Executive summary
PRISE-hATG plans a 108-participant efficacy dataset: about 72 participants recruited prospectively and 36 external participants borrowed from the SAFEGUARD study. This assessment asks whether the 72 prospective participants can plausibly be recruited, using public information only.
- Overall. Potentially feasible, but not validated at the site or patient level.
- Target population. The four-state baseline implies roughly 4,000 modeled new diagnoses a year under age 20. That describes disease burden. It is not a count of eligible patients, and it does not cover the full 5–40 age range.
- Recruitment environment. Five registered studies have potentially overlapping populations. Registry overlap does not prove competition for the same patients.
- Timeline. Enrolling 72 participants at a sustained two to six per month would take 12–36 months. The rates are planning assumptions, not forecasts.
- Key uncertainties. Eligible-patient volume, screening yield, site capacity, and competition at each site.
- Next step. A bottom-up, site-by-site forecast built from real screening and referral data.
How to read this report. Each figure is labeled as evidence-supported, calculated from stated inputs, or an analyst-selected assumption. Treat the three differently.
Study overview
PRISE-hATG (NCT07670650) is a Phase 3, multicenter, randomized, double-blind, placebo-controlled study of SAB-142 in people aged 5–40 with Stage 3 type 1 diabetes (T1D). Prospective participants are diagnosed more than 100 days and up to two years before enrollment and randomized 2:1 to SAB-142 or placebo. The external cohort comprises SAFEGUARD participants (NCT07187531) diagnosed within 100 days. Treatment is given at baseline and Month 6, with follow-up through Month 12.
The primary endpoint is change from baseline in log-transformed C-peptide AUC+1 during a two-hour mixed-meal tolerance test. The random C-peptide used for screening is a separate entry criterion and should not be confused with it.
What the external cohort does and does not do. It reduces how many participants PRISE must recruit. It does not show that the remaining 72 are achievable, and it makes baseline and assessment comparability important for the combined analysis.
The age range spans pediatric, adolescent, and adult care, which affects referral pathways, consent and assent, investigator requirements, and logistics.
Target-population assessment
Geographic baseline
The baseline covers California, Colorado, Florida, and Indiana, the states of the PRISE-hATG locations in the public record. Inputs are 2024 U.S. Census population estimates and published state-level T1D incidence rates.
| State | Population under 20 | Modeled new T1D diagnoses a year, under 20 |
|---|---|---|
| California | 9,333,226 | 1,923 |
| Colorado | 1,384,000 | 334 |
| Florida | 4,999,479 | 1,265 |
| Indiana | 1,781,214 | 474 |
| Combined | 17,497,919 | 3,996 |
Diagnoses are modeled by applying state incidence rates to population denominators. They are not observed counts.
From population to eligible patients
The incidence rates apply only to people under 20, while the study spans 5–40. Census counts can supply denominators for ages 5–19 and 20–39, but adult diagnoses need adult age-specific incidence rates. Three populations need separate estimates:
- Newly diagnosed people who may fall in the recent-onset window (incidence-based).
- Existing T1D population in the longer duration window (prevalence-based).
- Potentially eligible people who also meet C-peptide and other clinical criteria.
Without patient or site-level data, eligibility is shown as labeled scenarios. For a hypothetical population of 1,000, assumed eligibility of 5%, 10%, and 20% gives 50, 100, and 200 people. This demonstrates the method only. The percentages are not published PRISE screening rates.
Not accounted for: residual C-peptide status, other eligibility criteria, referral patterns, catchment areas, competing studies, willingness to participate, and site screening capacity.
Recruitment environment
Residual C-peptide narrows the pool
TrialNet analyses show wide variation after diagnosis. In one analysis of 407 participants from five intervention studies, 93% still had detectable C-peptide two years after diagnosis, with faster average decline in year one than year two. In a separate analysis of participants selected for stimulated C-peptide of at least 0.2 nmol/L, the share maintaining baseline secretion fell over four years. Maintaining baseline is not the same as staying above a trial threshold. Gitelman and colleagues found beta-cell glucose sensitivity declined faster in children than in adolescents and adults. Age and disease duration should be considered together.
Referral pathways and screening barriers
Likely pathways include pediatric and adult endocrinology clinics, diabetes centers, hospital research units, and clinical-research networks. Because recent-onset windows are short, timely referral after diagnosis matters. Kinney and colleagues reported screening uptake of roughly 9% to 50% across five TrialNet prevention trials, and higher-volume sites screened more. Those are presymptomatic prevention trials, so the figures should not be applied to PRISE. Participant burden (laboratory screening, MMTTs, repeat visits, travel, caregiver involvement) is also unquantified.
Potentially overlapping studies
| Study | Population and criteria | Relevance to PRISE-hATG |
|---|---|---|
| SAFEGUARD NCT07187531 | Recent-onset Stage 3 T1D; SAB-142 | Related study supplying the external cohort; not simply a competitor |
| CNP-103 NCT06783309 | Ages 12–35; recent diagnosis; residual C-peptide required | Possible overlap in age, duration, eligibility |
| βETA PRESERVE NCT07088068 | Recent Stage 3 T1D; ages 1–25 | Possible overlap among younger participants |
| BARICADE-PRESERVE NCT07222332 | New T1D; ages 1–35; residual beta-cell function | Possible overlap in age and objective |
| WAVE T1D NCT07061574 | Ages 9 to under 21; recent-onset Stage 3 T1D | Possible overlap among pediatric and adolescent participants |
This is a targeted sample, not an exhaustive landscape. A registered study is not necessarily recruiting, and overlap in disease area does not prove two studies draw on the same patients. Verify status, criteria, and locations before operational use.
