Supplement safety statistics show two different realities: most available measurements concern selected adverse events or flagged products, while broad denominator-based risk remains difficult to establish. The figures below cover U.S. emergency care from 2004–2013, FDA warnings from 2007–2016, clinical laboratory samples, and historical international testing.
Key Supplement Safety Statistics
The most quotable findings include:
- In the United States during 2004–2013, an estimated 23,005 emergency-department visits per year involved dietary-supplement adverse events.
- In the United States during 2004–2013, an estimated 2,154 supplement-related emergency visits led to hospitalization each year.
- In the 2004–2013 U.S. surveillance sample, 3,667 emergency-department cases were recorded across 63 hospitals.
- During 2004–2013 in the United States, unsupervised child ingestions accounted for 21.2% of supplement-related visits.
- During 2004–2013 in the United States, adults aged 20–34 accounted for 28.0% of supplement-related emergency visits.
- After child-ingestion exclusions in U.S. 2004–2013 surveillance, herbal or complementary products represented 65.9% of single-supplement adverse-event visits.
- After child-ingestion exclusions in U.S. 2004–2013 surveillance, micronutrients represented 31.8% of single-supplement adverse-event visits.
- After child-ingestion exclusions in U.S. 2004–2013 surveillance, weight-loss products were implicated in 25.5% of single-supplement adverse-event visits.
- During 2004–2013 U.S. surveillance, weight-loss or energy products were involved in 71.8% of adverse events with palpitations, chest pain or tachycardia.
- Among U.S. adults aged 65 or older during 2004–2013, choking or pill-induced swallowing symptoms caused 37.6% of supplement adverse-event visits.
- The FDA-warning analysis identified 776 adulterated supplement products in the United States from 2007–2016.
- In the FDA-warning sample from 2007–2016, sexual-enhancement products represented 45.5% of flagged products.
- In the FDA-warning sample from 2007–2016, weight-loss products represented 40.9% of flagged products.
- In the FDA-warning sample from 2007–2016, unapproved drug ingredients were absent from the product label in 97.6% of products.
- In the FDA-warning sample from 2007–2016, 79.8% of products contained one unapproved pharmaceutical ingredient.
- In a U.S. liver-injury clinical sample, 140 of 272 herbal and dietary supplements, or 51%, were mislabeled.
- In the same U.S. liver-injury sample, 55 of 272 products contained compounds not listed on the label.
- In a historical international study of samples collected in 2001–2002, 94 of 634 non-hormonal supplements, or 14.8%, contained anabolic steroids or steroid derivatives.
Contents
- Emergency visits and hospitalizations
- Age, symptoms and product purpose
- FDA adulteration and hidden ingredients
- Label accuracy and liver-injury samples
- International contamination and athlete safety
- Surveillance limits and how to read the data
Emergency Visits and Hospitalizations: U.S. Supplement Safety Data
The main national burden estimate comes from a study of U.S. emergency-department surveillance during 2004–2013. It used 3,667 observed cases from 63 hospitals to estimate medically attended events, not every adverse reaction occurring in the population. (Emergency Department Visits for Adverse Events Related to Dietary Supplements)
| Measure | U.S. estimate for 2004–2013 | Interpretation |
|---|---|---|
| Emergency-department visits | 23,005 per year | Estimated annual burden |
| 95% confidence interval | 18,611–27,398 per year | Sampling uncertainty |
| Visits leading to hospitalization | 2,154 per year | Estimated admissions from ED visits |
| Hospitalization 95% confidence interval | 1,342–2,967 per year | Sampling uncertainty |
The confidence intervals matter because these are surveillance estimates rather than a census. They also exclude events that did not reach emergency care, so the figures should not be presented as current global incidence.
The hospitalization estimate describes a serious subset of medically attended events. It does not measure all supplement-related hospital care or establish that a particular product caused every reported outcome.
Age, Symptoms and Product Purpose in Supplement Adverse-Event Statistics
The U.S. emergency-department study shows why age and symptom context are essential. A percentage of visits is not the same as a person-level risk rate, because the analysis does not provide a denominator of supplement users for each subgroup.
Key age findings from 2004–2013 U.S. surveillance include:
- Unsupervised child ingestions accounted for 21.2% of supplement-related visits, with a 95% confidence interval of 18.4%–24.0%.
- Adults aged 20–34 accounted for 28.0% of visits, with a 95% confidence interval of 25.1%–30.8%.
- Among adults aged 65 or older, choking or pill-induced dysphagia or globus caused 37.6% of supplement adverse-event visits, with a 95% confidence interval of 29.1%–46.2%.
After excluding unsupervised child ingestions, the distribution of single-supplement visits was:
| Product purpose or category | Share of single-supplement visits | 95% confidence interval |
|---|---|---|
| Herbal or complementary products | 65.9% | 63.2%–68.5% |
| Micronutrients | 31.8% | 29.2%–34.3% |
| Weight-loss products | 25.5% | 23.1%–27.9% |
| Energy products | 10.0% | 8.0%–11.9% |
These categories can overlap in how products are marketed and analyzed, so the shares should be read as study-defined surveillance categories rather than a ranking of individual-user risk.
Cardiovascular-type symptoms concentrated in specific product-purpose groups. Weight-loss or energy products were associated with 71.8% of events involving palpitations, chest pain or tachycardia, with a 95% confidence interval of 67.6%–76.1%; adults aged 20–34 represented 58.0% of that symptom subgroup, with a 95% confidence interval of 52.2%–63.7%.
Among older adults with choking or pill-induced swallowing symptoms, micronutrients were implicated in 83.1% of visits, with a 95% confidence interval of 73.3%–92.9%. These are associations in emergency surveillance, not proof of product causality.
