Preliminary paper presented at American College of Emergency Physicians (ACEP) Research Forum. ACEP Scientific Assembly; October 12, 1999; Las Vegas Convention Center, Las Vegas, NV.

Introduction

Injuries are the leading cause of death among adolescents and young adults. Sixty-seven percent of the mortality in adolescents is due to unintentional injury.₁ In the United States the association between alcohol use and injury in adults has been well documented.₂,₃,₄ Studies show that approximately 50% of adult patients admitted to a trauma center are under the influence of alcohol, and their mean blood alcohol level is 187 mg/dl, over twice the legal driving limit in most states.₃

Alcohol use and abuse is also a significant public health problem in the adolescent age group. According to the most recent statistics from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) survey, 26% of eighth graders, 40% of tenth graders, and 51% of twelfth graders reported drinking alcohol within the past month.₅ The Youth Risk Behavior Surveillance System (YRBSS) 2001 results showed that 47% of all students in grades 9-12 had one or more drinks at least once in the 30 days preceding the survey.₆ Adolescents are at high-risk for unintentional injury as they indulge in impulsive risk-taking behavior, and alcohol use increases this type of behavior.₇ Previous studies of adolescent trauma have shown wide variations in the prevalence of alcohol use. Results range from as low as five per cent ₈ to as high as forty-five per cent.₉ These variations have been attributed to multiple factors such as geographic location, selection bias, and methods of testing for alcohol use. Studies on the effect of blood alcohol on severity of injury and mortality have also shown varying results.₁₀

The importance of identifying adolescents who use alcohol and present in the emergency department has been previously demonstrated. Maio et al showed that injured adolescents who had ingested alcohol had a greater probability of having a psychiatric history and more frequently had a prior or subsequent injury.₁₁ In a subsequent study, some of the same authors also showed that alcohol abuse is a significant problem among injured adolescents regardless of the severity or mechanism of injury.₁₂ These studies indicate that the emergency department is an important area where adolescents who are at risk for injury and alcohol abuse can be identified.

The purpose of this study is to attempt to determine the prevalence and distribution of alcohol use in both fatal and non-fatal trauma, across a large geographic area using a statewide trauma database. A high prevalence would strengthen the argument for universal blood alcohol screening for all injured adolescents. Injured adolescents under the influence of alcohol are likely at high risk for dependence, and early identification and interventions are of the utmost importance. The study will also examine the association of alcohol use and injury severity and mortality in our population.

Methods

An ASCII text database was obtained from the Illinois Department of Public Health (IDPH) on all patients presenting to an Illinois Level I or Level II trauma facility from January 1, 1997 through December 31, 1998. All the Illinois Trauma Registry (ITR) data was received in ASCII text format, merged into a Paradox (Corel; Toronto, Canada) database for querying and partitioning, and then re-formatted for population statistical analyses by SPSS (SPSS, Chicago, IL). During the study period there were 72 hospitals designated as trauma centers that reported their data to IDPH. Included on the trauma registry is an etiology code (E-code) for each patient entry. The E-code defines, in detail, the mechanism of injury. The database was designed to include only those patients who met local “trauma bypass” criteria to a level I or level II facility. It contains data on all patients who had formal consultations and/or admissions to the trauma surgery service. Therefore, it does not include most minor trauma that was treated and released from the emergency department as well as isolated (orthopedic) injuries cared for by other services. The database was extremely complete (> 99.9% for each parameter/variable).

The study was a retrospective analysis of all patients 12 – 24 years of age. The adolescent group was partitioned into two groups, 12-15 years old and 16-18 years old. This was based on the typical age of transition into adulthood, independence, and an age at which alcohol consumption might increase. For comparative purposes, the 19-24 year old age group was also analyzed. Each age group was then partitioned by gender, etiology (E-code), Injury Severity Score (ISS), mortality, number tested for blood alcohol, and blood alcohol concentration (BAC). Patients who had a positive BAC were considered to have been using alcohol or “alcohol users”. Relationships among variables were identified using SPSS 10.1 coefficient correlations along with multivariate regression analyses. Statistically significant differences were determined when the probability value (p) was equal to or less than 0.05.

