Lovage's Furanocoumarins and the Photosensitivity Question

Lovage root's furanocoumarin content is one of the few safety claims on this site that gets more precise, not less, the harder you look. The chemistry is real and lovage-specific: modern LC-MS work has identified and quantified the exact compounds, one of which is also a licensed prescription drug used deliberately to sensitize skin to light. The human case-report evidence, though, is thinner and more route-specific than the "celery-family plants cause sunburn" shorthand suggests — and Europe's own medicines regulator, after formally reviewing the question, concluded the concern does not hold up for the oral tea that lovage root is actually used to make.


Table of Contents

  1. What Is Actually in the Root: the Chemistry
  2. The Compound That Is Also a Prescription Drug
  3. The Documented Human Cases Involving Lovage Itself
  4. The Wider Apiaceae Family: Celery, Parsnip, and Giant Hogweed
  5. Why the Route of Exposure Changes the Answer
  6. Doing the Arithmetic, With the Assumptions Stated
  7. A Genotoxicity Signal in the Laboratory
  8. What the Findings Mean for Tea, Oil and Harvest Exposure
  9. Key Research Papers
  10. Connections
  11. Featured Videos

What Is Actually in the Root: the Chemistry

Furanocoumarins (also spelled furocoumarins) are a coumarin subclass built around an extra fused furan ring, and that ring is what makes them photoactive: on absorbing ultraviolet A (UVA) radiation, they form covalent bonds with DNA and with skin-cell membrane components, which is the molecular basis of both their photosensitizing and their photomutagenic activity. Lovage root has now been analyzed for these compounds several times, by independent groups, with results that agree on the broad picture even where the specific numbers differ.

A 2022 LC-MS and validated HPLC study identified 25 metabolites in fresh lovage root, with coumarins as the dominant compound class. Four major coumarins were isolated and quantified by name: apterin, xanthotoxin, isopimpinellin, and pimpinellin. Total coumarin content ranged from 1.7–2.9 mg/g in fresh root to a concentrated 15–24 mg/g in dried root — drying removes water weight but not the compounds themselves, so the concentration rises substantially in the dried form used to make tea. An earlier LC-DAD analysis of chloroform extracts of lovage root, cited in the EU regulatory assessment of the plant, separately detected psoralen and bergapten as the furanocoumarins present. The two analyses do not report identical compound lists — a normal feature of natural-product chemistry, where extraction solvent, cultivar, and analytical method all shift which minor compounds are detected and quantified — but every published chemical analysis of lovage root agrees that furanocoumarins of the photoactive kind are genuinely present, not a borrowed assumption from a related plant.

Back to Table of Contents

The Compound That Is Also a Prescription Drug

One of the four major coumarins identified in lovage root, xanthotoxin, is better known in medicine by its pharmaceutical name: methoxsalen, also called 8-methoxypsoralen (8-MOP). This is not an obscure phytochemical curiosity — it is a licensed drug, taken orally or applied topically before controlled ultraviolet A exposure, in PUVA photochemotherapy for psoriasis, vitiligo, and cutaneous T-cell lymphoma. A treatment history spanning decades confirms exactly the mechanism this page is describing: methoxsalen intercalates into DNA and forms ultraviolet A-activated crosslinks in skin cells, which is therapeutically useful for suppressing overactive immune and proliferative skin processes and is the same molecular event responsible for phototoxic sunburn-like reactions at uncontrolled doses.

This is the single strongest piece of evidence on this page, applying a method used elsewhere on this site for other herbs: when the proposed mechanism has already been developed into an approved drug, you can look up exactly how potent that mechanism is, rather than guessing. Methoxsalen's photoactivating effect is well-characterized, dose-dependent, and requires ultraviolet A exposure. Without light, the EU assessment report states, "the toxicity of furanocoumarins is low, with an LD50 in mammals of 300 to 600 mg/kg body weight" — low, not absent. That fact reframes the whole safety question: the concern is not "lovage root contains a toxin," it is "lovage root contains a compound identical to a photoactive drug, so both benefit and hazard depend on light exposure."

