How ASH grades evidence
The Tier A/B/C framework and GRADE-aligned ratings used across every page.
Every condition, RCT, compound, and biographical entry on hirudotherapysociety.org carries a regulatory-status badge (the colored tag near the page title) and, where applicable, an evidence-quality rating. This page explains what those badges mean.
The three-tier framework
ASH classifies every clinical use of medicinal leech therapy into one of three tiers:
FDA-cleared
The FDA has cleared Hirudo medicinalis as a medical device (K040187, June 21 2004) for venous congestion in surgical flaps. This is the only formally cleared indication. Tier A entries can refer to this clearance without off-label disclaimers.
Example: Venous Congestion in Surgical Flaps, Digit Replantation,Ear Replantation.
Off-label with clinical evidence
Uses outside the FDA-cleared indication that have at least one randomized controlled trial (RCT) or systematic review supporting clinical efficacy. Includes off-label disclaimer language and individualized informed-consent expectation.
Example: Knee Osteoarthritis (multiple RCTs), Chronic Venous Insufficiency, Lateral Epicondylitis.
Investigational
Mechanistic rationale and limited human evidence (case series, small pilots, or preclinical extrapolation). Tier C entries include explicit framing that hirudotherapy is investigational for this use, with patient-discussion expectations covering evidence limits.
Example: Fibromyalgia, Vitiligo, Chronic Tinnitus,Stroke Rehabilitation.
Preclinical / Mechanistic
Pages describing biology, biochemistry, mechanism-of-action research, or omics discovery science. Not a clinical efficacy claim. Common in /science/*,/leech-biology/*, and /research/genome*.
GRADE-aligned evidence ratings
On RCT pages, the evidence-quality badge follows the GRADE Working Group framework. Trial design alone doesn't determine grade \u2014 size, blinding, and consistency all matter.
Sources from FDA 510(k) submission file. Top tier for regulatory framing.
Aggregates multiple RCTs with quality assessment. Used for top-of-page recommendations.
Single-trial evidence. May be downgraded for small N, unblinded design, or selection bias.
Observational design. Suggests effect but does not establish causation.
Single-patient observation. Hypothesis-generating only.
In-vivo animal model. Suggests biological plausibility, not human efficacy.
Cell-culture or cell-free assay. Mechanism-of-action evidence only.
Biochemistry, structural biology, computational modeling. Pre-clinical context.
Pre-modern medical literature. Educational reference, not current clinical guidance.
Practice guideline or scope-of-practice statement from a professional society.
Editorial integrity
- PMID verification: Every PMID cited is verified to resolve to a real PubMed paper with title matching the claim. Three fabricated PMIDs were caught and corrected in May 2026 \u2014 documented in the public changelog.
- "What this does NOT mean" required: Every Tier B/C condition page carries an explicit
WhatThisDoesNotMeanBoxclarifying evidence limits in patient language. - FDA framing fixed-baseline: Every Tier A page references the exact device clearance (K040187, NRN, Unclassified Pre-Amendment). No inflated claims.
- Drug-vs-leech distinction: FDA-approved drugs derived from leech (bivalirudin, lepirudin, desirudin) are NOT the same as whole-leech therapy. Pages describing the drugs carry explicit separation.
- Public corrections queue: Errata are logged in the changelog, not silently revised. All commit history is on public GitHub.