The risk is concentrated. St John's wort with almost any serious prescription, grapefruit with certain statins, calcium or iron with thyroid pills, gram-dose niacin on top of a statin, and serotonin-boosting supplements stacked on antidepressants — those are the pairs that do real damage. Most other combinations are fine. The cheapest fix in medicine is a single written list of everything you swallow, shown at every appointment, with a four-hour gap between your minerals and your thyroid pill.
Your liver and the lining of your gut run a janitorial service that breaks down most pills before they reach the bloodstream. A family of enzymes called CYP enzymes does the breaking; a pump called P-glycoprotein does the ejecting. A small number of supplements work on those two systems directly.
Some supplements crank the system up. St John's wort is the strongest example — it tells the liver to make more CYP enzymes and more of the pump, and within two weeks the body is clearing drugs faster than the prescription was designed for. Cyclosporine slips below the level that keeps a transplanted kidney safe. Indinavir drops below the level that suppresses HIV. The birth-control pill drops below the level that prevents ovulation. The supplement isn't doing anything to the drug directly; it's training the cleanup crew to work overtime Nicolussi et al. 2020.
Other supplements and foods jam the system shut. Grapefruit juice has a chemical called bergamottin that disables the CYP enzymes in your gut wall for about a day after a single glass Lown et al. 1997. With the cleanup crew on strike, a normal dose of simvastatin can deliver ten times the intended exposure to muscle and liver. Goldenseal does something similar over a longer time horizon Gurley et al. 2008. So does the piperine added to many curcumin and turmeric supplements: the black-pepper extract that helps the curcumin absorb does it by slowing this same gut clearance, so it can raise the level of other drugs that ride the same path.
The third family doesn't touch CYP at all — it's pure chemistry in the gut. Calcium, iron, magnesium and aluminum form sticky complexes with thyroid hormone tablets, with most antibiotics in the fluoroquinolone family (ciprofloxacin, levofloxacin) and with the tetracycline family. The drug and the mineral grab onto each other in the stomach and the resulting clump can't cross the gut wall. Levothyroxine absorption drops by twenty to fifty percent if you swallow it with the calcium chew Wiesner et al. 2021.
The fourth family is what your body does after the drug is in. Vitamin K rebuilds the clotting factors warfarin is designed to block; a sudden week of leafy greens or natto can push the INR around enough to matter Violi et al. 2016. Serotonin precursors like 5-HTP and SAMe push the same chemistry SSRIs push; stacking them on top can cross a line into serotonin syndrome Boyer & Shannon 2005. Niacin at the gram doses found in cholesterol-lowering supplements multiplies the muscle-damage risk of statins, an effect large enough to detect in twenty-five-thousand-patient trials HPS2-THRIVE 2014.
Four pathways, a small cast of culprits. Everything that follows is bookkeeping.
How sure are we
The strongest evidence comes from a small set of controlled studies that fed healthy volunteers a supplement, then a prescription, and measured the difference. They are unusual in pharmacology — most herbal claims rest on case reports. These don't.
The same supplement does the same thing to immunosuppressants. Two patients on stable cyclosporine after heart transplants started St John's wort for low mood; both rejected their grafts within weeks, both recovered after stopping the supplement and resuming standard cyclosporine doses Ruschitzka et al. 2000. A trial in ten kidney-transplant patients showed why: two weeks of St John's wort cut tacrolimus exposure by more than half, and the immunosuppressant dose had to be nearly doubled to keep the graft safe Mai et al. 2003.
Warfarin patients lose anticoagulation the same way. Twelve healthy men took a single dose of warfarin alone, then again after two weeks of St John's wort. Warfarin was cleared from the body 29% faster, and the anticoagulation effect dropped by about a fifth — enough, in a real patient, to swing the INR from therapeutic to inadequate Jiang et al. 2004. A 2021 systematic review catalogued seventy-eight foods, herbs and supplements with documented warfarin interactions; about three in five strengthen warfarin and one in three weaken it Tan & Lee 2021.
Grapefruit's effect is the same story in reverse — the supplement strengthens the drug instead of weakening it. The mechanism was nailed down in the late nineties: a single glass of grapefruit juice halves the amount of CYP3A4 protein in your gut wall within four hours, and the effect lasts about a day after one drink or three days with daily intake Lown et al. 1997. The clinical consequences span eighty-five drugs, summarized in a 2013 review that tabulated a hundred and two clinical trials Bailey et al. 2013.
The chelation interactions live in a different evidence world — they don't need outcome trials because the mechanism is direct chemistry. A 2021 systematic review of sixty-three studies confirmed that calcium, iron and other multivalent cations reduce levothyroxine absorption by twenty to fifty percent when taken simultaneously, and that a four-hour separation eliminates the effect Wiesner et al. 2021. Fluoroquinolone and tetracycline absorption falls by similar margins under the same chemistry Manakova et al. 2021.
Who needs to care most
For most readers on no prescription drugs, this entry is background reading — useful before the next ear infection or thyroid diagnosis, otherwise filed away. Four populations face the steepest stakes.
If you take a drug with a narrow safety margin — warfarin, lithium, anticonvulsants like phenytoin or carbamazepine, antiretrovirals, organ-transplant medications (cyclosporine, tacrolimus, sirolimus, mycophenolate), oral chemotherapy pills like imatinib or ibrutinib, digoxin, or thyroid replacement — every new supplement is a question for the prescriber. A twenty-percent shift in drug level matters here in a way it doesn't for most other prescriptions Nicolussi et al. 2020.
If you take levothyroxine, the bottle of calcium chews on the kitchen counter and the multivitamin with iron in the bathroom cabinet are subtracting a sizeable fraction of your dose every morning Wiesner et al. 2021. The four-hour rule is the entire fix.
