Cilantro: the trials are on the seed, the outbreaks are on the leaf

A scoping review of clinical trials at culinary doses found 142 trials across 14 herbs and spices, coriander among them, and named cinnamon, fenugreek and ginger as the ones carrying both the most trials and the clearest signal for glycaemic control. Coriander was absent from that shortlist, and the review pools nothing and reports no effect size. What sits under the name Coriandrum sativum for the heart is thinner again. Of 3008 records screened, 18 studies were included, 12 of them animal work in vivo and only 2 clinical, so the cardioprotective label rests mostly on rodents and the review gives direction of effect alone. The human trials that exist tested dried seed rather than the fresh leaf you chop. Coriander seed powder at 2 g a day improved the lipid parameters of patients with hyperlipidemia in a single blind trial that randomised 80 of them to four groups of 20, so only 20 people took coriander on its own, against garlic powder, a garlic and coriander mixture and a control, and the trial reports no confidence intervals. Those were patients diagnosed with hyperlipidemia, and nothing in it speaks to a healthy person eating a handful of leaves. A 6 g blend of 13 spices that included coriander lowered postprandial IL-1beta at 240 minutes in 12 men aged 40 to 65 with overweight or obesity, while 2 g of the same blend did nothing. The share belonging to coriander alone was never measured. We found no study of absorption, bioavailability or digestibility for cilantro under any of its names, so that axis stays empty on this page.

How much is actually in it

• Across 242 vegetable samples from northern California farmers markets, 6.6% carried Salmonella with most of the positives in cilantro and squash leaves, and samples from growers who said they farmed organically had a twofold higher probability of contamination. — 242 samples of six Asian-cuisine vegetables from 21 vendors at six farmers markets, summer 2013 (n = 242) Evidence C source
  16 of 242 samples positive for Salmonella at 6.6%, twofold higher probability of Salmonella with stated organic practices, no E. coli O157:H7 detected in any sample

What cooking and storage change

• Chopped cilantro held at room temperature grew Salmonella Thompson faster and to a higher concentration than whole leaves, which is why the same investigation tells kitchens to serve it promptly after chopping. — cilantro samples in the laboratory arm of an outbreak investigation, not people Evidence D source
  Faster growth and higher final concentration on chopped cilantro than on whole leaf at room temperature, with no numeric counts given in the abstract

Safety, limits and interactions

• In a Salmonella Thompson outbreak in California, eating cilantro at a restaurant carried an odds ratio of 3.5 with a 95% confidence interval of 1.1 to 11.4, with 63% of cases exposed against 34% of controls. — 35 sporadic cases and a restaurant-associated outbreak of 41 cases in California, onset 6 to 31 March 1999 (n = 76 cases) Evidence C source
  Odds ratio 3.5, 95% CI 1.1 to 11.4, which is odds and not risk, 3 patients hospitalised
• Coriander can act as a hidden allergen, with a reported case of anaphylaxis in a young woman after drinking coriander-flavoured beer, a drink she otherwise tolerated. — one adult woman with anaphylaxis after coriander-flavoured beer (n = 1) Evidence C source
  Anaphylaxis, described by the authors as the first reported case of beer anaphylaxis caused by coriander added as a flavouring
• In a 5-day randomised dietary replacement trial, urinary DEHP metabolites rose from a median of 283.7 to 7027.5 nmol/g at P<0.0001 in the 21 people of the replaced-diet arm, and ingredient testing traced it to ground coriander at 21,400 ng/g. — 10 families, 21 people in the complete dietary replacement arm and 19 in the written-recommendations arm, 5 days (n = 40) Evidence B source
  Median urinary DEHP metabolites 283.7 to 7027.5 nmol/g at P<0.0001 in Arm 1 and no significant change in Arm 2, ground coriander measured at 21,400 ng/g of DEHP, estimated child intake above the US EPA oral reference dose and the EFSA tolerable daily intake
• Generic E. coli was found on 4.9% of Ontario-grown cilantro samples in a 2004 survey, below parsley at 13.4% and above field tomatoes at 0%, with no Salmonella and no verotoxigenic E. coli in the cilantro samples. — 61 cilantro samples within 1,183 samples of Ontario-grown fruits and vegetables, summer 2004 (n = 61) Evidence C source
  4.9% of cilantro samples positive for generic E. coli against 13.4% for parsley and 0% for head lettuce and field tomatoes
• Among 300 fresh vegetable samples from ten markets in northeast Thailand, cilantro among them, 36.0% carried intestinal parasites, most often Blastocystis hominis at 24.7%. — 300 samples of ten vegetables from 10 markets across five districts of Khon Kaen, Thailand (n = 300) Evidence C source
  36.0% of samples positive overall, Blastocystis hominis 24.7%, Strongyloides stercoralis 13.0%, Ascaris lumbricoides 8.7%, with peppermint highest at 70.0% and the rate for cilantro alone not reported in the abstract
• In a 2018 English outbreak of multidrug-resistant Shigella sonnei with 33 linked cases, 86% of the cases at one venue reported eating dishes with coriander as an ingredient or garnish, and all five outlets with two or more cases used fresh coriander. — 33 cases linked to seven catering outlets in England, April 2018 (n = 33) Evidence C source
  86% of cases at one venue exposed to coriander, 4 cases admitted to hospital and 1 with ciprofloxacin treatment failure, no shared supplier identified

