Clams: the trial used an extract, the risk is in the water

The only randomized trial on this page never put a clam on a plate. It gave 36 people with prediabetes in Taipei, aged 25 to 75, a 2 g capsule of freshwater clam extract or placebo every day for 180 consecutive days, and serum TNF-alpha was lower at six months and still lower six months after they stopped. The effect sat in the subgroup who started above 4.05 pg/mL of TNF-alpha rather than in everyone, which makes it a lead out of a small trial rather than a finding. Two grams of a concentrate is not a portion of shellfish, and nobody has measured a conversion between the two. Nothing on this page measures what that capsule does beyond TNF-alpha, or what amount of it would be too much. The extract is a dietary supplement, which in the US the FDA does not approve as a medicine before it is sold, and anyone with prediabetes who takes medicine for it should raise such a capsule with whoever treats them. Everything else known about clams as food comes from weaker designs, and the largest of them is about raw clams and hepatitis A. During the 1988 Shanghai hepatitis A epidemic the attack rate was 11.93 percent among people who had eaten clams against 0.52 percent among everyone else, a relative risk of 22.94, out of raw clams from sewage-contaminated water. In the matched case-control analysis of 1208 pairs from that epidemic, 3.5 percent of the people who fell ill had cooked their clams against 18.1 percent of those who stayed well. That is evidence about a virus. Nothing on this page shows that cooking removes the shellfish toxins described below, so a thoroughly cooked clam from water nobody has tested is not a safe clam.

The numbers, per 3.5 oz (100 g)

As people eat it, per 3.5 oz (100 g)

What a real portion looks like

• The dose in the only randomized clam trial was two grams of extract a day for 180 days in a row. — 36 patients with prediabetes, extract or placebo (n = 36) Evidence B source
  2 g a day for 180 consecutive days
• The large serving used in that safety calculation was 3.1 oz (88 g) of shellfish in a day, the 95th percentile among Korean shellfish eaters. — Korean shellfish consumers, KNHANES 2010 to 2015 Evidence C source
  88 g per day at the 95th percentile, paired with a maximum measured toxin level of 198.7 mcg/kg
• Pregnant women in Japan ate a median of 3.3 g a day of seafood and clams, against 28.7 g a day of fish overall. — 81428 pregnant women in the Japan Environment and Children's Study (n = 81428) Evidence C source
  Median 3.3 g/day of seafood and clams, interquartile range 0 to 8.6 g/day

What your body absorbs

• Feeding clams to 18 normolipidemic men lowered how much cholesterol they absorbed, alongside oysters and mussels. — 18 normolipidemic men, each shellfish fed so that its protein matched the animal protein of their usual diet (n = 18) Evidence C source
  Cholesterol absorption decreased on the oyster, clam and mussel diets

What cooking and storage change

• In the 1988 Shanghai hepatitis A epidemic only 3.5 percent of people who fell ill had cooked their clams, against 18.1 percent of those who stayed well. — 1208 matched case-control pairs from an epidemic of 292301 cases in Shanghai (n = 1208 matched pairs) Evidence C source
  3.5% of cases had cooked their clams vs 18.1% of controls

What it does to your health

• The same clam diet lowered VLDL triglycerides and cholesterol as well as LDL and total cholesterol in 18 men, while shrimp and squid diets changed nothing. — 18 normolipidemic men fed six shellfish species with fat restriction (n = 18) Evidence C source
  Lower VLDL triglycerides and cholesterol, lower LDL and total cholesterol on the clam diet; no change on squid or shrimp
• A 2 g daily capsule of freshwater clam extract lowered serum TNF-alpha in 36 people with prediabetes and the drop held for six months after they stopped. — 36 patients with prediabetes aged 25 to 75 in Taipei, 180 days of intake and 6 months of follow-up (n = 36 randomized, 19 extract and 17 placebo) Evidence B source
  Serum TNF-alpha reduced at 6 months and still reduced at 12 months
• The TNF-alpha effect of clam extract showed up in the people who started with raised TNF-alpha above 4.05 pg/mL, not in everyone equally. — the subgroup of the 36 prediabetic patients whose baseline serum TNF-alpha exceeded the 4.05 pg/mL cut-off (n = subgroup of 36) Evidence B source
  Suppression of serum TNF-alpha at 6 and 12 months in those with elevated baseline values