If you take an antidepressant (SSRI, SNRI, MAOI, tramadol for chronic pain, even high-dose dextromethorphan in cough syrup), do not stack St John's wort, 5-HTP, SAMe, popular adaptogens like ashwagandha or rhodiola, or any over-the-counter mood supplement on top without telling the prescriber. The supplement can blunt the prescription drug or, at the other extreme, push serotonin chemistry into a syndrome that lands in the emergency room Boyer & Shannon 2005.
If you take five or more prescriptions — the situation for most adults over sixty-five — the math compounds. A US analysis of community-dwelling older adults found two-thirds were on five or more medications or supplements concurrently, and roughly one in seven was at risk for a major drug interaction in any given year Qato et al. 2016. The risk concentrates in this group not because their supplements are unusual, but because they have more prescription targets to disturb.
What most people get wrong
"Natural means safe to combine." The strongest documented interactions are with the most natural-sounding supplements. St John's wort is a yellow wildflower; grapefruit is a fruit. Both are pharmacologically active at the same place statins, antibiotics, immunosuppressants and antidepressants are processed. The label doesn't help you here.
"Fish oil thins your blood — stop it before surgery." A 2024 meta-analysis pooled eleven trials and a hundred and twenty thousand patients and found no increase in bleeding from omega-3 supplements, even when patients were also on aspirin or warfarin Javaid et al. 2024. Only the very high pharmaceutical doses (around 4 grams of purified EPA) showed a small signal. A 1-gram fish oil capsule before a colonoscopy is not the problem your preoperative checklist thinks it is.
"Ginkgo causes bleeding." The case-report literature looks alarming, but eighteen controlled trials in two thousand adults showed no effect on platelet aggregation, prothrombin time, or any other clotting measure Kellermann et al. 2011. The case reports almost always involve patients also on warfarin or aspirin; ginkgo gets blamed because it's the one in the supplement bottle.
"Warfarin means no leafy greens." The opposite — consistent moderate vitamin K intake stabilizes the INR better than strict avoidance, because a sudden gap of vitamin K is as destabilizing as a sudden surge Violi et al. 2016. Eat your salad on Tuesday and Friday, not Tuesday this week and not next.
"Niacin is just a vitamin." At the vitamin doses found in a multivitamin (around 20 mg), yes. At the gram doses sold to lower cholesterol, niacin is a pharmacological agent that interacts with statins badly enough to ruin two large outcome trials HPS2-THRIVE 2014 AIM-HIGH 2011. The "no-flush niacin" bottle on the shelf is not in the multivitamin category.
"Red yeast rice is the natural alternative to a statin." It isn't an alternative — it is a statin, naturally containing the same compound as prescription lovastatin. Stacked on top of a prescribed statin, it quietly doubles the dose.
The pairs to never combine
The list below collects the combinations supported by trial data, large case series, or specialty-guideline consensus. None of these is exhaustive, and a prescriber should be the final word — but if you find yourself in any row, you should not be guessing alone.
What to actually do
The action shrinks to two habits and a timing rule.
For a narrow-therapeutic-index drug — warfarin, lithium, transplant immunosuppressants, antiretrovirals, oral chemotherapy, anticonvulsants, levothyroxine, digoxin — go further: do not start any new supplement without checking with the prescribing physician first. The free interaction checkers maintained by pharmacology databases (a search of the drug name plus the supplement name turns them up) catch most of the well-documented pairs. For anything serious, the prescriber's read is the one that matters.
Two weeks before any scheduled surgery, an honest conversation with the surgical team about every supplement you take is worth more than the generic "stop everything herbal" advice. St John's wort affects anesthesia metabolism; some supplements genuinely raise bleeding risk and most don't. The team can give specific guidance you can't get from a checklist.
What it looks like when this goes wrong
The version of this that ends up in the medical literature isn't dramatic in the moment. It's a stable transplant patient who started a yellow-wildflower tincture for low mood and felt nothing change for three weeks; the rejection paperwork arrives at the fourth week's clinic visit Ruschitzka et al. 2000. It's a woman on the same combined pill she's taken for two years walking into a gynecology appointment with an unexpected pregnancy that traces back to a supplement bottle on her counter Schwarz et al. 2003. It's the HIV patient whose viral load comes back detectable for the first time in years, who didn't think the herbal tea counted as medication Piscitelli et al. 2000.
For the more typical reader it's smaller and more annoying. Months of feeling like the thyroid replacement isn't working — tired afternoons, hair on the pillow, weight creeping back — because the morning calcium chew and the levothyroxine were sharing a stomach Wiesner et al. 2021. A second course of antibiotics because the first one was taken with the daily multivitamin Manakova et al. 2021. Muscle aches and dark urine after weeks of grapefruit at breakfast and a statin at night that suddenly hits ten times harder than designed Bailey et al. 2013.
None of these is rare in absolute numbers. A US analysis put fifteen percent of older adults at risk for a major drug interaction in any given year, with rates rising as polypharmacy and supplement use rise together Qato et al. 2016. The cost of prevention is the smallest in medicine: one list, one sentence at the appointment, one timing habit.
Related topics worth a separate read: the individual supplements themselves — St John's wort for depression, fish oil for cardiovascular risk, magnesium for sleep, vitamin D, creatine — each have their own entries with their own benefit and risk profiles. Warfarin, statins, and levothyroxine are the prescription drugs that show up most in the interaction literature, and dedicated entries on each cover what to expect from the prescription side. Alcohol and grapefruit share some metabolism with this story; the alcohol entry covers the rest. For the related but distinct topic of drug–food interactions outside of supplements, the food category index is the entry point.
- — Curcumin-with-piperine is a textbook offender — skip it if you're on meds that ride through the liver, and never on blood thinners.
- — St John's wort speeds up the liver and can drop the pill's hormone levels enough to fail — one of the riskier supplement-drug pairs.
