Magnet Ingestions in Children Presenting to Emergency Departments in the United States 2009–2019: A Problem in Flux : Journal of Pediatric Gastroenterology and Nutrition

Secondary Logo

Journal Logo

Rapid Communication

Magnet Ingestions in Children Presenting to Emergency Departments in the United States 2009–2019: A Problem in Flux

Reeves, Patrick T.∗,†; Rudolph, Bryan; Nylund, Cade M.∗,†

Author Information
Journal of Pediatric Gastroenterology and Nutrition: December 2020 - Volume 71 - Issue 6 - p 699-703
doi: 10.1097/MPG.0000000000002955


What Is Known/What Is New

What Is Known

  • The U.S. Consumer Product Safety Commission effectively banned the sale of high-powered magnet sets in 2012.
  • Magnet ingestions decreased 28% until a federal court decision allowed high-powered magnets to reenter the U.S. market in 2016.
  • Multiple public policy decisions, including proposed federal legislation, are pending the publication of additional ingestion data.

What Is New

  • Total magnet ingestions, likely high-powered magnet ingestions, and multiple magnet ingestions all substantially increased after 2016.
  • Escalation of care, defined as hospital admission or transfer, also increased after 2016.

Small rare-earth magnet sets (SREMs) are a novelty item that consists of extremely powerful, small, easily manipulated neodymium magnets. Sold in sets of up to 1000, they were first introduced in 2009 and marketed as stress-relieving desk toys, currently intended for consumers over the age of 14 (1). Despite the use of warning labels and multiple public education campaigns, the sale of SREMs has coincided with a significant uptrend of magnet ingestions in children of all ages (2,3).

When more than one is ingested, high-powered magnets frequently attach across bowel loops and cut off the vascular supply, leading to tissue perforation, fistulae, sepsis, or death (4–9). Given this unique health hazard and increased injury frequency, in 2012 the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN), the American Academy of Pediatrics (AAP), and numerous consumer groups lobbied the U.S. Consumer Product Safety Commission (CPSC) to take action. Within a matter of months, the CPSC did so, issuing a recall order and notice of proposed rulemaking to effectively ban high-powered magnet sets (3). These actions are credited for the decrease in pediatric magnet ingestions that followed (10).

One company, however, contested these actions and sued the CPSC in federal court (11). On November 22, 2016, the 10th Circuit Court sided with that company, Zen Magnets, and vacated the magnet rule set (MRS), once again allowing SREMs to be sold within the United States (12). The number of magnet ingestions in children is believed to have increased substantially since then, though national estimates are lacking (13).

In the interim, Zen Magnets has pushed for a new rule set that relies on warning labels and packaging requirements alone – steps that are ineffective in preventing injuries (2,3). A MRS is nevertheless close to passage within ASTM International (ASTM) – formerly the American Society for Testing and Materials – an organization responsible for developing product safety standards often adopted by the CPSC, over the objections of NAPGHAN, the AAP, consumer groups, and even CPSC staff from the Office of Hazard Identification and Reduction (14). The objections are that warning labels, as the CPSC has acknowledged, do not work. Therefore, NASPGHAN, the AAP, and others believe there should be a “performance standard” (ie, making magnets too large to swallow or too weak to cause harm) instead. Meanwhile, Congress is now considering legislation that would essentially ban high-powered magnet sets (15,16).

Additional data are urgently needed to inform these pending public policy decisions. We therefore utilized the National Electronic Injury Surveillance System (NEISS), a database of consumer product injuries that provides a stratified national probability sample, to investigate whether magnet ingestions have increased since high-powered magnets reentered the U.S. market.


We identified magnet ingestions from 2009 to 2019 utilizing the NEISS database, with a particular focus on 2 time periods: 2013–2016 (off-market), when the MRS was in place and SREMs within the United States were effectively banned, and 2017–2019 (on-market), when sale of SREMs resumed following the Tenth Circuit Court decision. Methods outlined in the proceeding paragraphs are similar to those previously published by members of this research group (10).

Data Source and Study Population

The NEISS database catalogs emergency department (ED) visits for injuries related to consumer products over a nationwide census. Specifically, the system functions as a stratified national probability sample of hospitals that have at least 6 beds and an ED. All institutions included in the database provide 24-hour services through their respective EDs and all possess the basic capabilities to treat children. NEISS thereby serves as a public health tool by pairing injury data with national census statistics, allowing for correlative conclusions on product risks.