Back to Table of Contents

The Documented Human Cases Involving Lovage Itself

More than one published case report names lovage specifically. A 2013 case in the journal Contact Dermatitis describes a 31-year-old woman who developed sharply demarcated, red, scaly patches on her arms and shoulder two weeks after applying lovage essential oil (alongside a second, unrelated essential oil) to her skin; the reaction resolved with topical and systemic steroids. Two further points matter for interpreting this case honestly. First, it involved the concentrated essential oil applied directly to skin, not a cup of root tea taken by mouth — a real difference in dose and route, discussed further below. Second, the case report's own title describes it as "contact dermatitis," and the published report's title and PubMed record (which carries no abstract) do not specify whether the reaction was confirmed as strictly phototoxic (light-dependent) versus a non-photo-dependent allergic contact reaction to another oil constituent; both are plausible for an essential oil applied topically, and the site's own doctrine is to state what is confirmed rather than assume the more dramatic mechanism.

An earlier case is indexed in PubMed under its own title: a 1992 report in Contact Dermatitis by Ashwood-Smith and colleagues, "Photosensitivity from harvesting lovage (Levisticum officinale)" (PubMed holds no abstract for it). The EU regulatory assessment of lovage root describes the harvest reaction: extensive handling of lovage during harvest under prolonged exposure to strong sunlight brought on dermatitis within a few hours, with itching and redness (erythema); after 36 hours, blisters formed on the exposed arms and legs, with marked darkening of the skin (hyperpigmentation) after 3 weeks. The assessment credits that description to Wolf (1995), and credits Ashwood-Smith and colleagues (1992) with measuring appreciable amounts of furanocoumarins in the fresh lovage specimens: 3.12 and 4.02 micrograms per gram wet weight of psoralen and 5-methoxypsoralen, respectively. It also notes a similar case of dermatitis after contact with lovage oil (Vollman 1988). The Wolf and Vollman sources could not be retrieved for this page and are reported here via the regulatory assessment that cites them. What these accounts share is instructive: every documented case involves either concentrated essential oil on skin, or prolonged occupational handling of the raw plant in strong sun — not a cup of tea made from the dried root.

Back to Table of Contents

The Wider Apiaceae Family: Celery, Parsnip, and Giant Hogweed

Lovage's furanocoumarin chemistry is not unusual within its own botanical family — it is typical of it. Apiaceae (the carrot/celery family, which also includes parsley, fennel, dill, and angelica) is one of the most common phototoxic plant families, which a 2024 review lists as Apiaceae, Rutaceae (citrus, rue) and Moraceae (the fig family). Recent dermatology case literature documents this repeatedly and recently: celery (Apium graveolens) causing phytophotodermatitis severe enough to be described in a 2025 clinical review as "not as bland as you may think"; parsnip mash causing a perioral reaction reported the same year; wild parsnip cases from agricultural and recreational exposure; and giant hogweed (Heracleum mantegazzianum) — by far the most severe member of the family, capable of second-degree burn-like injury — appearing repeatedly in the emergency and toxicology literature. A 2024 general review of phytophotodermatitis mechanisms places all of these, lovage included by chemical class, on the same furanocoumarin-driven mechanism described above.

This context matters for calibrating lovage specifically: it sits within a well-established hazard family, chemically, but the clinical case literature is overwhelmingly concentrated on giant hogweed (by far the most dangerous), celery, and parsnip — all more widely handled in bulk, gardened, or foraged than lovage, which explains at least part of why lovage-specific cases are so much rarer in the literature. Rarity of case reports is not the same as absence of mechanism; the mechanism is real and shared across the family. But it is also not the same as lovage carrying giant hogweed's severity, and nothing in the literature suggests it does.

Back to Table of Contents

Why the Route of Exposure Changes the Answer

This is the finding that most changes how this page should be read, and it comes from the primary source with the most reason to take the hazard seriously rather than minimize it: Europe's own medicines regulator, reviewing lovage root specifically for a therapeutic monograph. The HMPC's assessment report considered the furanocoumarin question directly, alongside the chemistry and case literature summarized above, and its clinical safety conclusion states plainly: "Concerns regarding phototoxicity are not supported by clinical data or pharmacovigilance signals to be relevant for the oral use of lovage root as recommended in the monograph." The same document elsewhere calls photoactivation by ultraviolet (UV) radiation "a concern" at the level of chemical hazard and non-clinical toxicology, so this is not a regulator waving the question away — it reviewed the mechanism, reviewed the case reports, and specifically distinguished the oral tea route, which is what its 2–3 g twice-daily dosing recommendation covers, from the exposure routes that actually appear in the case reports: concentrated essential oil on skin, and prolonged raw-plant handling in strong sun.