- — Vitamin E pills are an easy one to forget you're taking — they nudge bleeding risk alongside any anticoagulant.
- — Ashwagandha and rhodiola interact with common prescriptions — thyroid, anticoagulants, antidepressants. Tell your doctor you take them.
- — Red yeast rice is a hidden statin — stacking it on a prescribed statin doubles the dose unknowingly.
- — Vitamin K2 is a textbook example of why warfarin patients must vet supplements with a clinician.
- — Anticholinergic burden is one reason to lay out everything you take, prescription or not.
- — The medication review is the natural moment to surface risky drug-supplement pairs too.
- — The CYP3A4 enzyme grapefruit and St John's wort hit is a main place supplements collide with prescriptions.
- — Iron and calcium bind thyroid pills in the gut — space them four hours apart.
- — The fix for hidden interactions is one list every prescriber can see — that's what the health record is.
- — Interactions and inherited metabolism both bend how a drug hits you — the gene test explains the part diet and supplements can't.
- — Knowing a supplement is contaminant-free doesn't tell you if it clashes with your meds — check both before you take it.
Substance and claimed effects
"Drug-supplement interactions" is the umbrella for clinically meaningful pharmacokinetic and pharmacodynamic interactions between common over-the-counter dietary supplements (herbal extracts, vitamins, minerals, amino acids, fish oil) and prescription medications. The entry covers four mechanistic pathways flagged in the brief: cytochrome P450 (CYP) induction/inhibition (most clinically: CYP3A4, CYP2C9, CYP2D6); P-glycoprotein (P-gp / ABCB1) induction/inhibition at the intestinal and hepatic level; anticoagulation and antiplatelet pathways (vitamin K, platelet aggregation, INR effects); and serotonergic pathways (5-HT precursors, MAO inhibition, SSRI/SNRI synergy). The downstream consequences scored are loss of drug efficacy (e.g., transplant rejection, contraceptive failure, viral rebound, subtherapeutic INR), drug toxicity from raised plasma levels (statin myopathy, calcium-channel-blocker hypotension), bleeding events, serotonin syndrome, and reduced absorption of narrow-therapeutic-index drugs (levothyroxine, fluoroquinolones, tetracyclines). The interactions are documented in case reports, controlled pharmacokinetic crossover trials, systematic reviews, and regulatory advisories. The reader-facing entry covers awareness and decision-making, not protocols — most actions reduce to "disclose supplements at every prescription" and "separate timing or substitute". The substance is an information hazard rather than a take-this-or-don't intervention, so meta dimension scores are anchored on harm prevented rather than benefit conferred.
Evidence by addressing question
Mechanism
Four mechanism families carry essentially all clinically important interactions documented to date.
CYP induction. Several botanicals activate the pregnane-X-receptor (PXR), the master transcriptional regulator of CYP3A4, CYP2C9, CYP1A2 and the efflux transporter P-glycoprotein. St John's wort is the prototype: its bioactive constituent hyperforin is one of the most potent PXR agonists known in clinical pharmacology, with CYP3A4 induction magnitude correlating linearly with hyperforin content of the preparation Nicolussi et al. 2020. In a 14-day dosing trial in 8 healthy volunteers, 900 mg/day St John's wort increased duodenal CYP3A4 expression 1.5-fold and hepatic CYP3A4 functional activity 1.4-fold Nicolussi et al. 2020. Induction is dose-dependent, takes 1–2 weeks to develop, and persists 1–2 weeks after discontinuation as enzyme protein turns over.
CYP inhibition. Goldenseal (containing berberine and hydrastine) and grapefruit furanocoumarins (bergamottin, 6',7'-dihydroxybergamottin) are mechanism-based inhibitors of intestinal CYP3A4. Goldenseal extract (~1 g three times daily for 14–28 days) reduces CYP3A4 and CYP2D6 activity by 40–60% in healthy volunteers Gurley et al. 2008. A single 200 mL serving of grapefruit juice reduces intestinal CYP3A4 protein by ~50% within 4 hours Lown et al. 1997; inhibition lasts up to 24 hours after one dose and up to 72 hours after multiple daily intake Bailey et al. 2013.
P-glycoprotein modulation. St John's wort, garlic, and to a lesser extent ginseng induce intestinal P-gp via PXR. Garlic extract reduced saquinavir AUC by 51%, Cmax by 54%, and 8-hour trough by 49% in a controlled 21-day trial, mechanism attributed to intestinal CYP3A4/P-gp induction by allicin-derived constituents Piscitelli et al. 2002. P-gp induction is the main mechanism behind the SJW–digoxin interaction (digoxin is metabolized minimally but transported heavily by P-gp): 14 days of SJW comedication reduced digoxin AUC by ~25% and trough by ~37% in healthy volunteers, and the effect tracked hyperforin content of the preparation Mueller et al. 2004.
Pharmacodynamic interactions. Three distinct pharmacodynamic axes carry clinically important interactions:
- Antiplatelet additivity. Ginkgo biloba, garlic, ginger, high-dose vitamin E, and high-dose fish oil inhibit platelet aggregation to varying degrees, and may add to aspirin / clopidogrel / NSAID effects. Mechanism is heterogeneous (ginkgolide B is a platelet-activating-factor antagonist; allicin inhibits thromboxane synthesis; EPA competes with arachidonic acid in the thromboxane pathway) Tan & Lee 2021.
- Vitamin K – warfarin pharmacodynamic antagonism. Warfarin inhibits vitamin K epoxide reductase; dietary vitamin K (phylloquinone in leafy greens, menaquinone in natto) directly bypasses the block. A weekly change of 714 μg dietary vitamin K shifts INR by 1 unit in regression models Violi et al. 2016.
