Ten research investigations

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Is the clinically effective experimental butterbur root extract interchangeable with any retail butterbur product, including unverified “PA-free” products?

**Sources and study classes:** Lipton et al., Neurology 2004, RCT, [PMID 15623680](https://pubmed.ncbi.nlm.nih.gov/15623680/), DOI `10.1212/01.wnl.0000147290.68260.11`; [NIH LiverTox butterbur monograph](https://www.ncbi.nlm.nih.gov/books/NBK547997/), adverse-event case assessment; [NCCIH butterbur safety summary](https://www.nccih.nih.gov/health/butterbur), evidence/guideline context. Case-series attribution is not an efficacy RCT.

**Observed trial evidence:** Three-arm, 245 adults with migraine, four months. In the trial **per-protocol** analysis, 75-mg-twice-daily *specific extract* reduced migraine attacks by 48% from baseline vs 26% placebo (**p=.0012**); the tested lower 50-mg-twice-daily arm's 36% reduction **was not significant** versus placebo (**p=.127**). Fifty-percent responder rates were 68% for the higher-exposure arm vs 49% placebo; do not confuse within-arm percentage reductions with a 48-percentage-point placebo-adjusted effect.

**Safety boundary:** Butterbur plant parts can contain hepatotoxic pyrrolizidine alkaloids (PAs); preparation identity and removal/verification matter. LiverTox reviews rare clinically significant hepatitis associated with commercial preparations, while causality/contamination attribution is disputed. NCCIH notes the AAN withdrew its favorable recommendation in 2015 because of serious safety concerns. A “PA-free” label is not an independent laboratory certificate and does not prove every current lot is safe.

**Null/limits:** Lower-exposure arm did not outperform placebo; 4-month efficacy does not establish years-long safety, prevention of stroke, or product equivalence. Do **not** convert studied amounts into general dosing instructions.

**Disposition:** **P0** product-identity and liver safety editorial hold; no affiliate recommendation or efficacy promotion without independent review.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Can a 2005 trial establish broad migraine prophylaxis efficacy without mentioning the 2002 multi-dose failure?

**Primary sources:** Diener et al., Cephalalgia 2005, [PMID 16232154](https://pubmed.ncbi.nlm.nih.gov/16232154/), DOI `10.1111/j.1468-2982.2005.00950.x`; Pfaffenrath et al., Cephalalgia 2002, [PMID 12230594](https://pubmed.ncbi.nlm.nih.gov/12230594/), DOI `10.1046/j.1468-2982.2002.00396.x`. Both human randomized placebo-controlled trials of **CO2-extract MIG-99**; do not generalize to feverfew leaf teas or heterogeneous supplement powder.

**2005 trial:** 218 participants randomized; **170 in efficacy ITT set** (89 active, 81 placebo) after protocol-eligibility exclusions. Over defined months, migraine attacks fell ~1.9 per month active vs ~1.3 placebo (**p=.0456** on primary model). Some secondary outcomes were significant; adverse events possibly related to treatment 8.4% vs 10.2% placebo (no significant difference).

**2002 contrast:** Adaptive 147-person dose-ranging study over 12 weeks **failed to show significant efficacy overall and failed its dose-response hypothesis**; an approximately 49-person baseline-severity subgroup showed a positive signal not sufficient to reverse the overall null.

**Boundary:** Apparent benefits are extract-specific, study-design dependent, and modest; 2005 exclusions and borderline primary p-value affect certainty. Do not imply every available feverfew product, acute migraine treatment, or pregnancy safety is proved.

**Disposition:** P1 contradiction-observatory candidate; surface both trials and their inclusion populations before assigning evidence tier.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does the 8-week result justify saying the primary endpoint succeeded or substituting Asian ginseng?

**Source:** Barton et al., J Natl Cancer Inst 2013, [PMID 23853057](https://pubmed.ncbi.nlm.nih.gov/23853057/), DOI `10.1093/jnci/djt181`, phase-III randomized double-blind placebo trial, **364 participants** with cancer-related fatigue, eight-week **American ginseng (*Panax quinquefolius*)** treatment.

