NCT02424708 was a small phase 2b study comparing two intranasal reduced-glutathione exposures with saline in 45 people with Parkinson disease. All three cohorts improved on total Unified Parkinson’s Disease Rating Scale scores during the 12-week intervention, and neither glutathione group was superior to placebo on the primary clinical comparison. A nominally significant within-group change in the higher-exposure arm does not overturn that null between-group finding. The study also had five intervention-period withdrawals, mixed or null secondary results, a one-region all-White sample, and limited power. It supports a careful record of one formulation and protocol, not a general efficacy conclusion or equivalence to another glutathione form.[1][2][3][4]
Study at a glance
| Field | Study record |
|---|---|
| Research question | Over 12 weeks, did either of two intranasal reduced-glutathione study exposures improve Parkinson outcomes more than saline? |
| Design | Phase 2b, randomized, double-blind, saline-controlled, three-arm study in one US region |
| Population | 45 people with Hoehn and Yahr stage 1–3 Parkinson disease; 15 assigned per arm; all participants were White |
| Intervention | Historical identifiers: intranasal reduced glutathione at 100 or 200 mg/mL, 1 mL three times daily, corresponding to nominal 300- or 600-mg/day study exposures |
| Comparator | Intranasal saline under the same 12-week protocol |
| Follow-up | 12-week intervention followed by four weeks of washout observation |
| Primary clinical measure | Change in total and motor Unified Parkinson’s Disease Rating Scale scores; direct active-versus-placebo evidence remained null |
| Registration | NCT02424708; completed; enrollment 45; a results submission was returned after quality-control review and no structured results were posted |
| Publication | Journal of Parkinson’s Disease 2017; PMID 28436395; PMCID PMC5438472; DOI 10.3233/JPD-161040 |
| Sponsor/funding | Bastyr University; Michael J. Fox Foundation funding; founder interest and supplied device tips disclosed |
The recorded concentration, amount, route, frequency, and duration identify the historical study. They are not preparation, dosing, administration, device-use, self-use, or treatment guidance.[1]
Design, masking, and selected population
Investigators assigned 15 participants to saline, 15 to the lower glutathione exposure, and 15 to the higher exposure. A blinded clinician assessed outcomes. The statistician generated the allocation using Stata and stratified it by eligibility for magnetic-resonance spectroscopy; only the compounding pharmacy was unblinded. Fifteen people entered the spectroscopy subset, far fewer than the full cohort.[1][3]
Participants had Hoehn and Yahr stage 1–3 disease. The group contained 23 men and 22 women, mean age 60.9 years, with Parkinson disease diagnosed an average of about 3.5 years earlier. Every participant was White, and recruitment occurred in one US region. Existing Parkinson medications were part of the background and were intended to remain stable. The study therefore cannot represent all stages, racial and ethnic groups, medication patterns, or care settings.[1]
The intervention was a specific reduced-glutathione preparation delivered intranasally within this protocol. Intranasal identity matters: it is not interchangeable with oral, intravenous, topical, inhaled, or catalog forms. A shared ingredient name cannot establish matching composition, concentration, stability, device performance, exposure, or clinical behavior.[1]
Participant flow and five withdrawals
Forty-five participants were assigned, but the primary UPDRS table contains 14 placebo, 11 lower-exposure, and 14 higher-exposure participants, or 39 total. Five people withdrew during the intervention period: one placebo, two lower exposure, and two higher exposure. One additional difference between 40 people without an intervention-period withdrawal and 39 in the table reinforces why assignment, retention, and analyzed counts should not be treated as one denominator.[1]
The placebo withdrawal involved irritated sinuses and headache attributed to the intervention. Lower-exposure withdrawals involved a fall with broken bone judged unrelated and puffiness beneath the eyes attributed to intervention. Higher-exposure withdrawals involved logistical difficulties judged unrelated and tachycardia with newly diagnosed cardiomyopathy. These reasons matter because a headline saying merely “five withdrew” would hide both arm distribution and medical context.[1]
Primary analysis: within-group change is not superiority
The registry identifies UPDRS as the primary outcome. The paper says its a priori analysis focused on matched-pair change from baseline within each cohort for total and motor UPDRS. It also states that the study was not powered to test superiority over placebo. That choice makes direct between-group results essential to interpretation.[1][3]
Total UPDRS mean changes at week 12 were −3.4 points with placebo (SD 6.2; p=.059 within group), −2.6 with lower exposure (SD 7.4; p=.262), and −4.6 with higher exposure (SD 4.7; p=.0025). The paper used a multiple-comparison threshold of p=.0033 for the principal UPDRS comparisons, so the higher arm’s within-group total change crossed that threshold.[1]
However, the active groups did not outperform placebo. Direct week-12 comparisons yielded p=.558 for placebo versus lower exposure and p=.382 for placebo versus higher exposure. All three groups improved on total UPDRS. A change from baseline can reflect natural fluctuation, expectations, repeated measurement, background care, regression toward the mean, or other study effects. It cannot by itself identify which exposure caused the change.[1]
Motor-UPDRS changes were +0.4 points with placebo (SD 4.8; p=.783), −1.1 with lower exposure (SD 3.9; p=.374), and −2.2 with higher exposure (SD 3.8; nominal p=.0485). Between-group p values were .676 and .643. The motor finding therefore was neither significant under the paper’s multiplicity threshold nor superior to placebo.[1]
Secondary symptoms, medications, and washout