- Serotonergic synergy. 5-HTP (a serotonin precursor), SAMe (a methyl donor whose metabolism feeds monoamine pathways), and St John's wort (weak SSRI/MAO-A activity from hyperforin/hypericin) raise central serotonin and add to SSRIs, SNRIs, MAOIs, tramadol, dextromethorphan, triptans and linezolid. Serotonin syndrome is the syndromic endpoint: agitation, hyperreflexia, clonus, hyperthermia, autonomic instability Boyer & Shannon 2005.
Chelation / absorption. Divalent and trivalent cations (Ca²⁺, Mg²⁺, Fe²⁺/³⁺, Al³⁺, Zn²⁺) form insoluble complexes in the GI lumen with fluoroquinolones, tetracyclines, bisphosphonates, and levothyroxine, dropping bioavailability by 30–90% depending on cation and dose Manakova et al. 2021 Wiesner et al. 2021. Mechanism is well characterized: the 4-oxo and adjacent carboxyl group on every fluoroquinolone, and the β-diketone system of tetracyclines, chelate the metal cation; the resulting complex is poorly permeable across enterocyte membranes.
Evidence
The evidence base is heterogeneous: a small number of interactions have rigorous controlled trials, many rely on case reports and pharmacokinetic studies in healthy volunteers, and a tail of folk-wisdom claims has thin or contradictory data.
Tier-1: established by controlled trial. The St John's wort family of interactions is the best-documented herbal interaction in clinical pharmacology. The Piscitelli indinavir trial (n=8, 14 days SJW 900 mg/day) reduced indinavir AUC by a mean 57% and trough by 81%; the FDA and CDC issued physician advisories Piscitelli et al. 2000. Jiang's three-way crossover (n=12, 25 mg single-dose warfarin alone vs after 14 days SJW vs ginseng) showed SJW increased S-warfarin clearance by 29%, R-warfarin clearance by 23%, and reduced INR AUC by 21% — evidence of CYP2C9, CYP1A2 and CYP3A4 induction Jiang et al. 2004. Mai's renal-transplant trial (n=10, 600 mg SJW for 14 days) cut dose-corrected tacrolimus AUC from 180 to 76 ng·h/mL; tacrolimus dose had to be increased from 4.5 to 8.0 mg/day to maintain therapeutic trough Mai et al. 2003. The Mueller digoxin trial (n=96, 10 different SJW preparations) localized the effect to hyperforin-containing extracts and quantified the AUC reduction at ~25% Mueller et al. 2004. Grapefruit juice interactions are likewise extensively trialed: Bailey's 2013 review tabulates 102 RCTs across ~85 drugs Bailey et al. 2013; the foundational felodipine study demonstrated mechanism via downregulated intestinal CYP3A protein Lown et al. 1997.
Tier-2: case reports + plausible mechanism + supportive observational data. SJW-cyclosporine transplant rejection is supported by >10 published case reports — most strikingly two acute heart-transplant rejections in patients with previously stable cyclosporine troughs after starting SJW for depression, with cyclosporine returning to therapeutic range after SJW discontinuation Ruschitzka et al. 2000. SJW-oral-contraceptive interaction is supported by the Schwarz case report (unwanted pregnancy on stable combined OCP after SJW initiation) Schwarz et al. 2003, ~19 UK MHRA Yellow Card reports, and a mechanism trial showing 44% of women on COC+SJW developed ovarian follicles >30 mm vs 16% on COC alone. Garlic-saquinavir is established by controlled PK trial, but the broader garlic-warfarin pharmacodynamic interaction rests on case reports plus mixed crossover data Piscitelli et al. 2002 Tan & Lee 2021. The Tan & Lee 2021 systematic review catalogued 149 articles describing 78 distinct foods/herbs/supplements interacting with warfarin: 58% potentiated, 30% inhibited, 13% had limited impact; bleeding events ranging from gum bleeding to fatal intracranial hemorrhage were attributed to ~20 herbs/supplements Tan & Lee 2021.
Tier-3: theoretical concern with weak or negative trial data. The popular concern that fish oil causes clinically meaningful bleeding additive to warfarin or antiplatelets is not supported by recent meta-analysis: pooled data from 11 RCTs (120,643 patients) found no significant increase in overall, hemorrhagic stroke, intracranial or GI bleeding (RR 1.09, 95% CI 0.91–1.31), with a modest signal only at high-dose purified EPA (~4 g/day icosapent ethyl) — and even there the absolute risk increase is 0.6% and unrelated to antiplatelet background therapy Javaid et al. 2024. The ginkgo-bleeding story shows a similar pattern: 18-RCT meta-analysis (n=1985) found no effect of standardized Ginkgo biloba extract on platelet aggregation, PT, aPTT, or fibrinogen, despite a large case-report literature claiming bleeding events Kellermann et al. 2011. The vitamin-K-and-warfarin story is more nuanced: dietary vitamin K does shift INR (regression model: 714 μg/week → 1 INR unit), but low-dose supplementation (~150 μg/day) appears to stabilize INR in patients with previously erratic control rather than antagonize anticoagulation Violi et al. 2016.
Tier-4: regulatory / observational data on chelation and absorption. The levothyroxine-mineral interaction is supported by 63-study systematic review: simultaneous calcium or iron reduces levothyroxine AUC by 20–50%, with smaller effects from soy, fiber, and coffee; the chelation mechanism is direct and predictable, and the clinical workaround (4-hour separation) is well-established Wiesner et al. 2021 Skelin et al. 2017. Fluoroquinolone-cation chelation has been demonstrated for every cation tested; the regulatory advice is 2 hours before or 6 hours after the antibiotic Manakova et al. 2021.