Inclusion Criteria and Variable Classification

NEISS was queried to capture suspected magnet ingestion (SMI) cases in subjects 0–17 years of age using “ingested object” as the primary diagnostic search term. “Total” magnet ingestion cases were identified if the word magnet appeared anywhere in the case narrative section. These include any magnet ingestion, regardless of type or source. “Small/round” magnet ingestion cases, which are likely SREM related, were identified by narrative words, such as “small,” “tiny,” “1 cm or less (specified),” “toy,” “battery-like,” “marble,” “bead,” “BB,” “ball,” “sphere,” “round,” or “circular.” Suspected multiple magnet ingestions were identified by narrative words such as “multiple,” “number specified greater than 1,” or the plural word “magnets.” Clinical and demographic data were captured, as were any recorded outcome data (eg, hospitalization or discharge). Cases were designated as “escalation of care” if the disposition code for the encounter included “treated and transferred,” “treated and admitted/hospitalized,” or “held for observation.”

Statistical Analysis

NEISS supplied weights and variance variables used in all analyses. Taylor series linearization was used to generate national estimates with 95% confidence interval (CI). This procedure takes a component of interest (the actual number of SMIs in a single year) and uses a linear estimator (national census data) to approximate population-wide values. Of note, national estimates were not reported if fewer than 30 cases were identified, as low case numbers tend to produce uninterpretable estimates with high coefficients of variation and wide CIs. Population aggregate incidence rates were calculated using United States Census Bureau age-specific population estimates by year (17).

Regression analyses were modeled to determine if there were significant differences in age for total, small/round, or multiple magnet ingestions across both the time period of interest and since 2009. Rao-Scott chi-square was used for all categorical comparisons. Because the off- and on-market groups cover a different number of years (off-market included 4 years and the on-market period included 3 years), we divided the sample weights in each period by the number of years before calculating annualized estimates and corresponding CIs.

Multiple trend analyses were then performed using the design-corrected logistic regression method to examine trends in magnet ingestion rates over time (2009–2019) and the time period of interest (2013–2016 compared to 2017–2019) (18). In this evaluation, yearly census population data were used to evaluate changes in magnet ingestion rates over time. Next, we evaluated changes in the proportions of small/round ingestions to total magnet ingestions or multiple magnet ingestions to total magnet ingestions. These analyses were performed using NEISS national estimate data as the numerator and the national age-specific population per year (supplied by the United States Census Bureau) as the denominator.

A P < 0.05 was considered statistically significant. SAS 9.4 (SAS Institute, Cary, NC) was used for all analyses (19). The present study was reviewed and deemed exempt by the Uniformed Services University of the Health Sciences Institutional Review Board.


There were an estimated 23,756 (CI, 15,878–30,635) total SMI cases between 2009 and 2019 (Fig. 1). Of those, an estimated 3709 (CI, 2342–5076) cases involved small/round magnets and 6100 (CI, 3889–8311) involved multiple magnets (Table 1). The average annual increase in total cases was 6.1% over this time period (P = 0.01). There was also a statistically significant increase in small/round magnet ingestions (P < 0.001) and multiple ingestions (P = 0.02) between 2009 and 2019.