The study was Institutional Review Board exempt because it involved the use of existing publicly available data. When received from IDPH, the data were labeled in such a manner that subjects could not be individually identified. All demographic information such as patient name, telephone number, address, or any type of number that could be used to link the data to a medical record was omitted.

Results

There were a total of 78,252 trauma patient entries during the two-year study period. Of these, 7465 (9.5%) were 12 -18 years old and comprised our adolescent study group (Table 1).

Figure 1

Table 1: Blood alcohol testing by age group

Seventy-one percent were males and 29% were females. There were 3018 (40.4%) patients who were 12-15 years old and 4447 (59.6%) that were 16-18 years old. Of the patients 12-15 years old, 555 (18.4%) were tested for alcohol and 95 (17.1%) tested positive. Of the patients 16-18 years old, 1962 (44.1%) were tested for alcohol and 668 (34.0 %) tested positive. In comparison, for the 19-24 year old group, there were 8461 subjects. Of these, 4390 (51.9%) were tested for alcohol and 2137 (48.7%) of these tested positive. The adolescents in both groups were statistically less likely to be tested than the 19 -24 yr old patients (p <0.05).

Alcohol use and testing rates was analyzed by gender (Table 2). For the 12-15 year olds and the 16-18 year olds there was no statistically significant difference in testing rates by gender. However, the 16-18 year old males are more likely to test positive than their female counterparts (p<0.001).

Figure 2

Table 2: Percentage tested by gender

The pattern of blood alcohol use and testing and was studied for the seven most common mechanisms of injury - vehicle drivers, vehicle passengers, bicycle injuries, pedestrians, falls, self-inflicted injuries and assaults (Table 3). Overall, for adolescents, it was observed that there was a trend towards higher testing rates for assaults and road injuries. Across all age groups, assault victims had the highest testing rates. Those injured due to falls were tested the least. Wilcoxon rank correlation tests found no correlation of mechanism of injury between testing rates and actually testing positive. The mechanisms of injury tested most often did not yield the highest rates of testing positive.

Figure 3

Table 3: Mechanism of injury and alcohol use

Table 4 shows the mean ISS and percent mortality rate in association with frequency of alcohol testing and frequency of alcohol use for each age group. For the 12-15 year olds, the ISS was significantly lower in those testing positive compared to those testing negative (p< 0.05). This difference was not present in the other two age groups. There was, however, a significantly lower mortality in those who tested positive across all age groups (p<0.05). For the entire group of 16-18 year olds, as well as the 16-18 year olds testing positive, the ISS was significantly higher than for those 12-15 years old (p<0.05). The 16-18 and 19-24 year old groups were not significantly different in mean ISS or in mortality.

Figure 4

Table 4: Injury severity, mortality rate and alcohol

Discussion

The impact of trauma on our society is enormous. Injuries are the leading cause of death for those one to 37 years old.₁₃ There are more than 140,000 deaths per year due to injury, and for each death there are two additional patients with permanent and disabling injuries.₂ Annually the cost of all unintentional injuries is estimated at over $400 billion, of which medical expenses account for almost $75 billion.₁₄

For adolescents, injury is the most common cause of morbidity and mortality and accounts for one-half of their deaths.₈ Each year more adolescents are killed or injured due to motor vehicle accidents, homicides, suicides, or drowning than any illness.₁₅ Although overall mortality has declined for this age group during the past 30 years, mortality has actually risen for the injured adolescent population.₁₆

Alcohol use and injury are closely related and has been well documented in the adult population. In the United States, alcohol is the major risk factor for all injuries.₂ Prior studies of alcohol-related trauma have focused on hospitalized adults and have shown about one-third to one-half of the patients have positive alcohol screens and abuse/dependence rates that are nine times higher than the general population.₄,₁₇,₁₈,₁₉,₂₀

The role of alcohol in adolescent trauma is less clear. In adolescents, about one-third of trauma-related deaths involve alcohol.₂₁ However, a paucity of data exists regarding the association of alcohol use and non-fatal injuries. Previous studies have focused mainly on fatalities or hospitalized patients, and have shown a wide variation in the incidence of alcohol use, ranging from 5% - 44%.₉

Two studies of hospitalized adolescent trauma patients found that about one-third tested positive for the use of alcohol or drugs. ₁₁, ₂₂ In one of these studies, the investigators found that patients testing positive were more likely to have had a prior injury.₁₁ In the other, the authors found that those with intentional injuries were more likely to test positive than those with unintentional injuries.₂₂ A study by Mannenbach et al. from a Milwaukee trauma center reported that 40% of injured adolescents tested positive for alcohol. They found no relationship between mechanism of injury and alcohol use.₂₃ Two other recent studies showed a much lower prevalence (4 - 5%) of alcohol use among adolescent trauma patients. ₈, ₁₂ All of these studies sampled small numbers of patients (< 300), were from a small number of centers, and suffered selection bias.

