Pediatric Fatalities Associated With Over the Counter (Nonprescription) Cough and Cold Medications

Introduction 

The increasing use of nonprescription (over the counter) cough and cold medicines is a worldwide phenomenon. Sales of these products is approximately $3.5 billion annually in the United States.1 Approximately 4 million children younger than 12 years are treated with nonprescription cough and cold products each week in the United States.2

Cough and cold products have been sporadically associated with severe toxicity and death in children. The American Association of Poison Control Centers reported a total of 64,658 exposures to cough and cold products in children younger than 2 years in 2005; of these, 28 (0.04%) were associated with a major adverse effect or death.3 Recent reports of death associated with these products have brought this issue to the attention of regulatory bodies.4

The Food and Drug Administration (FDA) convened an external advisory committee to evaluate the efficacy and safety of nonprescription cough and cold products. On October 19, 2007, the committee recommended that the use of these medications be prohibited in children younger than 6 years. Despite the advisory committee's recommendations, a poll performed by National Public Radio, the Kaiser Family Foundation, and the Harvard School of Public Health after the recommendation reported that 20% of parents with children younger than 2 years and 30% of parents with children aged 2 to younger than 6 years plan to use cough and cold medicines for their children.5

A risk evaluation of the use of cough and cold products in children is needed but has been difficult because the information publicly available consists of a small number of case reports and retrospective case series. To evaluate the role of nonprescription cough and cold medications associated with fatal cases, an expert panel was convened to compile all available pediatric fatalities associated with cough and cold medications, assess the causal relationship to the medications involved, estimate the dose involved, and identify factors contributing to the poisoning.

Materials and Methods 

The panel assessed all reported pediatric fatalities that could be gathered from 5 sources (Figure 1): the National Poison Database System of the American Association of Poison Control Centers (January 1983 through June 30, 2007), adverse event reports submitted to the major manufacturers of nonprescription cough and cold products (1980 to 2007), the FDA briefing materials for its advisory committee meeting on October 18 to 19, 2007,6 a Citizen Petition to the FDA submitted March 1, 2007,6 and the medical literature (1950 to 2007). The medications evaluated included the most commonly used nonprescription cough and cold ingredients: antihistamine (brompheniramine, chlorpheniramine, diphenhydramine, doxylamine), antitussive (dextromethorphan), expectorant (guaifenesin), and decongestant (pseudoephedrine, phenylephrine). Because the searches were based on ingredient, reports involving prescription medications were also identified. Cases involving prescription medications were included in the evaluation process because they often involved both nonprescription and prescription medications and helped characterize the toxicity of the individual components.

View Large Image Download to PowerPoint

Figure 1. Disposition of fatal cases considered by expert panel. *See methods section for definitions. AAPCC NPDS, American Association of Poison Control Centers National Poison Data System; FDA, United States Food and Drug Administration.

The following medical literature databases were searched: MEDLINE (1950 to July 2007), using Ovid and PubMed, EMBASE Drugs and Pharmacology (1980 to third quarter 2007), and International Pharmaceutical Abstracts (1970 to July 2007). The search was conducted by querying each drug separately, limiting to human and English language, and then combining with death by using the Boolean operator AND drug name: brompheniramine, chlorpheniramine, dextromethorphan, diphenhydramine, doxylamine, guaifenesin, phenylephrine, and pseudoephedrine.

From each source, a case was included if the event occurred in the United States, the subject was younger than 12 years, the outcome was death, and one or more of the 8 cough and cold ingredients of interest were identified by medical history or at postmortem analysis. There were numerous duplicate cases, which were combined to present a complete record of the event to the expert panel. Each fatal case was first independently abstracted by 2 trained abstractors who were not involved in the consensus process. The abstractors were nonmedical personnel who made no medical judgments on the cases, but simply evaluated the case for inclusion and organized the data in a systematic fashion across disparate data sources. The abstracted fields included date of death, case source, demographics, product names, individual active ingredients, dose, duration of exposure, drug concentration (antemortem and postmortem), concomitant medical conditions, and results of scene investigation. Disagreements were resolved by discussion with their supervisor. When information was inconsistent, all reported data were captured and the panel was responsible for determining how the inconsistent information affected the interpretation of the case. The study was exempted from review by the Colorado Multiple Institutional Review Board.