The niacin-statin interaction. Pharmacodynamic rather than pharmacokinetic: both niacin and statins independently cause skeletal myopathy at low frequency; combined use multiplies that risk. The HPS2-THRIVE trial (n=25,673) found adding 2 g/day extended-release niacin to simvastatin 40 mg increased definite myopathy rates ~4-fold overall and ~10-fold in Chinese patients, alongside increased diabetes, GI events, infections and bleeding HPS2-THRIVE 2014. AIM-HIGH (n=3414) reached similar conclusions on net harm and was stopped early for futility AIM-HIGH 2011. Niacin is sold over-the-counter as a B-vitamin supplement and as "no-flush niacin" at multi-gram doses; the trial evidence sits squarely in the "drug-dose B-vitamin causing real harm in statin users" category.
Practice / clinical guidance
Specialty consensus is consistent across transplant medicine, HIV care, anticoagulation clinics, and anesthesiology preoperative guidelines:
- Transplant centers uniformly screen for St John's wort, garlic supplements, and grapefruit consumption pre-listing and at each clinic visit; SJW is on most transplant pharmacies' "never combine" list alongside antifungals like fluconazole Nicolussi et al. 2020.
- HIV clinics screen for SJW, garlic and other PXR-active botanicals on every protease-inhibitor or integrase-inhibitor prescription Piscitelli et al. 2000 Piscitelli et al. 2002.
- Anticoagulation clinics counsel warfarin patients to keep vitamin K intake stable (range 90–150 μg/day; avoid wild swings, not avoid greens entirely) Violi et al. 2016, and to disclose every new supplement at every INR check.
- Preoperative guidelines (American Society of Anesthesiologists, AAOS) advise discontinuing ginkgo, garlic, ginseng, vitamin E, and high-dose fish oil 1–2 weeks before elective surgery, though the controlled-trial evidence for clinically meaningful intraoperative bleeding is weaker than the case reports suggest Kellermann et al. 2011.
- Endocrinology instructs levothyroxine patients to take the tablet on an empty stomach 30–60 minutes before food, and to separate calcium / iron / multivitamins with minerals / proton pump inhibitors by at least 4 hours Wiesner et al. 2021.
- Psychiatry warns against SAMe, 5-HTP, SJW, and tryptophan in patients on SSRIs, SNRIs, MAOIs, tramadol, dextromethorphan, or triptans Boyer & Shannon 2005.
Community / lay evidence
Public awareness is substantially below clinical awareness. The Qato 2016 JAMA Internal Medicine analysis of two NSHAP cohorts (2005 vs 2011) of community-dwelling US adults aged 62–85 found concurrent use of ≥5 prescription medications and supplements rose from 53% to 67%, and ~15% of older adults were at risk for at least one major drug-drug or drug-supplement interaction Qato et al. 2016. UK community surveys put concurrent prescription + supplement use among ≥65-year-olds at ~34%, with females nearly twice as likely as males to be concurrent users. The chronic gap is patient-clinician communication: most surveys show under 40% of supplement users tell their primary-care physician what they take, and only a minority of physicians routinely ask. Supplement industry packaging carries "consult your doctor" boilerplate but typically not specific drug-class warnings — the SJW–contraceptive failure pathway in particular is rarely flagged on consumer-facing labeling outside Germany.
Failure modes
Common patterns in the case-report literature:
- Self-medication for depression on prescription medication. The Schwarz 2003 pregnancy and the Ruschitzka heart-rejection cases share a structure: stable on prescription drug, started SJW for low mood, did not tell the prescribing physician, presented with rebound disease or unintended pregnancy weeks later Schwarz et al. 2003 Ruschitzka et al. 2000.
- Coincident calcium and levothyroxine. Patients re-stabilize on levothyroxine, then start calcium for bone health, then present months later with rising TSH and fatigue. Separation by >4 hours fixes the problem Wiesner et al. 2021.
- Multivitamin with iron + ciprofloxacin. Single missed timing causes treatment failure in UTI; rarely catastrophic but a leading reason for clinical relapse Manakova et al. 2021.
- Tramadol + SSRI + 5-HTP "for sleep." Subset of serotonin syndrome cases in the toxicology literature.
- Grapefruit + simvastatin. Daily grapefruit at breakfast plus evening simvastatin can raise statin exposure 10-fold; classic presentation is unexplained muscle pain or rhabdomyolysis weeks after starting the juice Bailey et al. 2013.
Misconceptions
- "Natural means safe to combine." The strongest documented herbal interactions (SJW, grapefruit, goldenseal, garlic) involve plant compounds that are pharmacologically active at very low concentrations precisely because they hijack the same regulatory pathways as prescription drugs.
- "Fish oil thins your blood — stop it before surgery." The 2024 JAHA meta-analysis of 120,643 patients found no increased bleeding risk at typical doses (1–2 g/day), and no increase even when co-administered with antiplatelets Javaid et al. 2024. The instruction lingers in preoperative checklists.
- "Avoid all vitamin K if you're on warfarin." The data favor consistent moderate intake over avoidance; abrupt avoidance destabilizes INR as much as abrupt over-consumption Violi et al. 2016.
- "Niacin is just a vitamin." At cardiovascular doses (≥1 g/day) niacin is a pharmacological agent with statin-interaction risk demonstrated in two outcome trials covering 29,000 patients HPS2-THRIVE 2014 AIM-HIGH 2011.
- "Ginkgo causes bleeding." Case reports yes, controlled trial meta-analysis no; the case-report signal may be confounded by concurrent aspirin or warfarin in the same elderly population Kellermann et al. 2011.
Audience
Three populations carry disproportionate risk and warrant separate emphasis:
- Solid-organ transplant recipients on cyclosporine, tacrolimus, sirolimus, mycophenolate. SJW alone has caused documented acute rejection at 4–8 weeks of co-administration Ruschitzka et al. 2000 Mai et al. 2003.
- Patients on warfarin. Narrow therapeutic index plus >78 catalogued interacting foods/herbs/supplements Tan & Lee 2021.
- Older adults on ≥5 medications. Polypharmacy plus high supplement use plus age-related pharmacokinetic changes amplify every interaction; ~15% are at risk for a major interaction in any given year Qato et al. 2016.