Safety, limits and interactions

• In the same epidemic the attack rate was 11.93 percent among clam eaters and 0.52 percent among everyone else, a relative risk of 22.94. — historical cohort of the Shanghai population, December 1987 to January 1988 (n = about 31.7% of the population had eaten clams at least once) Evidence C source
  Attack rate 11.93% vs 0.52%, relative risk 22.94, attributable risk 11.41%
• In a 2015 outbreak in Tamil Nadu the odds of illness were 314 times higher among people who had eaten clams, and 199 people fell ill with a 95 percent attack rate. — 64 selected cases and 66 controls in Cuddalore District, India, clams identified as Meretrix meretrix (n = 199 cases in the April outbreak) Evidence C source
  Odds ratio 314 (95% CI 39 to 512) for eating clams; these are odds, not risk
• A clam left over from an Alaskan meal that put one woman in intensive care measured 277 micrograms of saxitoxin per 3.5 oz (100 g), more than three times the regulatory action level. — one elderly woman admitted to intensive care with respiratory depression and shock, discharged after 6 days (n = 1) Evidence C source
  Saxitoxin 277 mcg/100 g, neosaxitoxin 309 mcg/100 g, gonyautoxins 576 to 2490 mcg/100 g in the leftover clam
• The FDA acts against a lot of clams, mussels or oysters once paralytic shellfish poison reaches 80 micrograms per 3.5 oz (100 g) of meat. — fresh, frozen or canned clams, mussels and oysters in United States commerce Evidence E source
  Action level 80 micrograms per 3.5 oz (100 g) of meat, equal to 0.8 mg/kg saxitoxin equivalent
• Modeled Korean intake put the worst realistic single serving of shellfish at 43 to 60 percent of the acute reference dose for paralytic shellfish toxins. — Korean population intake from KNHANES 2010 to 2015 matched to measured toxin levels in mussel, clam, oyster and ark shell Evidence D source
  Acute exposure 0.30 mcg/kg body weight against acute reference doses of 0.5 or 0.7 mcg STX eq/kg, that is 43 to 60% of them
• Shellfish allergy affects an estimated 0.5 to 2.5 percent of people and runs higher in coastal Asian countries where shellfish are eaten often. — general population estimates pooled in a review of shellfish allergy Evidence C source
  Prevalence 0.5 to 2.5% of the general population
• Fish and seafood including carpet shell clam brought the median Italian eater to 20 percent of the tolerable weekly intake for cadmium, with the 95th percentile going over it. — twelve fish and seafood products sold in Italy, including carpet shell clam, matched to Italian consumption data Evidence D source
  Estimated weekly intake at 20% of the cadmium TWI for the median of the population, 1.5% and 10% of the lead and mercury provisional limits, and over the cadmium limit at the 95th percentile
• The 1988 Shanghai case-control analysis of 1208 matched pairs put the odds of hepatitis A at 9.47 times higher for clam eaters. — 1208 matched case-control pairs during an epidemic with an attack rate of 4083 per 100000 (n = 1208 matched pairs) Evidence C source
  Summary odds ratio 9.47, P less than .001; these are odds, not risk

