The TITRATE study enrolled 25 patients on standard IFX dosing (5 mg/kg at weeks 0, 2, and 6) and 23 patients on personalised IFX dosing using iDose. Precision dosing in patients with acute severe ulcerative colitis guided by iDose’s Bayesian algorithm achieved a composite response rate of 78% (18/23 patients) compared to just 36% (9/25 patients) with standard dosing as measured by AI-based endoscopy scoring (p=0.0047), according to results from the new TITRATE study.
This marks the first IBD study where AI assessment overturned expert evaluation to meet the primary endpoint, demonstrating both the efficacy of pharmacokinetic-guided dosing and the critical role of measurement precision in detecting treatment differences.
Statistically Superior Outcomes Across Multiple Endpoints
The prospective, multicenter trial randomized 48 infliximab-naive, steroid-refractory ASUC patients 1:1 to personalized dosing using iDose or standard dosing protocols. Personalized dosing targeted infliximab concentrations above 28 µg/mL during days 0-28 and above 15 µg/mL during days 29-42, with additional 5 mg/kg infusions administered based on the algorithm’s predictions.
Clinical outcomes diverged substantially. Clinical response at day 42 reached 91% in the personalized arm versus 64% in the standard arm (p=0.039). By 6 months, endoscopic remission measured by AI was 52% versus 20% (p=0.0337). The proportion of patients achieving combined clinical and endoscopic remission at 6 months was 65% versus 36% (p=0.082) based on the AI assessments.
Notably, serious adverse events were less frequent with personalized dosing: 9% versus 20% in the standard arm, suggesting that achieving appropriate drug exposure through therapeutic drug monitoring improves both efficacy and safety.
The Measurement Question
Expert endoscopists reading trial videos found no significant difference between treatment arms for the primary composite endpoint: 57% response in personalized dosing versus 44% in standard dosing (p=0.564). However, AI-based assessment using the DovaVision UC tool revealed the 74% versus 32% difference (p=0.0047) that met the study’s primary endpoint.
This discordance raises fundamental questions about measurement precision in clinical trials. If AI can detect treatment responses that expert readers miss, particularly in diseases where endoscopic evaluation drives treatment decisions, the implications extend beyond this single study to clinical trial methodology itself.
Efficient Drug Utilization
The median cumulative infliximab dose through day 42 was 18.41 mg/kg for personalized dosing versus 13.79 mg/kg for standard dosing. This modest increase of approximately 33% in total drug exposure translated into dramatic differences in outcomes, demonstrating that precision dosing isn’t about administering more drug blindly but about achieving the right drug exposure at the right time based on individual pharmacokinetics.
Why TITRATE Succeeded Where PREDICT-UC Failed
The TITRATE results stand in sharp contrast to those of the PREDICT-UC trial, published in The Lancet Gastroenterology & Hepatology in 2024. PREDICT-UC tested two empirical dose intensification strategies (accelerated induction strategy (AIS) and intensified induction strategy (IIS) together with the standard induction strategy (SIS) in 138 ASUC patients and found no significant benefit. The assignment was randomized and not reflective of the patients’ individual characteristics.
For the initial dose, patients were randomized to receive an initial dose of 5 or 10 mg/kg IFX. Following this, patients were randomized into one of the three doses with the potential for an additional dose administration. The primary outcome was clinical response by day 7 (reduction in Lichtiger score to <10 with a decrease of ≥3 points from baseline, improvement in rectal bleeding, and decreased stool frequency to ≤4 per day).
Patients randomized to an initial 10 mg/kg infliximab dose showed no improvement in clinical response at day 7 compared to 5 mg/kg: 30 [65%] of 46 vs 56 [61%] of 92 (p=0.62). The three different induction strategies (intensified, accelerated, and standard) produced similar outcomes at 3 months. Clinical remission rates were 23/46 (50%), 25/48 (52%), and 21/44 (48%), respectively (p=0.92).
Even in patients with high inflammatory burden markers like low albumin (below 25 g/L) and elevated CRP (50 mg/L or higher), the 10 mg/kg dose showed only numerically higher response rates without statistical significance. AI was not utilized for these assessments.
The Critical Distinction
PREDICT-UC delivered higher total drug exposure without pharmacokinetic targets, providing up to 20 mg/kg over 6 weeks. TITRATE delivered individualized exposure based on predicted concentrations and patient-specific clearance. However, this is fairly similar to the drug exposure in TITRATE at 6 weeks. The key difference is that the timing of the administered dose was determined by iDose in TITRATE based on individual IFX trough concentrations.
Clinical Context
Up to 40% of ASUC patients fail to respond to infliximab, presumably due to insufficient drug exposure. Hypoalbuminemia and high inflammatory burden correlate with increased infliximab clearance. The question for clinicians has been whether these patients need more drug or whether they need individualized dosing to achieve therapeutic exposure.
PREDICT-UC demonstrated that empirical escalation does not provide a systematic advantage. TITRATE demonstrates that pharmacokinetic-guided escalation does.
Implications for Practice
For the approximately one-third of ASUC patients who fail standard therapy, these data indicate treatment failure reflects inadequate exposure in specific patients rather than a universal need for higher doses. Without pharmacokinetic guidance, clinicians cannot identify which patients require intensification or distinguish between inadequate drug exposure and non-TNF-mediated inflammation.
The TITRATE investigators noted this as the first IBD study where precision dosing required precision measurement to demonstrate its full impact. The combination of individualized dosing algorithms and AI-based assessment may fundamentally improve both treatment strategies and clinical trial methodology.
Dashboard-driven therapeutic drug monitoring using tools like iDose provides adequate information to achieve target drug concentrations from the start of therapy, potentially preventing the treatment failures that lead to colectomy in this critically ill patient population.
Experience Precision Dosing in Your Practice
The TITRATE study demonstrates what dashboard-driven therapeutic drug monitoring can achieve in one of gastroenterology’s most challenging scenarios. To see how iDose’s Bayesian algorithm can optimize infliximab therapy for your ASUC patients and broader IBD population, schedule a personalized demo with Baysient’s Chief Science Officer, Dr. Diane Mould.
Discover how pharmacokinetic-guided dosing translates into better patient outcomes and more efficient biologic utilization in your clinical practice.
