ANTIHISTAMINES AND DRIVING-RELATED BEHAVIOR:

A REVIEW OF THE EVIDENCE FOR IMPAIRMENT BY
FIRST- VERSUS SECOND-GENERATION H ₁-ANTAGONISTS

1. INTRODUCTION

1.1 Statement of The Problem

The single largest contributing factor in fatal motor vehicle crashes in the United States is alcohol-induced impairment (AMA Council on Scientific Affairs, 1986). While this has been the case for many years, there also has been an increasing awareness of the traffic safety risks due to the behavioral toxicity of drugs other than alcohol. These include not only illicit drugs, such as cocaine and marijuana, but also medicinal drugs available by prescription or over the counter. In particular, the widespread use of antihistamines (i.e., histamine H ₁ -receptor antagonists, or H ₁ -antagonists for short) presents a particular focus for concern since the 1st -generation H ₁ -antagonists are well recognized for often causing sedation and central nervous system (CNS) dysfunction which can jeopardize safe driving. Moreover, these drugs also have additive effects with alcohol and other CNS depressants. An awareness of such safety risks actually was known more than 50 years ago with the initial introduction of clinically-useful H ₁ -antagonists. For example, in the same year that it received marketing approval by the Food and Drug Administration (FDA), 1946, diphenhydramine (Benadryl) was implicated as a contributing cause of a workplace accident involving impaired driving of a platform cargo truck (Slater & Francis, 1946). And more recently, a study of the association of 3,394 work-related injuries and prior usage of medication (as determined from actual pharmacy records) found a statistically significantly increased risk of injury (odds ratio = 1.5) among users of sedating antihistamines (Gilmore et al., 1996).

Currently, there are more than 60 antihistamines available for oral administration (Maibach, 1988) and many of these are freely available without prescription (i.e., over-the-counter). Commonly, antihistamines are the primary active ingredients in the myriad of cold and flu preparations. Antihistamines also are used individually as 1st -line treatment for the prevalent allergic conditions of rhinitis and chronic urticaria. Other treatment indications for these
H ₁ -antagonists include motion sickness, vertigo associated with Meniere's disease, vascular headaches, and tremors of Parkinsonism. These drugs also are used for their antipruritic (i.e., for itching), antiemetic (i.e., for nausea), antitussive (i.e., for cough), anxiolytic (i.e., for anxiety) and sedative effects (i.e., for insomnia). Such widespread use underscores the increasing scope of the potential safety risks associated with their use by the driving population.

Notably, most states have enacted laws which prohibit driving under the influence of any drug that impairs driving (U.S. DOT, 1996); this, of course, would include sedating antihistamines that disrupt alertness, perception and performance. At the federal level, recent reports have focused on safety standards relating to the use of antihistamines both by workers in the transportation industry as well as by the driving public (cf. Office of the Assistant Secretary for Transportation Policy, Office of Environment, Energy and Safety, 1998). In brief, there have been increasing traffic safety concerns about the possible detrimental effects of medicinal drugs including the widely used antihistamines. But what evidence is there? The answer requires an examination of the problem from several perspectives. As suggested in an early review of alcohol, drugs and traffic safety (Smiley & Brookhuis, 1987; p. 83), “epidemiological studies, laboratory tests of driving-related skills, simulator studies and on-road studies each provide a vital part of the evidence establishing the role of any given substance to traffic safety.” The current review will focus on each of these perspectives, but will only provide a brief summary below of the epidemiological data and its limitations.

1.2 Limitations of Epidemiological Data

The scientific literature regarding impairment of driving-related skills performance by antihistamines consists primarily of experimental studies. These are studies where subjects or patients are administered known doses of antihistamines and then their performance is compared with that under placebo treatment or under comparable antihistamines. The emphasis on experimental studies in this report is due to the paucity of epidemiological studies and the difficulties in interpreting their results.

One of the earliest epidemiological studies of drugs and traffic safety was performed by Skegg, et al. (1979). The authors reviewed the prescription history for more than 43,000 patients over a two-year period. During that period, 57 people in the sample were injured or killed while driving either an automobile, motorcycle or bicycle. For these victims, the drugs prescribed in the preceding three months were compared with those in 1,425 control patients who were selected from the overall sample population as having the same gender, age and prescribing physician. Three of the crash-involved drivers, or 5.3% of the crash group, had been prescribed an antihistamine. Forty-three control drivers, or 3.0% of the control group, had received an antihistamine prescription. The relative risk is 1.8, but obviously this is not significant since it is based on only three injured drivers. It should be noted that in this study, tranquilizers and sedatives as a class showed a statistically significant, relative risk of 5.2.