Who this works differently for

• Among 513 Native American adults, those eating 15 or more razor clams a month scored lower on some memory tests, though the scores stayed inside the normal range. — 513 men and women from three Pacific Northwest tribes with regular access to razor clam beaches, followed over four years (n = 513) Evidence C source
  Isolated decrements on some memory measures at 15 or more clams a month, p = .02 to .03, other cognitive functions unaffected
• Tropomyosin is the major shellfish allergen and it cross-reacts between crustaceans and molluscs, which is why a shrimp allergy puts clams and mussels in question rather than in the clear. — people with IgE-mediated shellfish allergy Evidence C source
  Cross-reactivity within crustaceans, within molluscs, between the two groups and with fish
• A boy given boiled short-neck clam under supervision vomited two hours later with abdominal pain, the first reported case of food protein-induced enterocolitis from shellfish. — a 6-year-old boy with vomiting after short-neck clam since age 2, negative for other food and shellfish allergy (n = 1) Evidence C source
  Abdominal pain and vomiting 2 hours after ingestion, raised peripheral leukocytes at 6 hours
• Across 81428 Japanese pregnancies, eating seafood and clams carried no association with preterm birth, while fish paste did. — 81428 mother-child pairs in the Japan Environment and Children's Study, diet measured by food frequency questionnaire during pregnancy (n = 81428) Evidence C source
  No significant association for seafood and clams; fish paste in the highest quintile OR 1.11 (95% CI 1.04 to 1.17)
• CDC says most vibriosis comes from eating raw or undercooked shellfish, and that liver disease, cancer, diabetes, HIV, immune-suppressing therapy or stomach acid reducers raise the risk of infection and severe complications. — people with liver disease, cancer, diabetes, HIV or thalassemia, people on immune-suppressing therapy, people taking medicine to decrease stomach acid Evidence E source
  Higher risk of Vibrio infection and severe complications (no size given)

The bottom line

What clams do to blood lipids rests on 18 normolipidemic men fed six shellfish species in 1990, with no randomization. On the clam diet, VLDL triglycerides and cholesterol fell along with LDL and total cholesterol, cholesterol absorption fell as well, and the squid and shrimp diets changed nothing. Fat in that study was cut by more than half against the usual diet, so part of what happened belongs to the fat restriction. A realistic portion is small. Korean shellfish eaters at the 95th percentile ate 3.1 oz (88 g) of shellfish in a day. Among 81428 pregnant women in Japan, the median intake of seafood and clams was 3.3 g a day, with an interquartile range of 0 to 8.6 g, against 28.7 g a day of fish overall. In the country that eats the most shellfish, clams are an occasional food. Toxins are the real subject here. FDA acts against a lot of clams, mussels or oysters once paralytic shellfish poison reaches 80 micrograms per 3.5 oz (100 g) of meat, which is a regulator threshold for monitored commercial shellfish and not a dose-response study. A clam left over from an Alaskan meal that sent one elderly woman to intensive care with respiratory depression and shock measured 277 micrograms of saxitoxin per 3.5 oz (100 g), with neosaxitoxin at 309 and gonyautoxins from 576 to 2490. That is one case, nobody else at the same meal had symptoms, and it shows what a recreationally harvested clam can carry when nothing has tested it. The FDA threshold protects only shellfish that pass through that monitoring, and a clam you dig yourself has not. Set against measured toxin levels, a Korean worst-case serving reached 0.30 micrograms per kg of body weight, which is 43 to 60 percent of the acute reference doses of 0.5 or 0.7 micrograms saxitoxin equivalent per kg. The margin is real and it is not wide. Cadmium is the metal that matters for bivalves, because they filter it. Twelve fish and seafood products sold in Italy, carpet shell clam among them, brought the median Italian eater to 20 percent of the tolerable weekly intake for cadmium, with the 95th percentile going over it. Several groups have a reason to be careful, and the reasons differ. Shellfish allergy is estimated at 0.5 to 2.5 percent of the general population and runs higher in coastal Asian countries where shellfish are eaten often. Tropomyosin is the allergen that cross-reacts between crustaceans and molluscs, so anyone who reacts to shrimp should treat clams as a question for a clinician rather than as settled either way. A negative allergy test does not close the matter. A 6-year-old boy who tested negative for shellfish allergy vomited with abdominal pain two hours after a supervised challenge with boiled short-neck clam. That was the first reported case of food protein-induced enterocolitis from shellfish, and the reaction is delayed rather than IgE-driven, which is why the blood test missed it. Among 513 adults from three Pacific Northwest tribes followed over four years, those eating 15 or more razor clams a month showed isolated decrements on some memory measures (p = .02 to .03), with other cognitive functions unaffected and the scores still within normal limits. In 81428 Japanese pregnancies, seafood and clams carried no significant association with preterm birth, though clams were grouped with other seafood rather than analyzed alone. That cohort measured one outcome at a median intake of 3.3 g a day, and a null on preterm birth is not a clearance for raw or undercooked clams in pregnancy. This page has no card on which other groups are at higher risk from raw or undercooked clams, and the absence of a warning here is not evidence of safety for anyone. An outbreak in Cuddalore District in 2015 put the odds of illness 314 times higher among clam eaters (95 percent CI 39 to 512), which are odds rather than risk and come out of near-total exposure among the cases.