Secondary at-risk groups: women on hormonal contraception (SJW-OCP failure documented in case series); patients on HIV antiretrovirals (any PXR-active supplement risks viral rebound) Piscitelli et al. 2000; patients on hypothyroid replacement (chelation pathway); patients on SSRIs/SNRIs (serotonergic pathway).
Contraindications
The interactions themselves act as contraindications when the drug is required and the supplement is not. Hard "do not combine" pairs supported by trial-level or strong case-report evidence:
- St John's wort + cyclosporine, tacrolimus, sirolimus, indinavir, any protease inhibitor, any NNRTI, any hormonal contraceptive, warfarin, digoxin, anticonvulsants, oral oncology agents (imatinib, irinotecan).
- Grapefruit/grapefruit juice + simvastatin, lovastatin, atorvastatin (less so), felodipine, nifedipine, amiodarone, certain calcium-channel blockers, tacrolimus, cyclosporine.
- 5-HTP, SAMe, SJW + SSRIs, SNRIs, MAOIs (including linezolid), tramadol, meperidine, dextromethorphan, triptans, MDMA.
- Calcium, iron, magnesium, multivitamins-with-minerals + levothyroxine (separate by 4+ hours), oral bisphosphonates (separate by 30–60 min from breakfast, alendronate is empty-stomach), fluoroquinolones and tetracyclines (separate by 2–6 hours).
- Niacin ≥1 g/day + statins (significantly increased myopathy and diabetes risk; no offsetting cardiovascular outcome benefit).
- Licorice (glycyrrhizin-containing) + thiazides, loop diuretics, digoxin, antihypertensives (pseudo-hyperaldosteronism via 11β-HSD2 inhibition raises BP and drops potassium) Omar et al. 2012.
Protocol
The actionable behaviour for the reader is informational: disclose, time-separate, or substitute. There is no dosing protocol per se; the entry's "protocol" is the disclosure ritual.
- Maintain a single written list of every supplement (including doses) in the same place as the prescription list; share it at every clinical encounter.
- Before starting any new prescription drug, search the drug name + each supplement on a reputable interaction checker (e.g., LiverTox, Lexicomp, NIH-funded drug interaction database).
- If on a narrow-therapeutic-index drug (warfarin, anticonvulsants, immunosuppressants, oral chemotherapy, antiretrovirals, levothyroxine, digoxin), do not start any supplement without checking with the prescribing physician.
- For chelation interactions, the workaround is timing: take the prescription drug on an empty stomach, separate mineral supplements by 4 hours, separate fluoroquinolones by 2 hours before / 6 hours after.
Stakes
What goes wrong when this is missed:
- The Ruschitzka heart-transplant case: stable on cyclosporine for years, started SJW for low mood, acute graft rejection 4 weeks later, hospital readmission and IV antirejection therapy Ruschitzka et al. 2000. The kidney-transplant patient in the parallel series returned to dialysis after chronic rejection Mai et al. 2003.
- The Piscitelli HIV cases prefigured viral rebound and protease-inhibitor resistance — a permanent loss of treatment option Piscitelli et al. 2000.
- Statin-grapefruit rhabdomyolysis: 10-fold simvastatin exposure increases the rhabdomyolysis risk from rare to clinically meaningful Bailey et al. 2013.
- Calcium-levothyroxine: subclinical hypothyroidism mimicking depression, fatigue, hair loss, weight gain — months of "the medication isn't working" before the food-timing question is asked Wiesner et al. 2021.
- Levo-fluoroquinolone in pyelonephritis: treatment failure and progression to sepsis.
- Serotonin syndrome from 5-HTP + tramadol + SSRI: hyperthermia, clonus, agitation, ICU admission, occasional fatality Boyer & Shannon 2005.
Credibility range
The optimist case
The mechanistic pathways are real but their clinical penetrance is exaggerated. Most documented cases involve narrow-therapeutic-index drugs (transplant immunosuppressants, antiretrovirals, warfarin) where the receiving population is already heavily monitored — exactly the population in which any small change in drug exposure will be detected as a rejection or INR event. Across the broader prescription pharmacy, the typical adult on an SSRI, antihypertensive, statin, or thyroid replacement has wide enough therapeutic margins that supplement-induced 20–30% AUC changes are pharmacodynamically silent. The Javaid omega-3 meta-analysis demonstrates how strong the "natural therefore dangerous to combine" narrative is: 11 RCTs, 120,643 patients, no bleeding signal — and yet preoperative checklists still routinely tell patients to stop fish oil two weeks before surgery Javaid et al. 2024. The ginkgo case-report literature has a similar structural problem: elderly patients on ginkgo are also on aspirin and warfarin; bleeding gets attributed to the supplement Kellermann et al. 2011. The pragmatic rule — disclose what you take, monitor narrow-therapeutic-index drugs, separate chelating cations by hours — captures essentially all the avoidable clinical risk without requiring the reader to learn pharmacology.
The skeptic case
The interactions are real, the case-report literature substantially under-counts events (passive surveillance captures perhaps 1–10% of attributable events), and the mechanism-based predictions point at a much larger problem than the trial literature has time to characterize. SJW alone induces >30 drugs across CYP3A4, CYP2C9 and P-gp substrates; each interaction is plausible from PXR pharmacology, but only a handful have been formally trialed in humans Nicolussi et al. 2020. The Tan & Lee systematic review found 78 supplements/foods interacting with warfarin alone — and warfarin is one prescription drug among hundreds Tan & Lee 2021. The Qato data show that >50% of older US adults are on ≥5 concurrent medications + supplements; with ~15% at risk for a major interaction, the population-level harm rate is large even if individual case rates are low Qato et al. 2016. The chelation interactions are mechanistically certain and clinically common — every patient taking levothyroxine plus calcium loses 20–50% of their thyroid replacement until they fix the timing Wiesner et al. 2021. The reader who hears "your supplement may interact with your prescription" and shrugs because it sounds like a hedged caveat is the reader who shows up in the case-series literature.