Ray, et al. (1992) performed a similar study examining the relationship between psychoactive drugs and the risk of a motor vehicle injury crash in elderly drivers in a medicaid program. The advantage of using elderly drivers, over age 65, is that objective data were obtained from the Tennessee medicaid program regarding prescription drug use. Only drivers involved in an injury crash were included in the study, because it was believed that collisions involving only property damage are substantially under-reported and therefore would be less reliable. More than 16,000 people were in the study group which reported 495 injury crashes in a four-year period. Considerable information was available, both from the medical records and the drivers license records. The study employed a multiple regression analysis which controlled for many of these factors. The relative risk of involvement in an injury crash was 1.2 for current antihistamine use. The 95% confidence interval ranged from a relative risk of 0.6 to 2.4. Again there appears to be only a trend (i.e., statistically insignificant effect) to suggest that the use of antihistamines actually results in an increased crash rate. As noted, this study examined an elderly population. Whether or not an interaction exists between the effects of antihistamine use and age, however, has not been determined.

In a 1992 study by Terhune, et al., blood samples were collected from 1,882 fatally injured drivers from seven states during fourteen months in1990 and 1991. The prevalence of antihistamines in body fluid samples from these drivers was 0.6%. In order to determine the significance of the presence of antihistamines, since no comparable control group was available, the authors used a culpability/responsibility analysis which relied on expert raters utilizing police reports of the crash to assign responsibility. Only six drivers had antihistamine present and the responsibility rate was not explicitly stated by the authors, except to indicate that it was not significant.

A 1993 study by Crouch, et al., of 168 fatally injured truck drivers failed to uncover any drivers with an antihistamine present. In contrast, in a study by Warren, et al. (1981) of 768 fatally injured drivers from Ontario, Canada in 1978 to 1979, nine drivers were found to be using antihistamines. A culpability rate analysis indicated a 1.5 culpability rate.

It should be noted that there is considerable difficulty inherent in the attempts to use culpability analysis to compensate for the difficulty of obtaining adequate control groups. Shinar, et al. (1983) compared traffic crash reports by the police with those generated by a university-based investigational team, for example, and found that the police reports frequently omitted important information especially with regard to human factors. In addition, Waller (1982) criticized epidemiological studies of drug effects in driving which relied on culpability/ responsibility analysis because they failed to control for important determinants of driving crash rates such as time and place of collision and characteristics of the drivers. Waller compared studies using culpability analysis with studies utilizing the data of the Grand Rapids alcohol study (Borkenstein, et al., 1964). The Grand Rapids study provided information regarding covariates from both the crash-involved and control groups. This enabled researchers examining the Grand Rapids findings to extract the specific effect of alcohol on crash probability from the influence of variables such as age, gender, drinking practices, etc., which all contribute to an overall crash probability.

It would appear that epidemiological studies involving known populations with verifiable drug use are more likely to produce secure information than epidemiological studies that begin with drivers injured or killed on the road. These latter types of epidemiological studies have no comparable control groups even were we to rush to the scene of crashes, such as was done in the Grand Rapids study. While the Grand Rapids study was able to obtain breath alcohol samples from both crash and control drivers, efforts to obtain blood or urine samples from drivers have been notably unsuccessful. Moreover, even if we had blood samples from both groups, crash and control drivers, interpreting the behavioral implications of plasma drug levels is extremely difficult, as others have already elucidated in detail (e.g., Chesher, 1985).

We typically know the most about drugs detected in fatally injured drivers. However, we also know from studies on alcohol that the probability of being involved in a fatal crash is highly dependent on the blood alcohol concentration (BAC). It is not merely the probability of being involved in a crash that increases with BAC level; but given that you are involved in a crash, there is an additional interacting factor that the probability of death increases with BAC. There is nothing about the studies on antihistamines, however, that would suggest that the magnitude of behavioral effects are comparable with those associated with moderate to higher BAC levels. Thus, the lower magnitude of impairment by the antihistamines would be unlikely to show up in studies of fatal crashes unless the numbers were huge.

We conclude that the epidemiological evidence obtained from studies where 1st-generation antihistamines were commonly used suggests a trend toward some impairment, but not of great magnitude compared with the increased risks associated with alcohol. In summary, given the limitations of epidemiological studies, we believe that experimental studies provide the fundamental method for investigating the direct relationship between a given medication dose and driving efficiency in actual practice. That is, our evaluation of the effects of antihistamines on driving must rest primarily on experimental laboratory studies where we have known dose levels, placebo controls and established experimental response measures. As a background for evaluating such experimental studies of the effects of antihistamines on driving-related performance, a brief description of the clinical pharmacology of the H ₁ -antagonists is presented next.

1.3 Clinical Pharmacology & Issue of Drug Choice

Although the exact mechanisms of action for the histamine H ₁ -receptor antagonists remain unknown, the role of histamine as a neurotransmitter is now firmly established. Histaminergic pathways are widespread in the CNS and appear to be related to mechanisms that support alertness and vigilance during the wakeful state and the balance between wakefulness and slow-wave activity during sleep (Nicholson et al. 1985). Histamine, an endogenous substance first recognized in 1927, has strong vasodepressant and smooth muscle stimulant actions (Garrison, 1990). Considerable research since then has elucidated histamine's roles in mediating the immediate allergic response [via H ₁ -receptors], regulating gastric acid secretion [via H ₂ -receptors] and possibly functioning as a neurotransmitter [via H ₃ -receptors] (White, 1990). The focus of the current review is limited to the H ₁ -receptor antagonists.