Compared with other foods

Sources

1. clam-r5-07 · RCT
   Two grams of FCE or placebo was consumed daily for 180 consecutive days.
   Mar Drugs, 2022 · checked 2026-09-24
2. clam-r5-13 · dietary exposure assessment
   The acute exposure using a large portion size of 88 g/day (95th percentile for consumers only) with maximum toxin level of 198.7 μg/kg was 0.30 μg/kg bw.
   Food Res Int, 2018 · checked 2026-09-24
3. clam-r5-19 · cohort dietary assessment
   The median (interquartile range) consumption of total fish, fatty fish, lean fish, fish paste, seafood and clams, and other seafoods during pregnancy was 28.7 (15.3, 46.4) g/day, 13.3 (6, 22.8) g/day, 0 (0, 2.8) g/day, 1.3 (0.4.0) g/day, 3.3 (0, 8.6) g/day, and 6.7 (2.3, 10.7) g/day, respectively.
   Environ Health Prev Med, 2023 · checked 2026-09-24
4. clam-r5-03 · controlled clinical trial
   Cholesterol absorption was decreased on the oyster, clam, and mussel diets.
   Am J Clin Nutr, 1990 · checked 2026-09-24
5. clam-r5-01 · case-control
   Analysis of subsets who had eaten clams indicated that only 3.5% with hepatitis A had cooked their clams compared with 18.1% without hepatitis A, and those with the disease consumed more clams.
   J Infect Dis, 1991 · checked 2026-09-24
6. clam-r5-04 · controlled clinical trial
   Oyster, clam, crab, and mussel diets, low in cholesterol and high in n-3 fatty acids, lowered VLDL triglycerides and cholesterol and, except for the mussel diet, LDL and total cholesterol.
   Am J Clin Nutr, 1990 · checked 2026-09-24
7. clam-r5-05 · RCT
   FCE intervention effectively reduced serum TNF-α and persistently sustained the effects for half a year in patients with prediabetes.
   Mar Drugs, 2022 · checked 2026-09-24
8. clam-r5-06 · RCT subgroup
   Moreover, FCE significantly suppressed serum TNF-α concentrations at 6M and 12M and inhibited TNF-α release with time series in subjects with elevated TNF-α values.
   Mar Drugs, 2022 · checked 2026-09-24
9. clam-r5-02 · historical cohort
   The estimated attack rates in those who had and had not eaten clams were 11.93% and 0.52%, respectively (relative risk, 22.94; attributable risk, 11.41%).
   J Infect Dis, 1991 · checked 2026-09-24
10. clam-r5-09 · case-control
    Illness was associated with a history of eating of clams (odds ratio = 314, 95% confidence interval = 39-512).
    Indian J Public Health, 2021 · checked 2026-09-24
11. clam-r5-10 · case report
    A clam remaining from the meal was collected and analyzed for paralytic shellfish toxins (PST) by the Alaska Department of Environmental Conservation Environmental Health Laboratory; the clam tested positive for saxitoxin (STX; 277 μg/100 g), neosaxitoxin (NEO; 309 μg/100 g), multiple gonyautoxins (GTX; 576-2490 μg/100 g), decarbamoyl congeners (7.52-11.3 μg/100 g) and C-toxins (10.8-221 μg/100 g) using high-pressure liquid chromatography with post-column oxidation (AOAC Method 2011.02).
    J Anal Toxicol, 2018 · checked 2026-09-24
12. clam-r5-11 · agency action level
    Actionable if one sub from a lot shows paralytic shellfish poison value of 80 micrograms, or more, per 100 grams meat when bioanalyzed by current Association of Official Analytical Chemists procedure.