National estimated number of pediatric magnet ingestion emergency department visits in the United States from 2009 to 2019. From 2009 to 2010, the raw number of cases of multiple suspected magnet ingestion were less than 30 (raw range, 0–20) and thus not accurately estimable. The time between the double vertical dotted lines represents the study period when the magnet rule set was in effect and small rare-earth magnets were off the market.
TABLE 1 - Pediatric magnet ingestion-related emergency department visits in the United States between 2009 and 2019
Total Magnet Ingestions Small/Round Magnet Ingestions Multiple Magnet Ingestions
Actual (%) Estimated (95% CI) Actual (%) Estimated (95% CI) Actual (%) Estimated Cases (95% CI)
Total 1,113 (100) 23,756 (16,878–30,635) 148 (100) 3,709 (2,342–5,076) 333 (100) 6,100 (3,889–8,311)
Age, years
 0–4 599 (45.0) 14,590 (10,902 - 18,278) 83 (7.6) 2,477 (1,511 - 3,444) 118 (7.2) 2,323 (1,504–3,142)
 5–8 451 (31.3) 10,152 (7,320–12,984) 57 (4.2) 1,371 (666–2,075) 78 (4.5) 1,464 (792 - 2,137)
 9-13 349 (21.1) 6,843 (4,839–8,847) 48 (3.4) 1,099 (582–1,615) 147 (7.9) 2,575 (1,545–3,605)
 14-17 39 (2.6) 842 (402–1,281) 5 (0.6) 199 (0–409) 21 (1.3) 436 (120–751)
 Male 642 (57.7) 14,079 (9,559–18,598) 82 (55.4) 1,827 (814–2,841) 193 (58) 3,513 (2,379–4,647)
 White 611 (54.9) 13,526 (9,932–17,120) 90 (60.1) 2,398 (1,377–3,419) 208 (62.5) 4,070 (2,390–5,751)
 Black 61 (5.5) 822 (374–1,270) 7 (4.7) 108 (0–238) 11 (3.3) 239 (25–453)
 Unreported 328 (29.5) 7,114 (2,451–11,777) 32 (21.6) 731 (316–1,146) 84 (25.2) 1,398 (463–2,333)
 Other 113 (10.2) 2,294 (994–3,594) 19 (12.8) 472 (96–849) 30 (9) 393 (59–726)
 Home 649 (58.3) 16,403 (10,605–22,201) 87 (58.8) 2,550 (1,592–3,509) 184 (55.3) 4,203 (2,372–6,034)
 Admitted 238 (21.4) 2,785 (1,135–4,434) 31 (20.9) 271 (94–447) 134 (40.2) 1,535 (600–2,470)
 Released 853 (76.7) 19,780 (14,260–25,300) 112 (75.7) 3,149 (1,924–4,373) 187 (56.2) 3,861 (2,382–5,340)
 Transferred 22 (2) 1,192 (537–1,846) 5 (3.4) 290 (0–652) 12 (3.6) 705 (215–1,195)
 Escalation of care 260 (23.4) 3,976 (2,226–5,726) 36 (24.3) 561 (62–1,059) 146 (43.8) 2,240 (1,227–3,252)
Denotes that raw counts for this injury type were less than 30 but generated a 95% confidence interval with lower limit greater than 0. These results should still be interpreted with a certain degree of caution.
Denotes that confidence interval generates included 0.CI = confidence interval.

The median (interquartile range) age of children suspected to have ingested any magnet, a small/round magnet, or multiple magnets was 4.9 (2.8–8.8), 5.4 (2.8–9.5), and 7.7 (3.3–11.2) years, respectively. Cases since 2009 are reported by age in Supplemental 1, 2, and 3, Across all groups, the majority of SMI occurred in males and white children (56.9%), though a sizeable percentage did not have a reported race. Escalation of care was estimated to occur in 3976 (CI, 2226–5726) total magnet ingestion cases, 561 (CI, 62–1059) small/round magnet ingestion cases, and 2240 (CI, 1227–3252) multiple magnet ingestion cases (Table 1).

When stratified by time period, there were 6391 (CI, 4181–8601) estimated total magnet ingestion cases during the off-market period, or 1598 (CI, 1045–2150) estimated cases per year (Table 2). Conversely, there were 8478 (CI, 5472–11,485) estimated total cases during the on-market period, or 2826 (CI, 1824–3828) each year. This represents a 32% increase (P < 0.001) in total magnet ingestions after 2016. There was also a statistically significant increase in the number of estimated small/round (P < 0.01) and multiple (P < 0.001) magnet ingestions across these two time periods, with 164 (CI, 66–263) small/round and 350 (CI, 200–500) multiple magnet ingestions during the off-market period compared to 541 (CI, 261–822) small/round and 797 (CI, 442–1152) multiple magnet ingestion cases in the on-market period. The median age of children suspected of having ingested any magnet(s) during off-market time period was 4.6 (2.6–8.5) years compared to 4.8 (3.0–9.2) years in the on-market period (P < 0.001). There was no statistically significant age difference between children with small/round (P = 0.71) or multiple (P = 0.68) magnet ingestions when stratified by time periods.

