM). UNITI also counted a CDAI below 150 as a response.
- Patients who need corticosteroids to stay well are steroid-dependent, not in remission (ACG 2018).
PRO2: what current trials use
Regulators now want patient-reported outcomes plus endoscopy rather than the CDAI alone (ACG 2018). PRO2 uses just two diary items, so the tool calculates it from the same entries.
- Unweighted PRO2 (the trial endpoint): the daily averages of stool frequency (SF) and abdominal pain (AP). Clinical remission in the upadacitinib and risankizumab programmes is SF ≤2.8 and AP ≤1.0, neither worse than baseline. STRIDE-II and the mirikizumab trials use SF ≤3 and AP ≤1. Response per PROs (upadacitinib) is a fall of 30% or more in SF and/or AP, with neither worse than baseline. STRIDE-II defines clinical response as a fall of 50% or more in PRO2.
- Weighted PRO2 (Khanna 2015): 2 × average SF + 5 × average AP, the CDAI weights applied to daily means. Values of 8, 14 and 34 correspond to CDAI 150, 220 and 450. Some papers instead report the weighted 7-day sums, which are 7 times larger (and equal the stool and pain points of the CDAI). Check which form a paper uses before comparing numbers.
In the upadacitinib trials, CDAI remission was the co-primary endpoint for the FDA and PRO2 remission for the EMA. Expect to see both.
Conventions that change the number
The 1976 formula leaves some details open, and current trial protocols fill them in differently. The tool uses the most common choices and says so in the result and the note:
- Hematocrit term set to 0 when hematocrit is at or above the reference (47% men, 42% women). Upadacitinib, ontamalimab and brazikumab protocols do this.
- Weight term no lower than −10 (ontamalimab, CT-P13 SC and brazikumab protocols). The upadacitinib protocol sets it to 0 instead, and the brazikumab protocol also caps it at +10. When a limit changes the total, the tool also shows the unlimited arithmetic.
- Standard weight comes from Metropolitan Life tables in some protocols, an ideal-body-weight formula in others, or height² × 22.1 (men) / 20.8 (women). Pick one and keep it the same for each patient. The difference is usually a few points.
- Total never below 0.
For a trial patient, use that protocol’s rules. The tool’s convention is fine for following one patient in clinic.
Pitfalls
- It measures symptoms, not inflammation. Stool frequency and PRO2 correlate only moderately with SES-CD (R = 0.36), and less in ileal disease (Lewis 2020). Confirm activity with calprotectin, CRP, endoscopy or intestinal ultrasound, as STRIDE-II advises.
- Overlapping causes of symptoms such as IBS, bile-acid diarrhoea, SIBO after resection, infection and strictures raise the score without active inflammation (ACG 2018).
- The weight term. The index was built when undernutrition was common. It rewards weight above standard, so in obesity it can lower the CDAI without any change in symptoms (Treton 2026). The −10 limit softens this, which is one reason protocols use it. Standard weight is defined differently across protocols.
- Subjectivity and observer variation. Well-being and pain are self-rated. Abdominal mass and complications depend on the examiner, and experienced clinicians scoring the same patients have produced very different totals (Freeman 2008).
- Perianal disease is under-weighted. A fistula or abscess adds 20 points whatever its severity. Use the PDAI.
- Stoma, short bowel or prior resection raise the stool count regardless of activity.
CDAI and HBI
The Harvey-Bradshaw Index is a one-day simplification without the diary, laboratory test or weight. It is easier in clinic. In the PRECiSE trials an HBI of 4 or less corresponded to a CDAI of 150 or less, and a 3-point HBI change to a 100-point CDAI change (Vermeire 2010). STRIDE-II accepts HBI below 5 or PRO2 for clinical remission. Use the CDAI when you need trial-comparable numbers, and record Montreal phenotype alongside it.