The panel was formed of a nonvoting moderator and 8 experts in the fields of pediatrics (I.M.P., R.E.K.), pediatric critical care medicine (W.B.), and pediatric toxicology (B.H.R., G.R.B., W.B.), as well as clinical toxicology (A.S.M., B.H.R.), forensic pathology (D.C.W.), and forensic toxicology (R.B.P.). Selection of panel members was based on evidence of previous research and clinical experience involving the use of cough and cold preparations in children. The panel reviewed the case abstract, as well as all source materials on each case individually before the meeting. All decisions by the panel were formed during one face-to-face meeting and one additional conference call. Panel members were asked to base their decisions on the entire body of information available for each case. Each case was then debated by the group until consensus was reached.

The panel classified each case by using explicit definitions to assess the causal relationship:

Definitely related to medication: History of ingestion or drug concentrations consistent with exposure, clinical course consistent with exposure, and no other cause of death evident.

Likely related: History of ingestion or drug concentrations may be consistent with exposure, clinical course consistent with exposure, other cause of death possible, drug may have been secondary cause of death.

Possibly related: History or drug concentrations may be consistent with exposure, clinical course unknown, other cause of death possible or unknown, drug may have been secondary cause of death.

Unlikely related: History of ingestion or drug concentrations may be consistent with exposure, clinical course inconsistent or unknown, other cause of death likely.

Definitely not related: History of ingestion or drug concentrations not consistent with exposure, clinical course inconsistent or unknown, other cause of death evident.

Unable to determine: Not enough case detail was available to evaluate relationship of drug to death.

All cases assigned to the categories of possibly related, likely related, or definitely related were included as “related” for further analysis.

The panel also categorized the dose involved (therapeutic, supratherapeutic, unable to determine). To form this decision, the panel considered clinical information such as the medical history and blood concentrations performed before death and forensic information scene investigation, autopsy findings, postmortem drug concentrations, and the medical examiner's assessment of the cause and manner of death. Full information was not available for all cases. The panel also identified the person who likely administered the medication (adult, child, unable to determine), the intent of administration (therapeutic intent, nontherapeutic intent, unable to determine), and any other factors that contributed to the incident. The term nontherapeutic was defined as the documented use of the medication for reasons other than the treatment of cough and cold symptoms (eg, sedation). A child finding medication while exploring their environment was categorized as child administered.

Results 

A total of 189 US deaths in children younger than 12 years were identified by the search strategy.4, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29 The panel excluded 11 cases (Figure 1). Of the remaining 178 cases, 41 were judged “unlikely related” or “definitely not related,” and there were 19 cases in which the causal relationship could not be assessed. The net result was 118 cases in which the panel concluded that a relationship between the cough and cold ingredient and the fatality was at least possible. Of these 118 cases, 33 (28%) were judged definitely related, 33 (28%) were likely related, and 52 (44%) were possibly related. Cases judged not related typically included both an obvious alternative cause (eg, shock with positive blood culture) and a time course inconsistent with death from a cough and cold ingredient.

Of the 118 cases judged to be related to a cough and cold ingredient, 82 cases involved a nonprescription medication alone, 21 cases involved exposure to both a nonprescription and prescription medication, and the remaining 15 cases involved only a prescription medication. Prescription medications typically included an opioid (eg, hydrocodone, codeine) or prescription antihistamine (eg, promethazine, carbinoxamine), in addition to a nonprescription ingredient. Drugs for which the prescriptive status could not be determined were placed in the nonprescription drug category.

The 82 cases involving a nonprescription drug alone were combined with 21 cases (total 103 cases) involving both a nonprescription drug and a prescription drug for further analysis. Because many cough and cold products contain more than one ingredient, these 103 cases involved a total of 162 index drug mentions judged as at least possibly related. Three drugs accounted for most mentions: pseudoephedrine (n=45), diphenhydramine (n=38), and dextromethorphan (n=36). The remaining mentions involved chlorpheniramine (n=17), brompheniramine (n=13), doxylamine (n=7), and phenylephrine (n=6). There were no mentions of guaifenesin judged to be related to death.

Male children were involved in 53 (62%) of 86 cases in which the sex was reported. Most of the 17 cases with unreported sex occurred in children younger than 2 years. Age at the time of death ranged from 28 days to 10 years. When stratified by age groups defined by FDA-approved labeling, 77 (75%) cases involved a child younger than 2 years, 22 (21%) involved a child aged 2 to younger than 6 years, and 4 (4%) involved a child aged 6 years to younger than 12 years (Figure 2). Except for the age distribution, there were no apparent differences between age groups. All groups contained a variety of cough and cold ingredients, involved both male and female children, and involved high blood concentrations of cough and cold ingredients (when levels were available). The cases involving children aged 6 to 11 years included a case of diphenhydramine and smothering by a pillow and a case of intracranial hemorrhage involving a product containing phenylephrine and phenylpropanolamine.