TABLE 2 - Annualized age distribution of magnet ingestion-related emergency department visits in the United States, 2013–2016 (off-market period) and 2017–2019 (on-market period)
Total Magnet Ingestions Small/Round Magnet Ingestions Multiple Magnet Ingestions
2013-2016 2017-2019 2013-2016 2017-2019 2013-2016 2017-2019
Total 315; 1,598 (1045–2150) 415; 2,826 (1,824–3,828) 33; 164 (66–263) 65; 541 (261–822) 95; 350 (200–500) 136; 797 (442–1,152)
Age (Years)
 0–4 136; 715 (455–974) 164; 1,180 (716–1,644) 13; 68 (5–131) 20; 182 (62–302) 35; 157 (68–246) 38; 229 (88–371)
 5–8 90; 453 (259–647) 118; 775 (392–1,158) 10; 75 (4–146) 24; 172 (54–290) 24; 69 (17–121) 31; 179 (51–307)
 9-13 79; 390 (199–581) 119; 773 (512–1,035) 10; 21 (3–40) 19; 167 (48–286) 31; 113 (43–183) 62; 376 (195–558)
 14-17 10; 40 (0–83) 14; 98 (12–184) 0 2; 20 (0–57) 5; 11 (0–21) 5; 13 (0–28)
Results are expressed as number of raw cases; national estimate with associated 95% confidence interval.
Denotes that raw counts for this injury type were less than 30 but generated a 95% confidence interval with lower limit greater than 0. These results should still be interpreted with a certain degree of caution.
Denotes that confidence interval generates included 0.


The present study provides an update on the estimated national trends of SMI by children and demonstrates, for the first time, a rapid increase in national magnet ingestions since the 10th Circuit Court decision in 2016. Ingestion frequency has increased each year, with an estimated 4013 cases in 2019 alone. Importantly, this trend in total ingestions corresponds to an increase in the number of small/round ingestions, multiple magnet ingestions, and in the escalation of care, which strongly suggests that a large percentage of “total” magnet ingestion cases since 2017 are from SREMs. This is especially concerning given that previously published reports indicate that 32.5–75% of cases of multiple magnet ingestion result in perforation, which could be extrapolated to suggest there were approximately 1983–4575 cases of perforation related to multiple SREM ingestions from 2009 to 2019 (15).

These data may be useful in enacting public policy measures within ASTM, the CPSC, and/or Congress, as myriad regulatory actions on high-powered magnet sets remain in flux. After the 10th Circuit Court decision, Zen petitioned the CPSC to adopt a new, mandatory rule set that would require warning labels and packaging requirements for SREMs but not the performance standards favored by NASPGHAN (ie, making magnets either too large to swallow or too weak to cause harm). It also began working with industry partners to create a rule set within ASTM, the primary agency through which voluntary industry standards are created.

At the time of this article submission, an ASTM ballot favored by industry was being circulated. This rule is similar to that called for by Zen Magnets in its petition to the CPSC and is expected to pass over the objections of NASGPHAN, the AAP, multiple consumer groups, and even CPSC staff. Although there is no indication that companies will comply with the voluntary ASTM rule, passage will pressure the CPSC into issuing a new, mandatory rule set more favorable to industry.

Meanwhile, the CPSC is weighing action. On April 22, 2020, Zen Magnets unexpectedly withdrew its CPSC petition and, in response, the CPSC released its findings anyway. In a comprehensive rebuttal to Zen's petition, CPSC staff performed their own analysis of NEISS data and reached the same conclusion as our research group: high-powered magnet ingestions have increased since reentering the market in late 2016. Of note, the CPSC analysis utilizes stricter case identification methods by including only definite or highly likely cases. This conservative approach likely underrepresented national incidence and explains the somewhat disparate results between our studies (14). Nevertheless, both findings highlight a clear increase in magnet ingestions since the CPSC rule set was overturned and the urgent need for new intervention, such as that being considered in Congress.

In December of 2019, Senator Richard Blumenthal sponsored a new bill entitled “Magnet Injury Prevention Act” (S. 3143), which would legislate the performance standard currently favored by NASGPHAN into law. More recently, in March 2020, Congressional members Tony Cárdenas and Kim Schrier introduced a companion bill to the House. Both bills are currently in committee (16).