What people believe that the data does not support

• Coriander seed powder at 2 g/day for 40 days improved the lipid parameters of patients with hyperlipidemia, in one of three supplement arms of a trial of four groups of 20, and the tested material was dried seed rather than the fresh cilantro leaf. — 80 hyperlipidemic patients randomised to four groups of 20, receiving garlic powder, coriander seed powder, a 1:1 mixture of the two or control, so 20 took coriander alone, 40 days of supplementation and 20 days of follow-up (n = 80) Evidence B source
  Significant change in BMI, HDL, total cholesterol, triglycerides, LDL and systolic blood pressure at p<0.05, HDL increased, no confidence intervals reported
• A 6 g blend of 13 spices that included coriander lowered postprandial IL-1beta secretion at 240 minutes in 12 men with overweight or obesity, while 2 g of the same blend did nothing and the share belonging to coriander alone was never measured. — 12 nonsmoking men aged 40 to 65 with overweight or obesity and at least one cardiovascular risk factor, 3-period crossover (n = 12) Evidence B source
  Spice-by-time interaction on IL-1beta P<0.001, IL-8 P=0.020, TNF-alpha P=0.009, IL-1beta reduced at 240 min with 6 g but not with 2 g

Who this works differently for

• In a two-patient series, fresh coriander produced systemic anaphylaxis only when strenuous exercise came first, and the same food eaten without exercise left the patient with local mouth itching alone. — 2 patients with oral allergy syndrome to the implicated foods, one to fresh coriander and tomato (n = 2) Evidence C source
  Systemic anaphylaxis when exercise preceded the meal, only perioral pruritus without exercise, cross-reactivity described with birch and grass antigens
• A double-blind randomised trial in 60 fully breastfeeding mothers and their infants found no adverse effects from a herbal tea whose ingredients included coriander fruit, with follow-up to the infants age of 12 months. — 60 healthy exclusively or fully breastfeeding women with no milk insufficiency and their infants, 30 days of tea and calls until infant age 12 months (n = 60) Evidence B source
  No adverse effects attributable to the interventions at any time point and no difference from placebo on maternal or infant symptoms, quality of life, infant growth or satisfaction, all p above 0.05

What is still unknown

• A scoping review of clinical trials on culinary doses of herbs and spices found 142 relevant trials across 14 herbs including coriander, and named cinnamon, fenugreek and ginger as the ones carrying both the most trials and the clearest signal for glycaemic control. — clinical trials in healthy people and in people with metabolic syndrome or type 2 diabetes (n = 142 trials) Evidence A source
  142 trials across 14 herbs, coriander among them, with the promising signals attributed to cinnamon, fenugreek, ginger, cardamom and turmeric rather than to coriander
• The whole cardiovascular evidence base for Coriandrum sativum is 18 studies, of which 12 are animal in vivo and only 2 are clinical, so the cardioprotective label rests mostly on rodents. — 18 studies retrieved from 3008 screened, 12 in vivo animal, 2 in vitro, 2 clinical, 2 mixed (n = 18 studies) Evidence D source
  No pooled estimate reported, the review states direction of effect only