**Primary endpoint:** Week-four change in general Multidimensional Fatigue Symptom Inventory score **14.4** vs **8.2** placebo, **p=.07 (not statistically significant)**. **Week eight (secondary time point)** change **20.0** vs **10.3**, **p=.003**; the published subgroup suggests larger effect during active cancer treatment, but avoid treating subgroup as independent replication. Recorded adverse events did not significantly differ during this short trial.

**Evidence boundaries:** American (*P. quinquefolius*) is not interchangeable with Asian (*P. ginseng*) or retail mixed-ginsenoside products. Week-eight effect is not a successful week-four primary outcome, treatment of cancer, proven survival impact, long-term safety, or a recommendation for a patient receiving chemotherapy.

**Disposition:** P1 outcome-hierarchy and species firewall. Cancer-care interactions require independent review before live safety copy.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Is an acute pre-anesthesia signal sufficient for claims about chronic anxiety or sleep therapy?

**Source:** Movafegh et al., Anesth Analg 2008, [PMID 18499602](https://pubmed.ncbi.nlm.nih.gov/18499602/), DOI `10.1213/ane.0b013e318172c3f9`, randomized double-blind placebo-controlled **60-person** ambulatory surgery premedication trial.

**Results:** Anxiety numerical rating scores significantly lower during 90-minute preoperative observation in trial Passiflora group vs placebo (**p<.001**); measured sedation, postoperative psychomotor recovery and discharge times did **not** significantly differ.

**Null/boundaries:** A nonsignificant sedation finding in 60 people and one preoperative exposure does **not** prove lack of sedation or drug interactions in larger populations, nor equivalence to established therapies. The exact studied product was **Passipy**; passionflower tea/extract concentrations, context of anesthesia, GAD, insomnia and chronic dosing are separate evidence domains.

**Disposition:** P1 formulation, outcome and setting-specific research; do not promote as a generic anxiolytic safety guarantee. **Existing repository coverage includes PMID 18499602**, so plan source reuse, not novelty.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does symptom relief in five weeks demonstrate long-term nerve-disease modification?

**Sources:** Ziegler et al., Diabetes Care 2006 SYDNEY 2 RCT [PMID 17065669](https://pubmed.ncbi.nlm.nih.gov/17065669/), DOI `10.2337/dc06-1216`; Ziegler et al., Diabetes Care 2011 NATHAN 1 RCT [PMID 21775755](https://pubmed.ncbi.nlm.nih.gov/21775755/), DOI `10.2337/dc11-0503`.

**Short-term:** SYDNEY 2 enrolled **181** people with symptomatic diabetic distal symmetric polyneuropathy. At five weeks, symptom-score reductions in three active oral groups ranged about **48–52%** vs **32%** placebo (each active arm **p<.05** versus placebo). GI/vertigo adverse effects rose with higher studied exposures.

**Long-term contrast:** NATHAN 1 randomized **460** people for four years. **Prespecified primary composite endpoint did not differ significantly**, p=.105. Some neuropathy impairment secondary scales favored treatment; serious adverse events were **38.1% active vs 28.0% placebo**, an observed imbalance that cannot independently establish treatment causation.

**Null/boundaries:** Symptoms, secondary impairment tests, composite neurophysiology, nerve regeneration, and foot-ulcer prevention are *different* outcomes. Long-term primary null cannot be converted into “reverses neuropathy.” Exposure amounts are trial descriptors, not consumer dosing guidance.

**Disposition:** P1 precision correction; separate route (oral vs intravenous) and condition-specific population.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does a positive subgroup among treated uncontrolled hypertensive patients translate to lowering blood pressure in everyone?

**Source:** Ried et al., Maturitas 2010, [PMID 20594781](https://pubmed.ncbi.nlm.nih.gov/20594781/), DOI `10.1016/j.maturitas.2010.06.001`, placebo-controlled randomized **50-person** study across 12 weeks, one *aged garlic extract* product as **adjunct to existing treatment**.

**Results:** In the baseline subgroup with systolic BP **at least 140 mm Hg**, systolic BP was **10.2 ± 4.3 mm Hg lower than controls** (p=.03) averaged across treatment period. In participants with lower baseline systolic BP, between-group changes were **not statistically significant**.

**Limits:** The 50-person study's baseline-BP subgroup result does not establish a generic garlic-food or garlic-powder effect; endpoint was BP, **not** strokes/heart attacks. Does not establish drug equivalence, safety of replacing antihypertensives, or effect in normotensive adults.