View Large Image Download to PowerPoint

Figure 2. Age and sex distribution in fatal cases judged related to a nonprescription cough and cold medication (n=103).

The site of exposure (eg, child's home, daycare, home of babysitter, or health care facility) was reported in 59 of the 103 cases. For cases in which the site was reported, the medication was administered in the child's home in 45 cases (76%), a daycare facility or home of a babysitter in 12 cases (20%), and a health care facility in 2 cases (3%). Characteristics of the daycare facility (eg, licensure) could not be determined. All cases occurring in a daycare facility or babysitter's home involved a child younger than 2 years.

The panel concluded that the child received a supratherapeutic dose in 88 of 103 cases. In the remaining 15 cases, the panel concluded the information available was insufficient to estimate dose (Figure 1). In 38 of 103 cases, the medication contained multiple cough and cold ingredients. In 18 cases, more than one product containing cough and cold ingredients was involved. The number of doses administered in each case was reported in only 18 cases and was not analyzed further.

Of the 103 nonprescription medication cases, the person administering the medicine was an adult in 79 (77%) cases, was a child (self-administered) in 18 (17%) cases, and was not identified in 6 (6%) cases. The 79 cases of adult administration included 19 with therapeutic intent, 34 with unknown intent, and 26 with nontherapeutic intent (Table 1). In the cases judged to be therapeutic intent or unknown intent, several factors appeared to contribute to the administration of an overdosage: administration of 2 medicines containing the same ingredients, failure to use a measuring device, use of an adult product, use of the wrong product because of product misidentification, and 2 or more caregivers administering the same medication. In the cases of nontherapeutic intent, circumstances involved attempts at sedation and several included apparent attempts of overt child abuse and were under investigation by law enforcement authorities.

Table 1. Contributing factors to fatalities associated with adult administration of a cough and cold medication to a child.

		Expert Panel Assessment of Intent
	Contributing Factor	Nontherapeutic Intent (n=26)	Therapeutic Intent (n=19)	Unable to Determine Intent (n=34)
	Combination product	6	14	13
	More than 1 product with same ingredient	1	5	1
	Product intended for adults administered to child	2	0	1
	Site was daycare or babysitter home	10	1	0
	Incorrect or no measuring device used	3	3	0
	Attempting to sedate child	6	0	0
	Abuse (excludes sedation)	3	0	0
	Homicide suspected	10	0	0

Limitations 

The primary limitation of our analysis is the use of databases containing spontaneously reported information. It is likely that other cases with some association to cough and cold medications have occurred but have not been reported. We addressed this concern with extensive attempts to find cases, including sources not always available such as manufacturer's adverse event databases. The true total number of cases is unknown and likely to be higher than reported herein. Like all retrospective analyses, the data in some cases were limited and the accuracy of the information reported cannot be validated. We attempted to compensate for this characteristic by using an expert panel with diverse backgrounds and by including all cases with any suggestion of cough or cold toxicity. Finally, we addressed only deaths related to the nonprescription cough and cold products rather than all adverse events. In the overall risk-benefit evaluation of these products, death informs only a portion of the assessment.

Discussion 

Our results provide the most comprehensive evaluation available of pediatric fatalities associated with nonprescription cough and cold medications. Our search strategy discovered many fatalities not previously considered in the evaluation of these medications. The expert panel concluded that 103 cases involved a plausible relationship between a nonprescription cough and cold ingredient and death. In cases in which the dosage could be assessed, the panel concluded that evidence of an overdose was present. Deaths after use of cough and cold products in children appear to arise from product misuse rather than adverse effects resulting from recommended doses. There were 3 scenarios in which deaths occurred: adult administration with therapeutic intent, adult administration with nontherapeutic intent, or self-administration by a child.