Only one prior study by Porter utilized a statewide data base (Pennsylvania) and examined all injured adolescents admitted to a trauma center over a one year period (1996).₁₀ In this study of 4309 patients, 50% of the 12-17 year old group was tested for alcohol compared to 70% of the 18-20 year olds and 69% of controls (21-25 years old). Thirteen percent of the 12-17 year olds tested positive, as did 30% of the 18-20 year olds and 47% of the controls. In addition, Porter demonstrated a sharp rise in the rate of alcohol ingestion between 16 and 17 years of age. The difference in the prevalence of alcohol use between the two age groups was attributed to obtaining a driving a license at 16 years - the car affording mobility and privacy. Other factors that have been attributed to influence this rise in prevalence across age groups is the increase in the risk taking and sensation seeking seen during this age.₂₄ Though Porter sampled a much larger number of patients form over the entire state of Pennsylvania than previous studies, selection bias excluded all but the most seriously injured patients. Pennsylvania trauma center reporting criteria are death, initial admission to the ICU, length of stay > 3 days, or transfer to a trauma-specialty center.

Our study encompassed two years and included 78,252 patients from across the entire state of Illinois. Although it excluded minor injuries treated and released from the ED, inclusion criteria were much more liberal than prior studies. The patients included all admissions to the trauma service, regardless of the length of stay. It also included all formal trauma consultations regardless of whether the patients were admitted to the hospital.

Our results show that only a very small percentage of adolescent trauma patients are being tested for alcohol use. The rate of testing was only 33.7% for the 12-18 year olds, much lower than the 50% rate demonstrated by Porter and the 48%-92% rates shown by others.₈, ₉, ₂₂ In the 12 –15 year old age group, the rate of testing was only 18.4%. However, despite the low testing rates, the overall percentage of alcohol use in the entire adolescent population was 30.3%, 17.1% for the 12-15 year old group and 34% for the 16-18 year old group. This high prevalence reiterates that the legal age of drinking is not a sufficient barrier to the use of alcohol by adolescents who present with trauma to the ED, and suggests the need for more widespread or universal testing.

The decision whether to test a patient for alcohol is not uniform and probably based on clinical signs of intoxication as well as the nature of the injury. Clinical signs of intoxication, however, are often very difficult to detect, especially at lower concentrations. Gentiello published a study in which trauma center physicians could clinically identify only 23% of acutely intoxicated trauma patients.₂₅ This inability to clinically detect intoxication underscores the need for more vigilant screening of adolescent trauma patients. In fact, screening of all hospitalized trauma patients has been recommended by the Institute of Medicine and the Substance abuse and Mental Health Services Administration. Also, in a prospective study by Rivara et al, it was shown that trauma patients were at an increased risk for readmission or relapse and, therefore, all should be screened for alcohol use.₂₆

Our findings demonstrated a positive correlation between age and mean alcohol level that extends from the adolescent age groups into the 19- 24 yr age group. For those testing positive, the mean BAC was 88 mg/dl for the 12-15 year olds, 100 mg/dl for the 16-18 year olds, and 133 mg/dl for the 19-24 year olds. Though the blood alcohol concentrations were lower for the younger ages indicating less alcohol consumption, one study has raised the concern that adolescents have a much lower tolerance to the effects of alcohol.₂ An interesting observation to be noted was that the mean blood alcohol concentration in all age groups exceeded the “legal” limits of intoxication in most states.

With regard to gender, there was a trend in the data for adolescent males to be tested more than females. However the difference was not statistically significant for both adolescent age groups. We hypothesize that the trend toward a higher percentage of males tested could be due to sociological bias by emergency physicians to test more males. In the 12-15 year old group, both males and females were equally likely to test positive, unlike the 16-18 year old males who are more likely to test positive than females. Gender differences in the rate of testing and alcohol ingestion specific to adolescent trauma patients have not been previously reported in studies of adolescent trauma. These results argue for an unbiased testing strategy for adolescent trauma patients in the emergency room.