**Disposition:** P1 formulation and baseline-stratification enrichment; do not build product effectiveness rankings from a single studied preparation.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does a concentration-specific liver risk apply to ordinary brewed tea, and is the clinical risk measured?

**Sources:** Dostal et al., Food Chem Toxicol 2015, [PMID 26051348](https://pubmed.ncbi.nlm.nih.gov/26051348/), DOI `10.1016/j.fct.2015.05.019`, randomized double-blind **1,075 postmenopausal women**, one year, concentrated green tea extract versus placebo. Yu et al., Cancer Prev Res 2017, [PMID 28765194](https://pubmed.ncbi.nlm.nih.gov/28765194/), DOI `10.1158/1940-6207.CAPR-17-0160`, additional ALT analysis from **the same Minnesota Green Tea Trial**, **not** a second independent RCT.

**Primary safety observation:** ALT elevation occurred in **36 participants (6.7%) extract vs 4 (0.7%) placebo**, p<.001. Overall any-adverse-event reports were 75.6% vs 72.8% (not significantly different); serious adverse events 2.2% vs 1.5% (not significant), while GI symptoms differed by type. ALT-associated serious adverse events were described in 1.3% of the extract arm.

**Interpretation:** Liver enzyme elevation **is a biomarker/safety signal, not a measured liver-failure rate**. This exposure was a standardized high-catechin **extract** containing EGCG, not ordinary brewed green tea. Neither numeric rate is a current-market incidence estimate across products, ages, sexes or meal conditions.

**Null/boundaries:** Overall AE and serious AE comparisons null; do not portray an ALT increase as clinical liver injury in every exposed person, and do not infer safety of all extracts if low exposure differs.

**Disposition:** **P0** risk-tier/formulation quality review; prevent green-tea-beverage vs high-EGCG-extract confusion. No consumption/dose recommendation.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does a BP effect persist across treated and untreated hypertension and across follow-up duration?

**Sources:** Kapil et al., Hypertension 2015, randomized 68-person trial, [PMID 25421976](https://pubmed.ncbi.nlm.nih.gov/25421976/), DOI `10.1161/HYPERTENSIONAHA.114.04675`; Bondonno et al., Am J Clin Nutr 2015, 27-person double-blind crossover, [PMID 26135348](https://pubmed.ncbi.nlm.nih.gov/26135348/), DOI `10.3945/ajcn.114.101188`.

**Positive RCT:** Kapil randomized **68 hypertensive participants**, stratified **34 drug-naive and 34 treated**, to a nitrate-rich juice preparation versus nitrate-depleted juice over four weeks. Placebo-adjusted clinic BP lower by mean **7.7/2.4 mm Hg**; 24-hour ambulatory BP **7.7/5.2 mm Hg lower**, with reported favorable endothelial surrogate tests.

**Contrary RCT:** Bondonno enrolled **27 treated-hypertensive participants** in one-week crossover: nitrate-rich juice markedly raised nitrate/nitrite in blood/saliva/urine, but **neither home nor 24-hour ambulatory BP differed** versus nitrate-depleted juice.

**Evidence boundaries:** Both are RCTs with different treatment populations and durations. Biomarker evidence of nitrate delivery is **not** synonymous with BP efficacy. Do not infer heart attack/stroke prevention, replacement of prescriptions, universal effect, or all beetroot powders contain the tested nitrate content.

**Disposition:** P1 contradiction-observatory and formulation/endpoint mapping.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Was the same benefit established across both randomized sequential-comparison studies?

**Source:** Papakostas et al., Am J Psychiatry 2012, [PMID 23212058](https://pubmed.ncbi.nlm.nih.gov/23212058/), DOI `10.1176/appi.ajp.2012.11071114`, report of **two separate** randomized double-blind sequential-parallel comparison trials in people with major depressive disorder not responding adequately to SSRIs.

**Results:** **First trial, n=148**, lower-dose initial augmentation strategy: **no significant treatment-group outcome differences**. **Second trial, n=75**, used 15-mg/day L-methylfolate throughout the periods and reported significant improvements in **two primary response/symptom outcomes** versus SSRI+placebo; reported number needed to treat approximately **six** for response. Trial dosing is historical trial exposure, **not** advice.