The higher-exposure arm had a mean 10.17-point improvement on the Non-Motor Symptom Scale (SD 13.18; nominal p=.0217 within group). The result was not a demonstrated active-versus-placebo effect and did not meet the stringent .0033 context used for the principal comparisons. Other participant-reported measures, including PRO-PD, Parkinson’s Disease Questionnaire-39, PROMIS instruments, and cognitive testing with MoCA, did not form a consistent pattern favoring active exposure.[1]
Levodopa-equivalent medication changes were not significant. That matters because symptom scores can be affected by background therapy, yet the trial did not establish that glutathione reduced medication requirements. It also did not test disease progression, durable function, falls, hospitalization, institutionalization, or survival.[1]
All three cohorts worsened during the four-week washout observation. Without a maintained randomized contrast capable of separating natural variability and other causes, this pattern does not establish withdrawal, persistent efficacy, or disease modification. A short rebound-looking trajectory is not proof that an exposure slowed Parkinson disease.[1]
Biomarkers and the small imaging subset
No biomarker remained significant after correction for multiple comparisons. Whole-blood total glutathione fell in all cohorts; decreases from baseline were significant in placebo and higher exposure at nominal p=.0142 and .0112. That result does not support a simple story that intranasal exposure raised systemic glutathione.[1]
In the 15-person magnetic-resonance-spectroscopy subset, the higher group showed a trend toward increased brain glutathione, but it was not statistically significant. A small exploratory subset is especially sensitive to missing data, baseline imbalance, measurement noise, and selective emphasis. It cannot establish central nervous system delivery, target engagement, or clinical benefit for the full 45-person cohort.[1]
Tachycardia and cardiomyopathy narrative
One higher-exposure participant had reported a prior episode of anxiety and tachycardia after a stressful event, described as resolved at screening, and was already receiving annual cardiology follow-up. Around study week 5, a scheduled cardiology visit led to a new cardiomyopathy diagnosis. The cardiologist recommended withdrawal, and the paper reports that tachycardia resolved after the study exposure stopped.[1]
The investigators said it was unclear whether the event was related. Temporal sequence alone cannot determine causation, but uncertainty also should not be rewritten as “unrelated.” The faithful record includes the prior history, timing, cardiology diagnosis, withdrawal recommendation, reported resolution, and investigators’ unresolved assessment.[1]
Sponsor, funding, disclosure, and registry state
Bastyr University sponsored the registered study, and the Michael J. Fox Foundation funded it. The paper discloses that lead author Laurie Mischley founded NeurRx. Teleflex provided the device tips used in the protocol. These relationships do not decide whether a result is valid, but they belong in a transparent evidence record.[1][3]
ClinicalTrials.gov listed the study as completed with enrollment of 45 and was last updated June 24, 2016. The history shows results submitted July 28, 2017 and returned after quality-control review August 25, 2017. The current page says “Results Submitted,” but it has no posted structured result tables.[3][4] A returned submission is not public results, and its existence does not let readers reconstruct an unpublished registry analysis.
Statistical significance, effect, uncertainty, and importance
- Statistical significance: one higher-arm within-group total-UPDRS change met the paper’s threshold; direct active-versus-placebo comparisons did not.
- Effect size: total-UPDRS changes were −3.4, −2.6, and −4.6 points, with substantial within-group variation.
- Uncertainty: only 39 appeared in the primary table, five withdrew, multiple outcomes were tested, and the imaging subset contained 15 people.
- Clinical importance: no superiority, durable function, disease-modification, hospitalization, or survival effect was established.
Important limitations
This was a short, small, single-region study in an all-White cohort. It was not powered for superiority, had five intervention-period withdrawals, used multiple clinical and biomarker outcomes, and included a very small spectroscopy subset. Background Parkinson medication and natural symptom variability complicate within-person change. Twelve intervention weeks and four washout weeks cannot resolve disease progression, durability, or uncommon harms.[1]
What this study does not prove
NCT02424708 does not prove that intranasal glutathione was superior to saline, modifies Parkinson disease, raises brain glutathione reliably, reduces medication needs, or provides durable functional benefit. It does not establish causality for the cardiomyopathy account. It also does not validate oral, intravenous, topical, inhaled, compounded, or catalog glutathione, or authenticate another preparation, supplier, device, or lot.[1][3]
Frequently asked questions
Did either glutathione group outperform placebo on UPDRS?
No. All cohorts improved on total UPDRS, but neither active group was significantly superior to saline in the direct week-12 comparisons.[1]
Why is the higher arm’s p=.0025 not proof of efficacy?
That p value describes change within one arm. The causal comparison is active versus placebo, and those comparisons were not significant.[1]
How many participants contributed to the primary table?
Thirty-nine: 14 placebo, 11 lower exposure, and 14 higher exposure, versus 45 originally assigned.[1]
What happened during the washout period?
All cohorts worsened during four weeks of observation, but that pattern cannot by itself establish withdrawal, durability, or disease modification.[1]
Was the cardiomyopathy caused by glutathione?
The investigators said the relationship was unclear. This page preserves the event chronology and does not independently assign or exclude causality.[1]
Are structured results posted on ClinicalTrials.gov?
No. A submission was returned after quality-control review; the live page shows a submitted state without posted result tables.[3][4]
Does this study support other glutathione forms?
No. It studied one intranasal reduced-glutathione preparation and does not establish equivalence to oral, intravenous, topical, inhaled, or other materials.[1]