Though our analysis may help influence these pending policy decisions, several limitations should be acknowledged. First, NEISS contains data only from EDs and excludes patients evaluated in other settings (eg, children who do not seek medical care). Second, all foreign body ingestions herein were “suspected” and cannot be confirmed. Cases from NEISS do not provide case descriptions of care beyond the ED (ie, endoscopy, surgery, complications, or other morbidity associated with SMIs). Lastly, as there is no NEISS product code for magnets, cases were obtained by reviewing brief, individual narratives and were incomplete for some entries. This precludes extraction of important product-specific data, such as magnet type, flux, and brand. An important strength of our study, and NEISS more broadly, is the ability to generate a national estimate and confidence limit for pediatric magnet ingestions.


After the MRS was overturned by the 10th Circuit Court in 2016, there was a significant increase of magnet ingestions by children. Small, round magnet ingestions and multiple magnet ingestions also increased, which highlights the likelihood that small, round magnet ingestions and multiple magnet ingestions also increased, which highlights the likelihood that small, rare-earth magnets are a major, if not primary, source of these injuries. Regulatory action is needed to prevent children from harm.


1. Kadaba L. The attraction of Buckyballs. Available at: 2009. Accessed May 1, 2020.
2. North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. Warning labels ineffective at preventing high-powered magnet ingestions. Available at: 2012. Accessed May 1, 2020.
3. Consumer Product Safety Commission. Federal Register. Available at: 2012. Accessed May 1, 2020.
4. Agbo C, Lee L, Chiang V, et al. Magnet-related injury rates in children: a single hospital experience. J Pediatr Gastroenterol Nutr 2013; 57:14–17.
5. Naji H, Isacson D, Svensson JF, et al. Bowel injuries caused by ingestion of multiple magnets in children: a growing hazard. Pediatr Surg Int 2012; 28:367–374.
6. Brown JC, Otjen JP, Drugas GT. Pediatric magnet ingestions: the dark side of the force. Am J Surg 2014; 207:754–759.
7. Cox S, Brown R, Millar A, et al. The risks of gastrointestinal injury due to ingested magnetic beads. S Afr Med J 2014; 104:277–278.
8. Valdes M. Toy firm settles in death of Redmond toddler. Available at: 2006. Accessed May 1, 2020.
9. O’Donnell J. CPSC may ban tiny magnets that killed a child. Available at: 2014. Accessed May 1, 2020.
10. Reeves PT, Nylund CM, Krishnamurthy J, et al. Trends of magnet ingestion in children, an ironic attraction. J Pediatr Gastroenterol Nutr 2018; 66:e116–e121.
11. Shawn N. Skolky. CPSC's Zen Magnets Mandatory Recall Reversed: A Mixed District Court Decision Gives Zen Magnets a New Life and the CPSC Favorable Precedent for Mandatory Recalls. Available at: 2018. Accessed May 1, 2020.
12. Consumer Product Safety Commission. Safety standard for magnet sets; removal of final rule vacated by court. Federal Register. 2017. Available at: Accessed May 1, 2020.
13. Frankel T. Number of children swallowing dangerous magnets surges as industry largely polices itself. Washington Post. Available at: December 25, 2019. Accessed May 1, 2020.
14. Consumer Product Safety Commission. Informational briefing package regarding magnet sets. 2020. Available at: Accessed June 6, 2020.
15. Hussain SZ, Bousvaros A, Gilger M, et al. A. Management of ingested magnets in children. J Pediatr Gastroenterol Nutr 2012; 55:239–242.
16. Blumenthal R. S.3143 - Magnet Injury Prevention Act. Available at: December 19, 2019. Accessed May 1, 2020.
17. 2014 United States National Census Data. Available at: 2014. Accessed May 1, 2020.
18. Durand WM, Stey PC, Chen ES, et al. Trend analysis of aggregate outcomes in complex health survey data. AMIA Jt Summits Trans Sci Proc 2018; 2018:1–8.
19. SAS, Institute. “SAS/STAT User's guide, Version 9.2.” N. Cary (2004).

advocacy; foreign bodies; ingestion; magnets

Supplemental Digital Content

Copyright © 2020 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition