South American bark with centuries of indigenous use for infections, pain, and dysmenorrhea. One small human trial found significant pain reduction in menstrual cramps. No RCTs exist. Primary safety concern is anticoagulant interaction.
Traditions: South American indigenous, South American folk medicine
Inner bark decoction for febrile illness, bacterial and fungal infections (including syphilis, Candida), malaria, gastrointestinal complaints, pain, and dysmenorrhea. Bark simmered in water and consumed as daily tea.
Argentina, Bolivia, Brazil, Paraguay: 'lapacho' consumed for infections, arthritis, skin conditions, and general tonic use. Cancer reputation drove NCI pharmaceutical interest in lapachol from the 1960s onward.
Pau d'arco has minimal human clinical evidence. One uncontrolled pilot in 12 women found significant pain reduction in dysmenorrhea at all time points (p<0.01), with no control arm. One Phase II trial of a multi-ingredient lapacho product in 40 radiotherapy patients showed historically low mucositis rates, though pau d'arco's individual contribution is impossible to isolate. No randomized controlled trials exist for the botanical. Ex vivo data show the aqueous extract suppresses human T-lymphocyte proliferation and pro-inflammatory cytokines, providing mechanistic support for the anti-inflammatory traditional use.
n=12 enrolled, n=9 completers. 1050 mg/day Tabebuia avellanedae for 8 weeks. Significant pain reduction at first dose, 4 weeks, and 8 weeks (p<0.01 at all time points). No control arm — placebo contribution unknown.
n=40 head-and-neck cancer patients on radiotherapy. Lapacho-based multi-ingredient product. Grade 3 mucositis in 10% vs. 70–80% historical rate. Multi-ingredient product; pau d'arco's individual contribution cannot be isolated.
Aqueous extract dose-dependently inhibited primary human T-lymphocyte proliferation and CD25/CD71 expression. Activity was IL-2-independent and not attributable to beta-lapachone — an unidentified polar constituent is responsible.
All five solvent extracts nontoxic to primary human PBMCs. Multiple extracts produced potent suppression of IFN-γ, IL-1β, IL-6, IL-8, TNF, and MCP-1. Some extracts outperformed positive clinical control.
Taste: Mildly astringent, earthy, faintly bitter. Less aggressively medicinal than most medicinal barks. Some drink it plain; honey or ginger complement it.
Primary traditional preparation. Lapachol is poorly water-soluble, so decoctions deliver lower naphthoquinone content than ethanol preparations. However, the immunosuppressive activity identified in human lymphocytes (Böhler 2008) used an aqueous extract — only the aqueous form showed activity, not ethanol — meaning the anti-inflammatory fraction is water-extractable.
Only preparation with any human safety data (McClure 2022). Significant caveat: commercial pills showed no meaningful cytotoxic activity in vitro (GI50 >400 μg/mL) while properly prepared methanolic extracts were active at 76–110 μg/mL (Pires 2015). Standardization and species identity are highly variable in commerce.
Higher ethanol percentage improves naphthoquinone extraction relative to decoction. No regulatory standard or human clinical data exist for this preparation; dosing based on botanical medicine convention.
Pau d’arco is an instructive contrast to well-studied herbs: centuries of traditional use, compelling preclinical activity, and almost no human clinical data. What exists is worth knowing precisely.
Available human evidence:
Mechanistic support (not clinical trials):
What doesn’t apply to the botanical: The NCI tested lapachol — an isolated compound from pau d’arco — as a potential chemotherapy drug in 1974. The Phase I trial was terminated when anticoagulant toxicity appeared before any antitumor concentrations could be reached [5]. Synthetic beta-lapachone analogues (ARQ 501, ARQ 761) reached Phase I/II in refractory solid tumors and showed some signal but remained investigational [6]. Neither pharmaceutical development program says anything meaningful about drinking bark tea — they used purified or synthetic compounds at pharmacological IV doses.
Where the evidence stands: One small uncontrolled pilot for dysmenorrhea. One uncontrolled multi-ingredient product trial. Preclinical and ex vivo mechanistic data. The traditional reputation substantially outpaces what’s been tested in humans.
Pau d’arco’s home is Amazonia and the forests of Brazil, Bolivia, Argentina, and Paraguay. The Guaraní and Tupinambá peoples documented its use before European contact, primarily as an inner bark decoction for febrile illness, bacterial and fungal infections (including syphilis and Candida), malaria, gastrointestinal complaints, pain, and dysmenorrhea [8].
The “lapacho” tradition — a daily bark tea consumed for general health — spread through Argentine and Bolivian folk medicine. Japanese and Western alternative medicine markets adopted it in the 1960s and 70s under the name “taheebo,” creating global commercial demand. The herb’s cancer reputation drove the US National Cancer Institute to investigate lapachol starting in the 1960s, which ultimately produced the aborted Phase I trial described above.
This is not an Ayurvedic or TCM herb. Pau d’arco has no documented place in either tradition — it is exclusively a New World botanical with South American indigenous roots. Its global popularity is a product of 20th-century commercial expansion, not cross-cultural traditional validation [8].
One naming issue worth knowing: Current taxonomy uses Handroanthus impetiginosus, but most research literature and commerce uses Tabebuia impetiginosa or Tabebuia avellanedae (these are synonyms for the same tree). What’s actually in commercial products isn’t always the right species — mislabeling is well-documented [8].
The inner bark — not the outer bark — is the traditional medicinal material, comprising around 88% of documented preparations [8]. This distinction matters when purchasing.
Bark decoction (traditional, and mechanistically interesting):
The decoction is the primary traditional preparation. Critically, the immunosuppressive activity identified in primary human T-lymphocytes used an aqueous extract — the ethanol extract did not show activity [3]. Whatever polar constituent produces this effect is water-extractable. Traditional decoction delivers it; capsules of ethanolic extracts may not.
Capsules (the only form with human safety data):
Tincture (unvalidated, higher naphthoquinone extraction):
No dose-titration data exist. Start at the lower end: one cup of decoction or 500 mg capsule daily. The safety trial found no serious adverse events at 1050 mg/day over 8 weeks in 9 women [1].
For dysmenorrhea: allow at least 2–3 full menstrual cycles to assess effect. The pilot showed benefits at first dose, but the lack of controls means the magnitude of the herb’s actual contribution is unknown.
Baseline (1 week before starting):
During the trial:
RED FLAGS — Stop immediately:
The anticoagulant mechanism of lapachol makes bleeding-related signals the key safety indicator. They didn’t appear in the small pilot [1], but that trial was 8 weeks in 9 people — not a comprehensive safety dataset.
What the dysmenorrhea pilot participants experienced: significant pain reduction at every time point measured across 8 weeks. Whether any of that was the herb is scientifically unknown, but the traditional indication aligns with the trial’s design.
Pau d’arco decoction has an unexpectedly mild character for a medicinal bark. It’s slightly astringent, earthy, and faintly bitter — less aggressive than valerian, less sharp than yellow dock. The woody, faintly sweet aroma is pleasant. Most people can drink it plain; honey, ginger, or cinnamon complement it without masking it.
Simmer, don’t boil. Heat during aggressive boiling may degrade phenolic constituents [7], and the traditional preparation specifically calls for a gentle simmer.
If you expect something dramatic and medicinal-tasting, you’ll likely find it underwhelming in flavor. That’s not a bad thing for a daily tea.
Pau d’arco has serious quality control problems in commerce.
Species mislabeling is documented and common [8]. Products labeled “pau d’arco” may contain Handroanthus impetiginosus, Tabebuia avellanedae, related Handroanthus species, or something else. Look for products that explicitly state Handroanthus impetiginosus or Tabebuia impetiginosa (same species, different accepted name) on the label.
Inner bark, not outer bark. Around 88% of traditional preparations documented across Brazil, Bolivia, and Argentina specify inner bark [8]. Labels should state “inner bark” (cortex). Outer bark products are less desirable.
Commercial bioactivity is substantially reduced. Commercial pills and syrups tested in vitro showed no significant cytotoxic activity at concentrations up to 400 μg/mL, while laboratory methanolic extracts were active at 76–110 μg/mL [7]. Whether this gap matters for the anti-inflammatory and dysmenorrhea uses is unknown — but it suggests the active constituents are not reliably present in many commercial formats.
No official standard exists. No WHO, Commission E, EMA, or USP monograph has been published for pau d’arco. There are no regulatory benchmarks for lapachol content. Given this vacuum, prioritize:
Pau d’arco carries a significant traditional reputation built over centuries of South American indigenous use — for infections, dysmenorrhea, and general tonic purposes. The human clinical evidence is thin: one uncontrolled pilot in 12 women found pain reduction for dysmenorrhea at all measured time points, and one uncontrolled multi-ingredient product trial showed better-than-historical mucositis rates in radiotherapy patients. No randomized controlled trials exist.
The one safety fact that matters most: lapachol inhibits the same enzyme as warfarin. If you take any anticoagulant medication, pau d’arco is contraindicated. Pregnancy is also contraindicated based on animal data.
If dysmenorrhea is what draws you: the pilot data align with the traditional indication, and even without controls the consistent effect across time points is worth noting. Track your own pain systematically across 2–3 cycles against a baseline and see what you observe.
If infections or cancer are what drew you here: the traditional reputation drove serious pharmaceutical research that ultimately didn’t pan out for isolated constituents. Whole-herb preparations have no clinical data for these uses.
Prioritize quality — commercial bioactivity varies dramatically, species mislabeling is common, and inner bark is the thing that’s been used. Respect the anticoagulant contraindication. Everything else is exploratory.
Duration: Traditional use: 2–4 weeks for acute conditions. The only human safety trial ran 8 weeks.
What to notice:
No dose-titration data exist. Start at the lower end: one cup of decoction or 500 mg capsule daily. The safety trial used 1050 mg/day without serious adverse events in 9 women over 8 weeks. For dysmenorrhea: allow at least 2–3 full menstrual cycles before assessing effect. Set your baseline carefully — the only human trial had no control arm, so your own before/after comparison is most of the evidence you'll have.
Generally considered: caution
Contraindications:
Pregnancy/Nursing: Contraindicated in pregnancy based on animal fetotoxicity data. No human pregnancy studies exist. Avoid during lactation.
Mild adverse events occurred in 75% of participants in the only human trial (n=9 completers, 1050 mg/day, 8 weeks) with no serious adverse events reported. The anticoagulant interaction risk is the primary clinical concern — lapachol's mechanism is mechanistically identical to warfarin's. Rat genotoxicity studies showed no significant DNA damage at 100 mg/kg (the dose range approximating traditional human use) but significant genotoxicity at 300 mg/kg; no human genotoxicity data exist. No official monograph, maximum safe dose, or regulatory classification exists from any major body (WHO, EMA, Commission E, USP).