**Interpretation:** The first null must accompany the second positive. This was an **adjunct** in a selected psychiatric population, not a replacement for antidepressants, correction of demonstrated folate deficiency, treatment of every depression subtype, or evidence for standard multivitamins. The higher-strength intervention is not equivalent to generic folic acid or food folate. Short follow-up, trial sponsorship/conflict assessment and rare harms require independent review.

**Disposition:** P1 efficacy/formulation precision; no changes to user medication or medical advice.

Original notes reproduced from the merged research dossier without converting preliminary interpretations into approved medical recommendations. Source links are provided above. Other limitations in the full text still apply.

**Question:** Does an oral lavender-oil trial justify claims for aromatherapy or equivalence to paroxetine?

**Source:** Kasper et al., Int J Neuropsychopharmacol 2014, [PMID 24456909](https://pubmed.ncbi.nlm.nih.gov/24456909/), DOI `10.1017/S1461145714000017`, **539 adult** GAD randomized double-blind double-dummy trial, ten weeks, **oral Silexan** two exposure levels, paroxetine active control, and placebo.

**Results:** Hamilton Anxiety score reductions were **14.1** and **12.8** for active Silexan levels, **11.3** for paroxetine, and **9.5** for placebo. Two Silexan vs placebo comparisons p<.01 in full-analysis set; paroxetine versus placebo **p=.10** in full set, but observed-cases analysis did indicate significance. This trial did **not** demonstrate Silexan is clinically superior to paroxetine or interchangeable with all prescribed therapies.

**Evidence boundaries:** Trial interventions were **specific standardized oral lavender-oil capsules**, not diffuser use, lavender perfume, essential-oil ingestion generally, or topical preparations. Sponsor affiliations include product manufacturer; effectiveness/harms beyond ten weeks, pediatric use, pregnancy, and long-term outcomes are not established by this trial.

**Disposition:** P1 branded/route outcome correction; PMID **24456909 already occurs in repository citation exports/guide material**, so reconcile and reuse.

## Admission, deduplication and QA contract

1. **Receipt**: 10 bounded subject investigations, distinct from DE10 and DE11, including positive/null contrasts. PMID references in research artifacts are not counted as new source admissions.
2. **Existing repo source spot-check**: PMID **18499602** and **24456909** already show up in repository content/exports. Other PMIDs returned no matches under limited default-branch GitHub code search, which **does not establish canonical novelty**. Runtime/workbook exact-ID verification remains mandatory.
3. **Evidence classes**: All 10 dossiers use human RCTs. The butterbur safety adjunct is NIH LiverTox observational/case narrative, not a randomized risk frequency. The green tea extract follow-on analysis reuses the original RCT cohort. Do not double-count related papers as independent trials.
4. **Mandatory nulls**: Butterbur lower arm null; 2002 feverfew overall null; American ginseng four-week prespecified primary null; passiflora sedation and psychomotor non-difference; NATHAN1 long-term primary composite null; aged garlic lower-baseline subgroup null; green tea overall serious AEs non-difference (despite ALT signal); treated-hypertension Bondonno nitrate null; first methylfolate trial null; paroxetine full-analysis comparison not significant in Silexan trial.
5. **Prohibited extrapolations**: Species/extract identity transfer, same-cohort independence inflation, turning within-arm declines into placebo-adjusted benefit, converting secondary findings into successful primary outcomes, extrapolating surrogate biomarkers to disease prevention, asserting safety from nonsignificant harm measures, or marketing trial exposures as prescribed doses.
6. **Source integrity**: All linked PMIDs correspond to the named original paper titles checked during this run. Independent second-person review, formal risk-of-bias assessment, current retraction surveillance, and canonical workbook/source-alias reconciliation **remain pending**.
7. **Required next work to create value on the live site**: identify specific existing public claims for each topic; mark only verified conflicts; separate safety-critical P0 edits; independently verify exact source passages and caveats; apply approved corrections in separate content PRs with tests; measure indexing/AI citations afterward. No assumption that an investigation necessarily implies a defective live claim.
8. **ROI evidence**: This research-only document is not a production deployment, not proof of added indexed citations, traffic, revenue, medical efficacy, or guaranteed AI citation growth.