We examined the testing rates and prevalence of alcohol use for adolescents presenting with the seven most common mechanisms of injury. Adolescents injured by an assault or as a vehicular passenger had tendency toward being tested more often in both age groups. Those with self inflicted injuries and falls had a tendency to test positive more often in the 12-15 and 19-24 year old age groups respectively. Overall, the highest rates of alcohol use in the injured adolescents were seen in those that were injured in road accidents, fall and self-inflicted injury. In the 12-15 year old group, however, alcohol use was highest for those with self inflicted injuries (27.8%). In the 16-18 year age group alcohol use was at its highest among those with falls, bicycle accidents, and vehicle drivers. Porter ₁₀ and Spain et al.₇ both observed the highest rates of alcohol use in vehicle drivers and falls. However, both of these studies had a more rigorous inclusion criteria and a possible selection bias towards patients with more severe injury. Overall, in our study patients, there was no correlation between the mechanisms of injury with highest prevalence of alcohol use and the most commonly tested groups. Although falls and self-inflicted injuries were least tested, they were more likely to test positive. In addition, pedestrians were frequently tested but least likely to show presence of alcohol. This suggests that emergency physicians should adopt a broader spectrum policy for testing of injured adolescents and change their pattern of testing towards those mechanisms more often associated with alcohol use i.e. self inflicted assaults, vehicular passengers, and driver-related injuries.

When looking at correlations between alcohol use and injury severity score (ISS), we found that 12-15 year olds testing positive had a significantly lower ISS (p<0.01) compared to those testing negative. This difference was not present in the other age groups. Porter did not observe a difference in mean ISS scores between alcohol positive and alcohol negative patients, but he showed a trend towards lower ISS in the alcohol positive group. Linear regression analysis of the Porter data did show that both the presence and amount of alcohol was associated with a small, but statistically significant, lower ISS.₁₀

Our results showed a significantly lower mortality for those testing positive compared to those testing negative in the 16-18 year old group (p<0.001) as well the 19-24 year old group (p<0.001). In the 12-15 year old group mortality was zero in those testing positive. However, possibly because of the low overall testing rate in this group, the difference was not significant. The association of alcohol use with decreased mortality was also observed by Porter.₁₀

Contrary to these results suggesting a possible protective effect of alcohol in our population, Hicks et al.₉ had found an increase in injury severity with alcohol use in adolescents. Other studies of general emergency department patients by Ward et al.₂₇ and Brikley and Shepherd ₂₈ showed no association between severity of injury and alcohol intoxication. These studies are difficult to compare because of the varied ages and selection criteria of the study populations. Because of the large number of patients in our study that were not tested and its retrospective design, we cannot definitively answer this question, and further prospective studies are needed.

Our study is limited in that it is a retrospective review of a trauma database. Only a relatively small percentage of the adolescents were tested for alcohol. The biases that might be involved in this selection process are unknown. Adolescents also are known to have poor utilization of health facilities and may not present to the hospital, especially if they have mild injuries associated with alcohol use. Therefore, we can only surmise how many of those patients not tested were using alcohol. Prospective studies could more accurately measure the true frequency of alcohol use in adolescent trauma patients. Outcome studies of these patients could then measure the impact of improved detection and prevention strategies.

Conclusions

Our results suggest that alcohol use may be a significant factor in adolescent trauma care. Adolescents at risk for alcohol dependence need to be identified and strategies implemented for those in potential need of counseling and other rehabilitative measures. We found a relatively high prevalence of alcohol use by injured adolescents presenting to trauma centers in the State of Illinois. More than one-third of the injured adolescents who presented with injury tested positive for alcohol. Despite this, overall testing rates for adolescents are significantly lower than for young adults presenting with injury. Only one-third of all eligible patients are tested for alcohol compared to over 50% for the younger adult population. This study showed that mechanisms of injury that have the highest testing rates do not correlate with the mechanisms associated with the highest detection rates. Until more definitive prospective data is available, emergency and trauma physicians should adopt a broader policy of alcohol testing for injured adolescents.