Interpretation of “An analysis of energy-drink toxicity in the National Poison Data System”

To the Editor:

The safety of energy drinks is currently debated in the US. Seifert et al. stated that they performed “analysis of energy drink toxicity in The National Poison Data Systems (NPDS)”.¹ Their article was actually an analysis of energy drink exposures reported to NPDS.

Information bias may have resulted from lack of conformation to NPDS definitions of outcome, adverse events, and toxicity. Seifert et al. state that NPDS classifies the” severity of adverse effects” in five ordinal categories (“no effect”, “minor”, “moderate”, “major”, and “death”).¹ These categories are actually 5 of the 10 categories of medical outcome in NPDS. The five additional medical outcome categories (“not followed, non-toxic”, “not followed, minimally toxic”, “not followed, potentially toxic”, “unrelated effect”, and “death, indirect report”) are categorized by the authors as “unknown” outcome. But, outcome in these categories is recorded and ranges from nontoxic to death. Although cases may not be followed when history and dose indicate minimal or no toxicity, this does not justify classifying these cases as “unknown outcome”. The authors subsequently use these same medical outcome categories to determine “severity” of toxicity” (Table 2).¹ Neither severity of adverse events nor severity of toxicity are NPDS fields. Medical outcome is predefined and may not correlate with severity of toxicity.

Amount ingested according to age and clinical course were not included in the analysis. Children below 6 years old unintentionally ingest small amounts. In this age group exposed to non-alcoholic energy drinks, 28% experienced minor, 1% moderate, and none experienced major effects (Figure 3).¹ (It is important to note that 345/717 cases in this age group were classified as unknown outcome. If these cases were not followed due to lack of toxicity, incidence of minor effects would be 14% (103/717)). Those above 13 years old were more likely to experience a moderate/major effect, potentially due to a larger amount ingested. Amount ingested and clinical course could have determined if toxicity correlated to dose.

Based on management site data, the authors conclude “most calls for alcoholic energy drink ingestion resulted in advice to seek medical care” and further assert that type of energy drink (nonalcohol or alcohol) determined health care facility (HCF) referral.¹ But suicidal intent, amount ingested, and age were not considered—all important co-determinants of HCF referral. Nor did the authors differentiate patients that go to a HCF without PC referral. The statement that “almost 10% of exposures to alcoholic energy drinks resulted in critical care unit admission” is confusing, as nine of 182 (4.9%) in this group were admitted to critical care units.¹ Again, this may have been due to intent, not toxicity.

The supposition that the incidence of energy-drink-related toxicity may be under-reported is controversial.¹ One could also conclude that this incidence represents over-reporting, as callers may be more likely to call if symptoms (even minor ones) were experienced.

The conclusion “about half the cases of energy drink-related toxicity involve unintentional exposures by children <6 years of age”¹ should read “about half the cases of energy drink related exposures are unintentional in this age group”. An exposure does not equal toxicity.

Safety is an important issue. If NPDS analysis inaccurately concludes that “safe” products are “unsafe”, the reverse will also occur- “unsafe” products will be inaccurately deemed to be “safe”. If NPDS data are to be analyzed for safety, inherent limitations must be recognized, standardized NPDS definitions respected, all relevant fields evaluated, and conclusions substantiated by the data. The data will then speak for itself.

Declaration of interest

Donna Seger has served as a consultant to the Red Bull energy drink company.