A Fatality at the Other Site
A first-in-human Phase 1/2 CAR-T trial at a single academic site enrolls adults with relapsed or refractory B-cell lymphoma. Participant 014, aged 34, is recovering from Grade 3 cytokine release syndrome and Grade 3 immune effector cell-associated neurotoxicity on tocilizumab and dexamethasone. The site then learns that a participant at another site died of a similar event six days earlier, at what appears to be a higher frequency than the Investigator's Brochure lists. The investigator wants to hold the next two scheduled enrollments; the sponsor's medical monitor replies that the event is "expected." The approved consent form lists CRS and neurotoxicity as risks but not the apparent fatality rate. Participant 014's brother, a hospital employee in the research pharmacy, witnessed the consent because the participant reads at roughly a third-grade level.
1. Within what timeframes must the fatal event be reported, and by whom? Walk through the expectedness assessment step by step.
Read the worked answer
Two separate clocks run here.
Investigator to sponsor (24 hours). The second site's investigator must notify the sponsor within 24 hours of becoming aware, under the protocol's serious adverse event reporting window and ICH E2A conventions. Causality is not the trigger at this stage-seriousness is. Waiting for a causality assessment before the initial notification is a common and serious error.
Sponsor to regulators and investigators. The sponsor then makes three judgements in sequence:
- Seriousness - death; yes.
- Causality - reasonable possibility that the investigational product caused the event. For CAR-T with CRS and neurotoxicity this will almost certainly be "related."
- Expectedness - judged against the current Investigator's Brochure, and judged on nature, severity, and specificity, not nature alone. CRS and neurotoxicity are listed, so a naive reading says expected. But a fatal outcome, or an observed frequency materially above the rate stated in the Brochure, changes the severity/frequency profile and renders the event unexpected.
If fatal or life-threatening and unexpected, the sponsor reports to FDA and to all participating investigators within 7 calendar days, with follow-up within 8 additional days. Other SUSARs go within 15 calendar days. An unexpected fatal event also triggers an IB update and a consent form revision, which in turn triggers re-consent of currently enrolled participants.
2. The investigator wants to pause enrollment at their site; the sponsor wants to continue. Who has the authority, and what is the correct process?
Read the worked answer
Both have authority, but over different scopes, and the sponsor's label of "expected" does not settle the question.
The investigator may pause at their own site. ICH E6(R2) permits an investigator to deviate from the protocol without prior approval where necessary to eliminate an immediate hazard to trial participants, with prompt written notification to the sponsor and the IRB. A pause pending review is the conservative exercise of that judgement, not a violation.
The sponsor is the only party that can suspend the trial as a whole, except where a regulator or an IRB acts. The IRB holds independent authority to suspend or withdraw approval at its site and to demand information at continuing review.
Correct process:
- The investigator notifies sponsor and IRB in writing within the protocol-defined window.
- The sponsor convenes the DSMB with aggregate data from both sites: the fatality, exposure-adjusted event rates, CRS grading distribution, and any dose-level clustering.
- Any pre-specified stopping rule is applied. If none exists, its absence is itself a protocol deficiency to remediate before further first-in-human work.
- The DSMB recommendation is documented and the sponsor acts on it.
Continuing enrollment while a fatal event is unadjudicated is defensible only if an independent body has reviewed unblinded aggregate safety data and recommended continuation. "Expected" addresses reporting obligations; it does not address whether it is ethical to expose the next participant.
3. Is Participant 014's consent valid? What must happen now, and what systemic fix is needed?
Read the worked answer
Probably not valid as executed.
When a participant cannot read, ICH E6 and FDA guidance require an impartial witness present throughout the consent discussion who signs the form attesting that the information was accurately explained and apparently understood. "Impartial" has two limbs: independent of the trial, and free of any relationship that could pressure the participant. A sibling fails the first limb by family relationship and the second by employment in the hospital's research pharmacy-he is inside the research enterprise. Using him as witness defeats the purpose of the safeguard.
Remediation now:
- Re-consent using a genuinely independent witness-not a family member, not study staff-with the form read aloud in the participant's primary language.
- Because the neurotoxicity may have affected capacity, assess decision-making capacity before re-consent. If capacity is impaired, involve a legally authorized representative under applicable law.
- Document the deficiency and its correction in a note to file; do not overwrite the original record.
- Notify the sponsor and the IRB.
- Revise the consent form to include the new safety information, since re-consent is required for that separately.
Systemic fix: revise the site SOP to define witness eligibility explicitly, exclude employees of the institution and family members, and add a witness-eligibility check to the consent checklist audited at monitoring.