This is a route-substitution distinction running in the protective direction, the same logic this site applies elsewhere when a hazard demonstrated for an injected or topically concentrated form does not automatically transfer to a diluted oral preparation. It does not mean the essential oil is safe on skin in sunlight — the 2013 case report says otherwise — and it does not mean furanocoumarins vanish when swallowed. It means the specific concentration and skin-contact pathway that produces a photosensitization reaction is largely absent when the same compounds are diluted into hot water, drunk, absorbed through the gut, and processed by the liver before reaching skin at all.

A 2025 review of celery makes the same route point for a close relative of lovage: when celery is eaten as part of a typical diet, the amount of furocoumarins in it generally is not enough to cause phytophotodermatitis, while people who regularly handle celery, eat produce that is infected with fungus, or use tanning beds are at increased risk.

Back to Table of Contents

Doing the Arithmetic, With the Assumptions Stated

It is possible to sanity-check the HMPC's conclusion with rough numbers, and doing so is more persuasive than restating the conclusion alone — provided the assumptions are stated plainly, since this arithmetic chains together figures from more than one secondary source rather than a single measured study.

The EU's food-safety authority, in a 2009 compendium of botanicals with substances of concern (cited within the same HMPC assessment report), lists specific furanocoumarin concentrations (printed in the assessment report under the label "root seeds", so it is not clear they are dried-root values): imperatorin 12.82 mg/kg, 5-methoxypsoralen 6.38 mg/kg, psoralen 3.8 mg/kg, and 8-methoxypsoralen (methoxsalen/xanthotoxin) 0.5 mg/kg — a combined total of roughly 23.5 mg per kilogram of plant material for these four named compounds. The HMPC's recommended average daily dose is 4–6 g of dried root. Multiplying those figures gives a rough total furanocoumarin exposure, from these four compounds, of approximately 0.1–0.14 mg per day from a standard therapeutic dose of tea.

Compare that to the toxicological threshold the same assessment report cites for furanocoumarin phototoxicity in the presence of ultraviolet light: a lowest-observed-adverse-effect level of 0.14–0.38 mg per kilogram of body weight. For a 70 kg adult, that threshold works out to roughly 10–27 mg. Every number in this paragraph is drawn from the regulatory assessment report rather than independently re-measured for this page, the underlying compendium figures may not distinguish fresh from dried weight precisely, and the LOAEL itself is a general mammalian furanocoumarin figure rather than one derived specifically from lovage — so this is an illustrative order-of-magnitude check, not a precise safety margin. It points in the same direction as the HMPC's own clinical conclusion rather than contradicting it, which is exactly what a sanity check should do.

Back to Table of Contents

A Genotoxicity Signal in the Laboratory

One further finding from the regulatory assessment deserves an honest mention rather than omission, per this site's rule against manufacturing false reassurance by leaving out an inconvenient result. Bergapten and a lovage root extract both showed photomutagenic activity — ultraviolet-dependent gene mutation — in a mutant strain of the green alga Chlamydomonas reinhardtii, a standard photomutagenicity screening organism, under long-wave ultraviolet exposure in laboratory conditions. This is a real, ultraviolet-dependent genotoxicity signal for the same compound class discussed throughout this page. It is also, explicitly, an algal screening assay, not a mammalian or human finding, and the same assessment report states directly that no published data on the carcinogenicity of lovage root itself could be found in either direction. The honest position is that this is a mechanistic plausibility signal consistent with everything else on this page — ultraviolet-dependent DNA reactivity is exactly what a furanocoumarin is expected to do — and not a demonstrated human cancer risk, which has simply never been studied for this plant.

