Consumer Healthcare Products Association response to Major et al., “Trends in rates of acetaminophen‐related adverse events in the United States”, Pharmacoepidemiology & Drug Safety, May 2016, 25: 590–8

In the May issue, Major et al. examined trends in adverse events associated with use of prescription and over-the-counter (OTC) acetaminophen products (2009–2012) from data within three national surveillance systems reflecting calls to U.S. Poison Centers, emergency department visits, and hospitalizations. The analysis included changes in the number of adverse events over time as well as trends in adverse event rates (per population and per 1 million drug units sold/distributed). The number of adverse events associated with acetaminophen and captured in each of these three databases decreased during the 2009– 2012 period. Adverse event rates (per population) also decreased for poison center calls and hospitalizations. Major et al. also estimated the rate of acetaminophen-associated adverse events using sales of acetaminophen-containing products as a proxy for consumer use, to understand how adverse event trends were affected by changes in medicines sold. Despite a reduction in both adverse event numbers and the adverse event rate per population, the authors found an increase in event rates (after 2009) when adjusting for changes in sales of acetaminophen (events per million units sold). The authors used IMS Health, National Sales Perspective data as a source of both prescription and OTC acetaminophen drug sales. IMS is an appropriate data source for prescription drug sales, using a scanner-based method that captures over 90% of U.S. pharmaceutical product sales. For OTC products, however, IMS uses a different sample-based method of retail orders from warehouses to project OTC sales to consumers1 which includes indirect sales of retail and non-retail channels currently collected from 322 wholesalers and 60 drug chain distributors, a panel of over 300 non-federal hospitals, and 142 mail service pharmacies. These “indirect sales” measure the orders to wholesalers and distributors (from retailers) but do not represent a measure of consumer demand/sales. Use of IMS data can lead to a significant gap compared to scanner-based sales data sources with significantly higher sample coverage available from other data providers such as Information Resources Inc. (IRI), with a sample method that captures over 90% of retail sales, and Nielsen.2 For perspective, in comparison to the 57% sales decrease estimated by Major et al. using the samplebased data projections of IMS, U.S. sales of OTC acetaminophen-containing products (internal analgesic and upper respiratory categories) decreased by 15% between January 2009 and December 2012 according to Nielsen’s ScanTrack services. This smaller reduction in U.S. sales of acetaminophen observed with Nielsen scan data was supported by a similar finding observed when IRI scan data were used.3 In an analysis completed in 2015 and shared with FDA,4 Rocky Mountain Poison and Drug Center (RMPDC) analyzed the rate of acetaminophen adverse events using data from the National Poison Data System (2007–2013) similar to what was done by Major et al. Similar to the Major et al. study, results from the RMPDC analysis demonstrated that the number of acetaminophen exposures reported to U.S. poison centers decreased for both adult single product and multiple product exposures across all exposure reasons (intentional, unintentional, adverse reaction, and other/ unknown), as well as within each product type (OTC single ingredient, OTC fixed combination, and prescription fixed combination products) and across all outcomes (minor, moderate, severe effect). In contrast to Major et al., the RMPDC analysis demonstrated that the rate of reported adverse events with consideration of the amount of product sold was decreasing for each acetaminophen product type. RMPDC calculated that the annual exposure rate per 1 units sold decreased 3.9% for OTC adult single ingredient acetaminophen products and 8.7% for multi-ingredient