The age group of children younger than 2 years was the most commonly involved. Their small size may facilitate inadvertent administration of an overdose. It is also harder for young children than for older children to communicate emerging adverse effects to their caregiver. A potential contributor to the predominance of young children is the fact that the package label of nonprescription cough and cold medications does not provide dosing information for children younger than 2 years. Instead, FDA regulation requires the label to instruct the caregiver to call a physician, a difficult and unlikely action for many caregivers in the United States. Based in part on the panel's analysis, cough and cold products for use in children younger than 2 years were voluntarily withdrawn from the US market by the manufacturers.30

In cases judged to involve therapeutic intent, several factors contributed to administration of an overdose: administration of 2 or more medicines that contained the same ingredients, failure to use a measuring device or use of an inappropriate device, use of a product intended for adults, use of the wrong product because of product misidentification, and 2 or more caregivers administering medications to the child. Potential solutions include clear, legible labels that explicitly name the ingredients and improved medication administration devices that clearly indicate the appropriate dosage and cannot be separated from the medication. In the past, not all manufacturers provided measuring devices with their products. Furthermore, measuring devices are not standardized across the pharmaceutical industry, and the device for each product is different. The possibility of repeated inadvertent administration of an overdose exists. The contributions to medication errors must be more accurately identified and standardized solutions implemented.

The deaths related to nontherapeutic intent is a new and particularly concerning finding. In many cases, the caregiver admitted their intent to sedate the child, a use for which these products are neither labeled nor intended. In other cases, the documentation explicitly stated that the child did not have cold symptoms, indicating a different intent of the caregiver. It is important to understand that many of the cases in which the intent was to produce sedation occurred outside the child's normal bedtime. In other words, these were not attempts to facilitate sleep by reducing nighttime distress caused by cough and cold symptoms. To minimize deaths that result from similar attempts, nonprescription labels should clearly advise against the use of these products for sedation. Further, educational interventions targeted at health care providers, parents, and child care facilities could reduce inappropriate use of these products.

The fatalities involving malicious intent are particularly challenging because it is unlikely that label changes or restriction of access would effectively deter perpetrators. Other inappropriate means of controlling a child's behavior would remain available: adult formulations of the same ingredients contained in children's cough and cold medicines, other medications and chemicals in general, and other methods entirely (ie, physical abuse). Notably, the age, sex, and location of death associated with cough and cold medications is similar to those in cases in the child abuse literature.31 It is recognized that the extent of fatal child abuse is underestimated.32 One prudent intervention could be to standardize investigations of pediatric deaths and to routinely include medication screening in these investigations. The concept of child fatality teams was developed to address the recognized inadequacy of current systems in identifying the cause of unexpected death among children. A child fatality team provides a standardized evaluation of each pediatric death.33 A small contribution from poisoning has been recognized; however, it is clear that many of these cases require blood testing for detection, which many postmortem examinations do not include routinely. After excluding suffocation and sudden infant death syndrome, the Arizona Child Fatality Review Program reported 48 unexpected infant deaths in Arizona in 2006. Ten of 21 infants who had autopsy and toxicology tests performed had positive test results for cough and cold medications.34 Postmortem viral diagnostic studies should also be encouraged to identify the role of serious viral infections such as respiratory syncytial virus. Children abused with medications likely compose part of the hidden cases of child abuse. Improved investigation of child deaths would allow more accurate ascertainment of cases and perhaps identify additional risk factors.

The results of Schaefer et al35 about adverse events in children complement our results. Using a nationally representative sample of emergency departments in the United States, Schaefer et al35 projected from a sample of 301 cases that there were approximately 7,091 pediatric visits related to cough and cold medications in the United States annually. The study included all adverse events and does not report severity of the event. The predominant age group affected was 2 to 6 years. Our study found that most reported deaths involved children younger than 2 years. Many of the reasons for an adverse event reported by Schaefer et al35 were similar to ours (excess dose administered, wrong medication given, etc), with the important exception that attempts to harm the child or sedate the child were not noted. Some deaths may have been routed directly to a medical examiner's office. Furthermore, deaths after use of a cough and cold product are likely rare, and their sample size may have been too small to capture outcomes such as death.

The nonprescription cough and cold medications offer an important opportunity for public health intervention. The relatively small number of deaths during several decades is somewhat reassuring, given their common use. However, these products are intended for symptomatic relief; therefore, the risk of harm must be reduced as much as possible. An earlier US FDA advisory committee concluded in 1972 that cough and cold medications were effective. This conclusion was reached using the concept of extrapolation of data obtained in adult studies to children.36 This method has been challenged and is still a matter of debate. The most recent advisory committee concluded in 2007 that cough and cold medications lack evidence of efficacy. Additional research addressing both the appropriateness of extrapolating data from adult studies to children and evaluating efficacy in children with better methodology reportedly is planned.

Many of the factors related to inadvertent overdose identified by the expert panel are preventable and interventions could potentially reduce the deaths associated with cough and cold products substantially. A successful public health intervention requires clear identification of contributing factors, implementation of effective tools to address each factor, and ongoing active surveillance to document the effect of interventions and to detect new developments.