For the drug itself, the human evidence comes from medical treatment at drug doses with repeated light exposure. A 30-year study followed 1,380 patients with psoriasis first treated with psoralen plus ultraviolet A (PUVA) in 1975 to 1976. The risk of developing one or more squamous cell skin cancers in a year was strongly linked to the total number of treatments: for 350–450 treatments compared with fewer than 50, the incidence rate ratio was 6.01. Fewer than 150 treatments had, at most, modest effects on squamous cell cancer risk, and even high-dose treatment did not greatly increase the risk of basal cell skin cancer. These are results of repeated drug treatment combined with ultraviolet A, not of lovage.

The EU assessment report draws three further conclusions. Because data on mutagenicity and carcinogenicity are lacking, "a list entry for Levisticum officinale root cannot be recommended." Citing the extract's weak estrogenic activity and the absence of sufficient data, it states that "lovage root use is not recommended during pregnancy and lactation." And it states that "some caution is needed in combination with UV radiation exposure due to possible photoactivation caused by furanocoumarins."

Back to Table of Contents

What the Findings Mean for Tea, Oil and Harvest Exposure

Back to Table of Contents

Key Research Papers

  1. Olennikov DN. Coumarins of Lovage Roots (Levisticum officinale W.D.J.Koch): LC-MS Profile, Quantification, and Stability during Postharvest Storage (2022). Metabolites. — PubMed PMID: 36676932
  2. Lapeere H, Boone B, Verhaeghe E, Ongenae K, Lambert J. Contact dermatitis caused by lovage (Levisticum officinalis) essential oil (2013). Contact Dermatitis. — PubMed PMID: 23948036 — A published human case naming lovage directly.
  3. Gasparro FP. The role of PUVA in the treatment of psoriasis. Photobiology issues related to skin cancer incidence (2000). American Journal of Clinical Dermatology. — PubMed PMID: 11702610 — Xanthotoxin/methoxsalen's mechanism as a licensed photochemotherapy drug.
  4. Pate HF, Hill K, McGovern TW. Not as Bland as You May Think: Celery (Apium graveolens) Commonly Induces Phytophotodermatitis (2025). Cutis. — PubMed PMID: 40315466
  5. Holt A, Gabel C, Flanagan K, Levin NA. Parsnip Phytophotodermatitis (2025). Dermatitis. — PubMed PMID: 39453618
  6. Walling AL, Walling HW. Phytophotodermatitis induced by wild parsnip (2018). Dermatology Online Journal. — PubMed PMID: 29630166
  7. Flanagan KE, Blankenship K, Houk L. Botanical Briefs: Phytophotodermatitis Caused by Giant Hogweed (Heracleum mantegazzianum) (2021). Cutis. — PubMed PMID: 35100530
  8. Downs JW, Cumpston KL, Feldman MJ. Giant Hogweed phytophotodermatitis (2019). Clinical Toxicology. — PubMed PMID: 30739522
  9. Grosu Dumitrescu C, Jîjie AR, Manea HC, et al. New Insights Concerning Phytophotodermatitis Induced by Phototoxic Plants (2024). Life. — PubMed PMID: 39202761
  10. Qu L, Zou W, Wang Y, et al. European regulation model for herbal medicine: The assessment of the EU monograph and the safety and efficacy evaluation (2018). Phytomedicine. — PubMed PMID: 29655689
  11. Miran M, Monsef Esfahani H, Jung JH, et al. Characterization and Antibacterial Activity of Phthalides from the Roots of the Medicinal Herb Levisticum officinale (2020). Iranian Journal of Pharmaceutical Research. — PubMed PMID: 33224223
  12. Ashwood-Smith MJ, Ceska O, Yeoman A, Kenny PG. Photosensitivity from harvesting lovage (Levisticum officinale) (1992). Contact Dermatitis. — PubMed PMID: 1395606
  13. Stern RS; PUVA Follow-Up Study. The risk of squamous cell and basal cell cancer associated with psoralen and ultraviolet A therapy: a 30-year prospective study (2012). Journal of the American Academy of Dermatology. — PubMed PMID: 22264671

Primary Regulatory Source

  1. EMA/HMPC/524623/2011 — Assessment report on Levisticum officinale Koch, radix (2012), Sections 3.3, 3.4, and 5.5 (toxicological and clinical safety conclusions on phototoxicity)

Back to Table of Contents

Connections

Back to Table of Contents