The author's call
The skeptic case wins on the cases that matter most (transplant, HIV, warfarin, narrow-therapeutic-index drugs, anesthesia) and on the absorption interactions that affect millions of routine levothyroxine + antibiotic prescriptions. The optimist case is correct that the population-level interaction signal for typical supplements at typical doses against typical prescriptions is smaller than the case-report literature suggests; fish oil is unfairly maligned, ginkgo's bleeding signal is weak. The article should make four asymmetric points: (1) the St John's wort family of interactions is a serious clinical hazard wherever it touches a narrow-therapeutic-index drug; (2) the chelation interactions are universal, predictable, and trivially fixable by timing; (3) the serotonergic interactions matter when readers self-stack mood supplements on top of antidepressants; (4) most fear-of-bleeding supplement warnings are not supported by controlled data, but the few that are (grapefruit-statin, niacin-statin, licorice-antihypertensive) are real. The reader's action is awareness and disclosure, not avoidance of all supplements.
Stakeholder + incentive map
- Supplement industry. US Dietary Supplement Health and Education Act (DSHEA 1994) framework means supplements are regulated as foods, not drugs; no premarket interaction testing required. Industry incentive is to keep "natural = safe" framing intact and avoid specific drug warnings on labels.
- Prescribing physicians and pharmacists. Liability incentive cuts the other way — interaction screening at the pharmacy is a major patient-safety function. But typical 15-minute primary-care visits and crowded community pharmacies frequently miss supplement disclosure.
- Transplant centers, HIV clinics, anticoagulation clinics. Specialty pharmacies are aggressive screeners; this is where the strongest interaction-counseling protocols exist.
- Pharmacovigilance bodies. FDA MedWatch, UK MHRA Yellow Card, German BfArM publish case-report aggregates; the SJW-OCP pregnancies were aggregated this way Schwarz et al. 2003.
- Wellness influencer culture. Strong incentive to recommend stacks (fish oil + magnesium + vitamin D + ashwagandha + 5-HTP + SAMe) without surfacing prescription interactions, since most followers don't disclose their prescriptions to the influencer or to each other.
Population variability
- Age. Older adults carry the highest interaction burden: more prescriptions, more supplements, reduced hepatic and renal clearance, and reduced physiologic reserve when interactions tip drug levels out of range Qato et al. 2016.
- Genotype. CYP2D6 and CYP2C9 poor metabolizers are more susceptible to inhibitor interactions (smaller margin of safety). MDR1 polymorphisms modulate P-gp induction by SJW (the 1236C>T / 2677G>T/A / 3435C>T haplotype attenuates induction).
- Ethnic background. The HPS2-THRIVE niacin-statin myopathy rate was ~10x higher in Chinese participants than European participants on identical doses; mechanism likely involves SLCO1B1 transporter genotype frequency HPS2-THRIVE 2014.
- Baseline status. Patients with stable trough levels (transplant, warfarin) detect interactions fastest; patients with episodic prescriptions (course of antibiotics) may suffer treatment failure without ever attributing it to the supplement Tan & Lee 2021.
- Sex. UK community surveys find females nearly twice as likely as males to be concurrent prescription + supplement users; women are also disproportionately exposed to the SJW-OCP pathway.
Knowledge gaps
- The DOAC class (apixaban, rivaroxaban, dabigatran, edoxaban) interactions with supplements are less characterized than warfarin's, despite DOACs now being the dominant oral anticoagulant. CYP3A4 and P-gp substrate behavior makes SJW a plausible inducer, but trial data are limited.
- Oral oncology agents (imatinib, ibrutinib, osimertinib, palbociclib) are CYP3A4 substrates with narrow therapeutic windows; supplement interaction data are mostly modelled, not trialed Polasek et al. 2011.
- Population-level outcome data on supplement-drug interactions (versus case-report or PK trial data) are sparse; pharmacoepidemiology has not run the equivalent of the Qato analysis with bleeding/rejection/treatment-failure endpoints Qato et al. 2016.
- Cannabinoid (CBD, THC) supplement interactions with CYP2C9 (warfarin) and CYP3A4 (tacrolimus, clobazam) are increasingly important; case reports are mounting but trial-level evidence is thin.
- Many traditional Chinese and Ayurvedic preparations have multiple active constituents whose individual PXR / CYP / P-gp profiles have never been characterized.
Scope vs brief. The brief named four mechanistic pathways (CYP, P-gp, anticoagulation, serotonergic) and four consequence types (drug levels, efficacy, bleeding risk, toxicity). All four pathways and all four consequence types are covered in the article, anchored by the marquee trials for each. The chelation pathway (calcium/iron/Mg with levothyroxine, fluoroquinolones, tetracyclines) was added beyond the brief because it is the most clinically common drug-supplement interaction in primary care and would be a notable omission for a reader on thyroid replacement; framed honestly in mechanism §4.
- Action chosen:
know, notdecideorrespond. The dominant reader action is awareness leading to disclosure. For the narrow-therapeutic-index subset (transplant, HIV, warfarin, lithium, oral chemo), the action shades intodecide(with the prescriber), but most readers will use this entry as background.responddidn't fit because the entry is preventive, not symptom-triggered. - Cadence:
as-needed. Triggered by starting any new prescription or supplement. Consideredonce(read the entry once, internalize it) but the relevant moment recurs. - Evidence score 4, not 5. The marquee interactions (SJW family, grapefruit-statins, niacin-statins, chelation pathway) have Cochrane-level support and consistent specialty consensus. A 5 would imply the same level of certainty across the broader supplement-drug landscape, where most pairs rest on mechanism-based prediction and case reports rather than completed RCTs. The PXR-induction pharmacology is predictive but the in-human trial tail is sparse for the long list of drug substrates.