    FDA Compliance Policy Guide Sec. 540.250, Clams, Mussels, Oysters, Fresh, Frozen or Canned, Paralytic Shellfish Poison · checked 2026-09-24
13. clam-r5-12 · dietary exposure assessment
    Even though we estimated the acute exposure with a conservative manner, it was below the ARfDs (0.5 or 0.7 μg STX eq./kg bw) proposed by the international organizations, representing 43-60% of the ARfDs.
    Food Res Int, 2018 · checked 2026-09-24
14. clam-r5-14 · review of prevalence
    The prevalence of shellfish allergy is estimated to be 0.5-2.5% of the general population but higher in coastal Asian countries where shellfish constitute a large proportion of the diet.
    Immunol Invest, 2016 · checked 2026-09-24
15. clam-r5-17 · dietary exposure assessment
    The EWIs for Cd, Pb and Hg related to the consumption of fish and seafood products by the median of the Italian total population accounted for 20%, 1.5% and 10% of the standard TWI for Cd as well as PTWIs for Pb and Hg, respectively.
    Food Addit Contam Part A Chem Anal Control Expo Risk Assess, 2012 · checked 2026-09-24
16. clam-r5-20 · case-control
    A case-control study of 1208 matched pairs supported that clams were the vehicle for the virus (summary odds ratio, 9.47; P less than .001).
    J Infect Dis, 1991 · checked 2026-09-24
17. clam-r5-08 · cohort
    Based upon GEE analysis, controlling for age, sex, race, year, education level, tribe, and employment status, findings indicated that high razor clam consumers (15 or more per month) had isolated decrements on some measures of memory (p=.02 to .03), with other cognitive functions unaffected. The relatively lower memory scores were still within normal limits, thus not clinically significant.
    Harmful Algae, 2016 · checked 2026-09-24
18. clam-r5-15 · review of allergens
    Tropomyosin is the major allergen and is responsible for cross-reactivity between shellfish and other invertebrates, within crustaceans, within molluskans, between crustaceans vs. molluskans as well as between shellfish and fish.
    Immunol Invest, 2016 · checked 2026-09-24
19. clam-r5-16 · case report with oral challenge
    An oral challenge test with boiled short-neck clam induced abdominal pain and vomiting 2 h after ingestion, and the patient presented with increased peripheral leukocytes after 6 h.
    Arerugi, 2010 · checked 2026-09-24
20. clam-r5-18 · cohort
    Consumption of other types of fish, except fish paste, was not significantly associated with preterm birth risk.
    Environ Health Prev Med, 2023 · checked 2026-09-24
21. sf-vibrio-clams-1 · agency advice
    Most people get vibriosis by eating raw or undercooked shellfish, particularly oysters...Some medical conditions and treatments can increase your risk for infection and severe complications. They include:...Having liver disease, cancer, diabetes, HIV, or thalassemia...Receiving immune-suppressing therapy for the treatment of disease...Taking medicine to decrease stomach acid levels
    About Vibrio Infection, US Centers for Disease Control and Prevention, People at risk · checked 2026-10-08