FDA Adulteration and Hidden-Ingredient Statistics
A separate analysis examined U.S. Food and Drug Administration warning records from 2007–2016. It identified 776 adulterated products after reviewing 781 raw database entries, adding 14 products and removing 19 duplicates; this is a warning-database sample, not prevalence among all supplements sold. (Unapproved Pharmaceutical Ingredients Included in Dietary Supplements Associated With US Food and Drug Administration Warnings)
The warning sample’s product-purpose breakdown was:
- Sexual-enhancement products accounted for 353 of 776 products, or 45.5%.
- Weight-loss products accounted for 317 of 776 products, or 40.9%.
- Muscle-building products accounted for 92 of 776 products, or 11.9%.
- Other indications accounted for 14 of 776 products, or 1.8%.
Undeclared pharmaceuticals were central to the warning records. Unapproved drug ingredients were not declared on the label in 757 of 776 products, or 97.6%; one unapproved pharmaceutical ingredient appeared in 619 products, or 79.8%, while more than one appeared in 157 products, or 20.2%.
Three or more adulterants were detected in 33 products. Within that small selected subgroup, 16 products, or 48.5%, were marketed for weight loss and 13, or 39.4%, for sexual enhancement.
The sexual-enhancement subset offers a more specific example of hidden-ingredient patterns:
- Sildenafil or a structural analogue appeared in 287 of 353 adulterated sexual-enhancement products, or 81.3%.
- Sildenafil itself appeared in 166 of 353 products, or 47.0%.
- Structural sildenafil analogues appeared in 134 of 353 products, or 38.0%.
Warning actions also changed within the period. Across 2007–2016, voluntary recalls were associated with 360 products, or 46.4%, and public-notification-only actions with 342, or 44.1%; in the 2012–2016 subset, public notifications covered 321 of 443 products, or 72.5%.
Label Accuracy in U.S. Liver-Injury Samples
Laboratory evidence provides a different safety lens from emergency surveillance. A U.S. analysis of herbal and dietary supplements implicated in liver-injury cases found accurate labels in 132 of 272 products, while 140 of 272, or 51%, were mislabeled. (The Contents of Herbal and Dietary Supplements Implicated in Liver Injury in the United States Are Frequently Mislabeled)
The sample was collected from people enrolled in a liver-injury study. It therefore cannot be used as a retail-market mislabeling rate.
In the same 272-product clinical sample, 55 products contained compounds not listed on the label. Not all unlisted compounds were necessarily pharmaceuticals, so “unlisted” and “adulterated with an unapproved drug” are not interchangeable definitions.
Mislabeling varied by product class:
| Product class or purpose | Mislabeling result | Sample context |
|---|---|---|
| Steroidal products | 82% | Selected liver-injury cases |
| Botanical products | 44% | Selected liver-injury cases |
| Vitamin products | 49% | Selected liver-injury cases |
| Athletic-performance products | 80% | 46 selected products |
| Weight-loss products | 72% | 36 selected products |
| Energy-boosting products | 60% | 5 selected products |
| Sexual-enhancement products | 100% | 4 selected products |
Among 26 products with known hepatotoxins, 13 contained anabolic steroids. The very small energy-boosting and sexual-enhancement groups make their percentages especially sensitive to individual products.
International Contamination and Athlete-Safety Statistics
Historical international testing shows why geography, sampling design and product definitions must stay visible. An analysis of 634 non-hormonal supplements collected in 2001–2002 across 13 countries found 94 samples, or 14.8%, containing anabolic steroids or steroid derivatives; the products came from 215 separate suppliers. (Prevalence of adulteration in dietary supplements and recommendations for safe supplement practices in sport)
The study was targeted and historical, so it is not a current estimate of global market prevalence. Within the tainted samples, products purchased in the Netherlands contributed 25.8%, while products purchased in Austria contributed 22.7%; those are contributions among tainted samples, not country-specific prevalence rates.
A broader review reported contamination rates ranging from 12% to 58% across 23 studies reviewed through 2017. The studies used different products, methods and adulterant definitions, so the range is not a pooled estimate and should not be treated as a direct country comparison.
These findings are particularly relevant to athletes because an undeclared steroid or derivative can create safety and eligibility concerns even when the marketed purpose is not explicitly steroid-related. The supplied evidence does not quantify current athlete exposure, current positive-test rates or the risk from any specific creatine product.
Supplement Safety Surveillance Limits and Evidence Gaps
The statistics above use several different denominators: emergency visits, FDA-warning products, clinically selected products and laboratory samples. Combining them into one overall “supplement risk” percentage would be unsupported.
The most important evidence limits are:
- Emergency-department surveillance estimates medically attended events from 63 hospitals and 3,667 cases during 2004–2013; it does not capture reactions that never reached emergency care.
- The 2004–2013 U.S. burden estimate is not current global incidence, and its product shares are not population risk rates.
- FDA-warning figures from 2007–2016 describe products associated with agency warnings, not the prevalence of adulteration across all products.
- The DILIN laboratory findings describe selected products collected from people with suspected liver injury, not ordinary retail shelves.
- The international steroid-contamination study used samples collected in 2001–2002, while its review range covered studies through 2017; neither establishes current prevalence.
- Adverse-event reports and poison-center calls can support signal detection but generally cannot establish causality or denominator-based risk.
Taken together, the evidence supports a cautious interpretation: supplement safety problems can involve acute medical visits, swallowing incidents, undeclared pharmaceuticals, mislabeled ingredients and contamination, but each statistic must remain attached to its population, period, geography and sampling frame. Broad supplement statistics also do not measure creatine specifically.