Sites
The four locations in the public review are UCSF Benioff Children’s Hospital, the Barbara Davis Center, the University of Florida in Gainesville, and Riley Hospital for Children. Listing does not establish activation, patient volume, or capacity. The CNP-103 registry reports locations in the same four states, which warrants a closer look without assuming it already affects recruitment.
Enrollment and timeline scenarios
The modeling target is the 72 prospective participants. Site-level targets, activation dates, historical yields, and monthly recruitment are unverified.
Recruitment rate
| Scenario | Per month, all sites | Per site per month* | Months to enroll 72 |
|---|---|---|---|
| Conservative | 2 | 0.50 | 36 |
| Base case | 3 | 0.75 | 24 |
| Accelerated | 4 | 1.00 | 18 |
| High-performance | 6 | 1.50 | 12 |
*Assumes four sites contribute equally. Real recruitment will be uneven and the model ignores activation delays, ramp-up, pauses, and seasonality. The base case is a planning reference, not a forecast.
Screening burden
| Assumed yield | Candidates screened for 72 randomizations | Not randomized |
|---|---|---|
| 50% | 144 | 72 |
| 60% | 120 | 48 |
| 70% | 103 | 31 |
The yields are analyst-selected assumptions, not published PRISE rates, and should not be derived from prevention-trial uptake. A full funnel separates patients identified, referred, approached, consented to screening, passing screening, and randomized, because each loss needs a different response.
Timeline risks
| Risk | Potential effect | Data needed |
|---|---|---|
| Site startup delays | Later recruitment start | Activation dates and startup milestones |
| Too few eligible referrals | Lower recruitment rate | New-diagnosis volumes and referral pathways |
| C-peptide or other eligibility failures | More candidates per randomization | Screening logs and failure reasons |
| Competing trials | Fewer available referrals | Current site-level competitor status |
| Unequal site productivity | Aggregate enrollment below plan | Monthly screening and randomization by site |
| Participant or caregiver burden | Delayed screening or withdrawal | Visit schedule and access information |
Estimated study dates in the registry should be reconfirmed before use. An estimated start is not a first-participant date, and primary completion is not the final randomization date.
Findings and open questions
Preliminary judgment: potentially feasible, not yet validated. The evidence supports a plausible recruitment pathway. It cannot show that the target is achievable at these sites or within the planned timeline.
Five findings are supported by the evidence:
- The eligible population is narrower than the general T1D population.
- Residual beta-cell function varies by age and disease duration.
- Screening performance depends on referral and site processes.
- Overlapping studies warrant site-level review.
- Enrollment timelines remain scenario-based.
| Open question | Data needed |
|---|---|
| Is the eligible population large enough? | Age-specific incidence and prevalence, duration distribution, eligibility estimates |
| Can sites identify patients early enough? | Monthly diagnoses, referral delays, referrals within the windows |
| What share will pass screening? | Historical screening logs, C-peptide distributions, failure reasons |
| Can sites hit the required rate? | Site forecasts, staffing, activation dates, historical randomization rates |
| How much will competing studies matter? | Recruitment status, site overlap, referral information |
| Is the timeline realistic? | Confirmed milestones, site forecasts, endpoint follow-up requirements |
Assumptions to validate: a 72-person prospective target separate from the external cohort; relevance of the four U.S. locations; the illustrative recruitment rates and screening yields; some overlap effect from other studies; and continued relevance of the public timeline. They are acceptable in a sample when labeled, and should be replaced with verified data before any decision.
Recommendations
Actions are ordered to reduce uncertainty before an operational decision.
1. Validate the eligible patient pool
- Request age- and duration-stratified estimates from sites, separating recent-onset candidates, longer-duration patients, and those unlikely to meet all requirements.
- Decision enabled: whether the accessible pool can support the target.
2. Obtain historical screening and enrollment data
- Request de-identified aggregates from comparable T1D studies: referrals, screening consent, screen failures by reason, randomizations per month, referral-to-randomization time. Where data are missing, ask for documented estimates and their basis.
- Decision enabled: replacing assumed yields with defensible site-level estimates.
3. Assess competition at each site
- Verify status, locations, and criteria of overlapping studies, and whether they run at the same institutions or referral networks.
- Decision enabled: whether to add sites or referral partners.
4. Build a bottom-up recruitment forecast
- Model each site with activation dates, ramp-up, referrals, conversion, capacity, and competing commitments. Produce conservative, base, and accelerated cases and test sensitivity to yield and monthly rate.
- Decision enabled: whether the plan meets required milestones.
5. Strengthen referral and access pathways
- Map relationships with endocrinology practices and research programs. Review diagnosis-to-referral delay, travel, caregiver logistics, and visit burden, staying within the protocol, consent requirements, and clinical oversight.
- Decision enabled: practical gains in identification and screening throughput.
6. Keep an evidence-update process
- Reconfirm the protocol, criteria, site status, timeline, and competitors before finalizing. Log each assumption with its source, confidence, owner, and next validation step.
- Decision enabled: a traceable assessment that stays current.
Decision gate. A confident forecast needs site-level eligible-population estimates, defensible screening-to-randomization assumptions, confirmed activation plans and competitor information, and a bottom-up model that meets milestones in the base case with acceptable contingency in the conservative case. If the conservative case misses the timeline, evaluate more sites, stronger referral pathways, or a revised schedule rather than relying on accelerated assumptions.
Methodology and limits
This sample combines published literature, public registry information, Census estimates, calculated epidemiological estimates, and labeled scenarios. It does not reproduce any sponsor’s feasibility assessment and has no access to the full protocol, patient-level data, screening logs, or confidential forecasts. Its purpose is to show what the evidence supports, make uncertainty explicit, and specify the data needed for a defensible recruitment decision.