The bottom line

Cilantro turns up in outbreak reports because it is eaten raw and handled a great deal. In a Salmonella Thompson outbreak in California in March 1999, eating cilantro at a restaurant carried an odds ratio of 3.5 (95% CI 1.1 to 11.4), with 63 percent of case patients exposed against 34 percent of controls across 76 cases. That is odds and not risk, and the source farm was never identified. The laboratory arm of that same investigation grew Salmonella Thompson faster and to a higher concentration on chopped cilantro than on whole leaves at room temperature, with no counts given in the abstract. Serving promptly after chopping is the one handling step those authors name. In a 2018 English outbreak of multidrug resistant Shigella sonnei with 33 cases across seven catering outlets, 86 percent of the cases at one venue had eaten dishes with coriander as an ingredient or garnish. All five outlets with two or more cases used fresh coriander, and no shared supplier was found. Dishes of that kind contain coriander almost always, which lifts the exposure share by itself. Surveys put numbers on the background. Generic E. coli was on 4.9 percent of 61 Ontario grown cilantro samples in the summer of 2004, below parsley at 13.4 percent and above field tomatoes at 0 percent, with no Salmonella found, and generic E. coli is a faecal indicator rather than a pathogen. Across 242 samples of six vegetables from northern California farmers markets in 2013, 6.6 percent carried Salmonella with most of the positives in cilantro and squash stems and leaves. Samples from growers who said they farmed organically had a twofold higher probability, with the practice taken from the seller statement and no certificate. Among 300 market samples of ten vegetables in Khon Kaen, Thailand, 36.0 percent held intestinal parasites, most often Blastocystis hominis at 24.7 percent, and the rate for cilantro on its own is not reported. Two allergy reports count for more than their size. A young woman had anaphylaxis to coriander flavoured beer she otherwise tolerated, the first such case its authors know of, which is one patient and tells you that coriander can arrive as an undeclared flavouring. In a two patient series, fresh coriander produced systemic anaphylaxis only when strenuous exercise came first, and the same food without exercise left nothing but itching around the mouth, on two cases whose mechanism the authors call a hypothesis. A dried batch can carry more than the plant. In a 5 day randomised dietary replacement trial, urinary DEHP metabolites in the 21 people of the replaced diet arm rose from a median of 283.7 to 7027.5 nmol/g (P<0.0001), while the other arm did not change. Ingredient testing found 21400 ng/g of DEHP in ground coriander, a figure that describes that batch and its packaging rather than the plant. From that batch the investigators estimated child intake above both the US EPA oral reference dose and the EFSA tolerable daily intake. For breastfeeding, a double blind randomised trial in 60 mothers found no adverse effects from a herbal tea whose ingredients included coriander fruit, with calls until the infants were 12 months old. That is the safety of a multi herb tea in 60 pairs, too few for rare harm.

