Trial abstract · §2
GSH-PD-PH2 — design and population
Design, allocation, arms and the population enrolled.
§2Design and population
§2.1Design and allocation
- Design class
- Randomised, double-blind, placebo-controlled, parallel-group
- Masking
- Double-blind (participant, investigator and sponsor)
- Arms
- 5
- Allocation
- Randomised across dose arms and a common comparator
- Endpoint adjudication
- Not applicable to the primary endpoint of this design
- Data monitoring
- As specified in the protocol
§2.2Arms
Table 2. Randomised arms. Illustrative: arm-level allocations are reconstructed by the Institute from the design class and the randomised total where the published report does not state them.
| Arm | Allocated | Share | Description |
|---|---|---|---|
| Glutathione, dose level 1 | 9 | 20.0 % | Randomised dose arm |
| Glutathione, dose level 2 | 10 | 22.2 % | Randomised dose arm |
| Glutathione, dose level 3 | 9 | 20.0 % | Randomised dose arm |
| Glutathione, dose level 4 | 9 | 20.0 % | Randomised dose arm |
| Placebo | 8 | 17.8 % | Matched placebo |
| Randomised total 45 as published. Arm-level splits are reconstructed and are not published figures. | |||
§2.3Population
Indication. Cognitive performance and neuroprotection
Domain-specific cognitive function or protection against cognitive decline, assessed by standardised neuropsychological batteries or clinical dementia rating.
Geographic footprint. United States · Italy · Poland.
§2.4Eligibility as the Institute reads it
- Included. Domain-specific cognitive function or protection against cognitive decline, assessed by standardised neuropsychological batteries or clinical dementia rating.
- Excluded. Participants for whom the intervention is contraindicated, including known hypersensitivity. Exclusion criteria narrow the population to which the estimate applies and are the principal source of indirectness where a trial estimate is applied to ordinary practice.
Compound Evidence Institute · CEI-TR-0128/2 · https://compoundevidence.com/trials/gsh-pd-ph2/design/ · retrieved 30 July 2026