- Bleeding-risk misconception handling. The fish-oil and ginkgo bleeding myths were given prominent placement in misconceptions because clinical practice still over-restricts these supplements preoperatively. The article should not be dismissed by surgical teams as wellness-friendly soft pedaling: the 2024 Javaid meta on omega-3 (n=120,643) and Kellermann ginkgo meta (n=1985) are the strongest data we have. If the bleeding claim moves in subsequent meta-analyses, this section is the one to update first.
- Excluded: drug-food interactions outside of supplements proper. Cruciferous vegetables and CYP1A2 induction, charbroiled meats, soy and thyroid, dairy as a tetracycline source. Pointed at via
out-of-scope. Including them would have doubled the contraindications callout list without adding to the supplement-specific story. - Excluded: cannabis / CBD interactions. Mounting case reports on CBD + tacrolimus, warfarin and clobazam, but trial-level evidence is thin and the legal/clinical framing is its own entry. Flagged in research dossier knowledge gaps. Future-link candidate when a CBD entry exists.
- Excluded: DOAC-supplement interactions. Apixaban, rivaroxaban, dabigatran, edoxaban are now the dominant oral anticoagulants but interaction data with supplements are dominated by mechanism-based prediction (CYP3A4 + P-gp substrates) rather than completed trials. Acknowledged in research §knowledge gaps. Will become a separate entry when DOAC anticoagulation gets its own coverage.
- Audience scoring nuance. Energy at 1 and mood at 1 are deliberate. The levothyroxine-chelation pathway is mechanistically certain and clinically common but only relevant to the subset on thyroid replacement; same logic for mood (SSRI/SJW). Scoring higher would over-promise for the typical reader who is on none of these drugs; scoring 0 would silence the audience that needs the entry most. The compromise lands at 1 and is named in justifications.
- Rating difficulty. This is an awareness entry, not an intervention; the benefit dimensions get scored against harm prevented rather than benefit conferred. The 0–5 ladder is calibrated for interventions, so the scoring required interpretation. The justifications carry the load — the reader-facing pitches translate the dimension into harm-prevented language.
- Separate-entry candidates surfaced during writing: Grapefruit and Statins (warrants its own entry — the interaction is large enough and the prescription population broad enough); St John's Wort (the supplement itself has a benefit profile worth covering, with interactions as one consequence among several); Niacin for Cardiovascular Risk (the AIM-HIGH / HPS2-THRIVE story alone supports a stand-alone entry); Levothyroxine Absorption Timing (the chelation and food-timing details warrant a dedicated entry in the thyroid track).
- Future-link candidates once those entries exist: St John's wort, fish oil, magnesium, vitamin D, creatine, niacin, levothyroxine, warfarin, statins, grapefruit, SSRIs as a class.
Drug-Supplement Interactions
Maintaining a single supplement-and-prescription list and sharing it at clinical encounters is a minor lifestyle shift. The chelation interactions add a one-time learning step (separate by 4 hours) that becomes automatic.
The marquee interactions rest on controlled crossover pharmacokinetic trials (Piscitelli 2000 indinavir; Jiang 2004 warfarin; Mai 2003 tacrolimus; Mueller 2004 digoxin; Piscitelli 2002 garlic-saquinavir; Lown 1997 felodipine), large outcome trials (HPS2-THRIVE 2014 n=25,673; AIM-HIGH 2011), and systematic reviews (Tan & Lee 2021; Wiesner 2021; Violi 2016; Javaid 2024 meta of 120,643 patients; Kellermann 2011). Specialty consensus is consistent across transplant, HIV, anticoagulation, anesthesiology and endocrinology guidelines. Falls short of 5 because the broader supplement-drug landscape is dominated by mechanism-based predictions rather than completed trials.
Awareness averts identifiable short-term harms: bleeding events on warfarin (Tan & Lee 2021 catalogued 78 interacting foods/herbs/supplements), statin myopathy from grapefruit (Bailey 2013) or co-prescribed niacin (HPS2-THRIVE 2014), and chronically undertreated hypothyroidism from co-administered calcium or iron (Wiesner et al. 2021 meta-review). For the typical reader on no narrow-therapeutic-index drug, no felt effect; for those on warfarin, levothyroxine, antiretrovirals or immunosuppressants, the prevented harm is the entire benefit.
Specific catastrophic events are demonstrably prevented by awareness: acute transplant rejection from St John's wort lowering cyclosporine/tacrolimus levels (Ruschitzka 2000 Lancet heart-rejection cases; Mai 2003), antiretroviral failure and protease-inhibitor resistance (Piscitelli 2000 Lancet, 57% indinavir AUC drop), rhabdomyolysis from grapefruit-potentiated statins (Bailey 2013), serotonin syndrome from stacked serotonergic supplements (Boyer & Shannon 2005 NEJM). Not population-wide mortality reduction; in the affected subset, life-altering.
The chelation pathway — calcium, iron, magnesium taken with levothyroxine reducing thyroxine absorption by 20–50% (Wiesner et al. 2021) — is a common, under-recognized cause of persistent fatigue in supplemented hypothyroid patients. Awareness and 4-hour timing separation restore baseline energy. Modest population-level effect.
Two specific mood-relevant pathways: St John's wort blunts SSRI response and antidepressant levels via CYP induction (Nicolussi 2020; Jiang 2004 showed CYP2C9/3A4 induction); and 5-HTP / SAMe / SJW added to SSRIs, SNRIs, MAOIs, tramadol, dextromethorphan or triptans risks serotonin syndrome (Boyer & Shannon 2005 NEJM). Awareness preserves antidepressant efficacy and prevents the worst self-stacking accidents.