Sources

1. cila-daily-r2-12 · cross-sectional microbial survey
   Sixteen (6.6%) samples had detectable levels of Salmonella serovars (Newport, Enteritidis, Agona, and Worthington), with the majority of positives found in cilantro and squash stems and leaves...There was a twofold higher probability of Salmonella contamination in samples from growers or vendors who stated that they used organic farming practices compared with samples from those using conventional farming practices.
   J Food Prot, 2015 · checked 2026-09-24
2. cila-daily-r2-06 · laboratory growth experiment
   At room temperature, Salmonella Thompson grew more rapidly and to a higher concentration on chopped cilantro, compared with whole-leaf cilantro...Cilantro should be served promptly after chopping.
   J Infect Dis, 2001 · checked 2026-09-24
3. cila-daily-r2-05 · case-control outbreak investigation
   A case-control study found a significant association between illness and eating cilantro at a restaurant (63% of case patients vs. 34% of control subjects; odds ratio, 3.5; 95% confidence interval, 1.1-11.4).
   J Infect Dis, 2001 · checked 2026-09-24
4. cila-daily-r2-07 · case report
   We report on a case of a young woman who experienced anaphylaxis due to coriander-flavoured beer. This is the first case of beer anaphylaxis due to coriander, which had been added to the beer as aromatising substance.
   BMJ Case Rep, 2018 · checked 2026-09-24
5. cila-daily-r2-09 · randomised dietary trial
   Urinary di(2-ethylhexyl) phthalate (DEHP) metabolite concentrations increased unexpectedly from a median of 283.7 nmol/g at baseline to 7027.5 nmol/g during the intervention (P<0.0001) among Arm 1 participants, and no significant changes were observed for Arm 2 participants...Arm 1 food ingredient testing for DEHP revealed concentrations of 21,400 ng/g in ground coriander and 673 ng/g in milk.
   J Expo Sci Environ Epidemiol, 2013 · checked 2026-09-24
6. cila-daily-r2-11 · prevalence survey
   The E. coli prevalence was highest in parsley (13.4%), followed by organic leaf lettuce (11.6%), leaf lettuce (6.5%), scallions (6.4%), cilantro (4.9%), muskmelon (1.3%), head lettuce (0%), and fresh market tomatoes (0%).
   J Food Prot, 2007 · checked 2026-09-24
7. cila-daily-r2-13 · prevalence survey
   A total of 300 samples, including cilantro, celery, Thai basil, lettuce, cucumber, Chinese kale, white cabbage, Chinese cabbage, peppermint, and yard-long beans, were collected from 10 markets across five districts in the province...The overall prevalence of IPs was found to be 36.0%, with Blastocystis hominis (24.7%), Strongyloides stercoralis (13.0%), and Ascaris lumbricoides (8.7%) being the most common.
   Am J Trop Med Hyg, 2025 · checked 2026-09-24
8. cila-daily-r2-14 · retrospective cohort outbreak investigation
   Five outlets were linked to two or more cases, all of which used fresh coriander although a shared supplier was not identified. An investigation at one of the venues recorded that 86% of cases reported eating dishes with coriander as an ingredient or garnish.
   Epidemiol Infect, 2021 · checked 2026-09-24
9. cila-daily-r2-01 · randomised controlled trial
   The groups were randomly assigned to three supplements i.e. garlic powder (GP), coriander seed powder (CSP) and mixture (1:1 dry weight basis) of GP and CSP at a dose rate of 2 g/day...It was concluded that consumption of garlic, coriander and their mixture at a dose rate of 2 g/day is improving the lipid parameters of the patients.
   Pak J Pharm Sci, 2018 · checked 2026-09-24
10. cila-daily-r2-02 · randomised controlled crossover trial
    The spice blend consisted of basil, bay leaf, black pepper, cinnamon, coriander, cumin, ginger, oregano, parsley, red pepper, rosemary, thyme, and turmeric...A HFCM containing 6 g spice blend attenuated HFCM-induced postprandial IL-1β secretion in men with overweight/obesity.
    J Nutr, 2020 · checked 2026-09-24
11. cila-daily-r2-08 · case series
    Patient A had FEIA attributed to fresh coriander and tomato and Patient B to fresh celery. These food allergens have been implicated in OAS and have structural antigenic similarity to that of birch and/or grass...In both cases, strenuous exercise was antecedent to the systemic anaphylaxis reaction and subsequent ingestion without exercise produced only local symptoms of perioral pruritus.
    Allergy Asthma Clin Immunol, 2013 · checked 2026-09-24
12. cila-daily-r2-10 · randomised double-blind controlled trial
    No adverse effects attributable to the interventions were reported at any time point...This composite tea appears to present no safety risk for mothers or their young babies.
    J Hum Lact, 2019 · checked 2026-09-24
13. cila-daily-r2-03 · scoping review of clinical trials
    Out of a total of 1738 papers identified, there were 142 relevant studies on black pepper, chilli, cardamom, cinnamon, coriander, cumin, fennel, fenugreek, garlic, ginger, nigella seed, rosemary, sage and turmeric...Cinnamon, fenugreek and ginger were the herbs/spices with the most published trials on them and that showed promise for glycaemic control.
    Nutrients, 2023 · checked 2026-09-24
14. cila-daily-r2-04 · review of in vivo animal and in vitro studies
    A total of 2977 studies were further excluded for not meeting the determined inclusion criteria, thus leaving only 18 studies to be retrieved and included in this review. Twelve in vivo, two in vitro, and two clinical studies in addition to two mixed in vivo and in vitro studies were retrieved.
    Molecules, 2021 · checked 2026-09-24