The H ₁ -antagonists bind to peripheral and central H ₁1 -receptors and thereby block or, more accurately, compete with histamine's effects. That is, the effectiveness of the H ₁ -antagonist medications is related to the relative concentrations of histamine and its antagonist at the receptor site: an adequately high and frequent enough dosage of the drug is required in order to maintain sufficient concentrations to compete with histamine. An effective dose, however, often is associated with deleterious side effects which include, at least for the classical or 1st - generation drugs, sedation and anticholinergic effects such as dry mouth, nose or throat. The sedative side effects of the 1st -generation H ₁ -antagonists are due to their affinity for central H ₁ -receptors and their liposolubility which enables them to cross the blood-brain barrier. The anticholinergic and other adverse side effects arise from the 1st -generation H ₁ -antagonists' affinity for muscarinic anticholinergic, " -adrenergic, and serotonin receptors.

Newer, 2nd -generation H ₁ -antagonists have been developed in the past decade. Their availability provides allergy patients the choice of new drugs which have little or no side effects such as the sedation and psychomotor impairment often found with the 1st -generation drugs. The 2nd -generation drugs penetrate poorly into the CNS and so are relatively non-sedating, in contrast to the 1st -generation drugs which readily penetrate the blood-brain barrier. Also, the newer drugs have little or no affinity for muscarinic cholinergic, " -adrenergic, and serotonin receptors. This is in contrast to the 1st -generation drugs which do possess such activity. These factors may contribute to the relative lack of adverse CNS or peripheral effects by the 2nd -generation drugs (Simons, 1994). Of note, in the 2nd -generation drugs, there appears to some difference in potential side effects associated with the piperidine class (e.g., astemizole, fexofenadine, loratadine, and terfenadine) versus the piperazine class (e.g., cetirizine).

In sum, the pharmacodynamics and side effects profiles of the 2nd -generation H ₁ -antagonists suggest that these newer drugs offer a safety advantage particularly for patients who drive, pilot aircraft or operate machinery and must avoid the sedation and impaired performance which are commonly found with the 1st -generation drugs. Prior reviews of the experimental studies which have examined the effects of H ₁ -antagonists on performance measures from laboratory tests, driving simulators and on-road driving generally have concluded that the 2nd -generation drugs do pose little or no risk to safe driving. The major prior reviews of those findings are summarized below.

1.4 Prior Reviews of H₁-antagonists

Starmer (1985) provided the earliest review of the evidence concerning antihistamines and traffic safety. He concluded that experimental studies found sedation, impaired performance skills and additive effects with alcohol and other CNS-depressant drugs to be prominent within the heterogenous group of 1st -generation H ₁ -antagonists. He noted, however, that these drugs were seldom identified as causative factors in traffic crashes, possibly due to inadequate reporting. Finally, the several newer, or 2nd -generation H ₁ -antagonists available for study at that time all appeared to have little CNS effect and so presented less risk of impaired driving.

More recent reviews have included those by Rombaut & Hindmarch (1994), Hindmarch (1995), and Adelsberg (1997). The most comprehensive evaluation, however, is provided by Simons (1994) who reviewed the comparative safety of the 1st - and 2nd -generation H ₁ -antagonists in terms of CNS function as well as for cardiovascular adverse effects (specifically seen with some of the newer drugs). Simons, as other reviewers, concluded that the 2nd -generation H ₁ -antagonists are relatively devoid of sedation and CNS impairment, and so they clearly do provide a better “benefit-risk ratio” than do the 1st -generation drugs. Nonetheless, most reviewers also noted that the findings for cetirizine, a 2nd -generation drug, were rather mixed, with some reports of sedation and performance impairment on laboratory tasks as well as on actual driving. The prior reviews also emphasized the difficulty in evaluating the safety profiles of a given drug since the doses, tasks and measures across the studies varied widely.

1.5 Focus of Current Review

Over five years have passed since the most comprehensive review of antihistamines' effects was published (Simons, 1994). Thus, the present review was undertaken to provide a current status of the experimental evidence for impairment of driving-related skills by 1st - versus 2nd -generation H ₁ -antagonists. Importantly, many more studies of the 2nd -generation drugs have been published during this time. Hopefully, these newer studies have employed refined methods and more sensitive measures to detect drug-induced sedation and impairment. Also of note, Simons' (1994) review included approximately 50 controlled studies which compared drugs from the two generations in a single design. However, there are many more studies of the H ₁ -antagonists if one also considers experiments which only examined drugs from one generation or the other. For example, the 1st -generation H ₁ -antagonists often are included as a positive control drug in studies of various drugs other than the antihistamines. Also, some study designs test only a single drug, from the 1st - or 2nd -generation, against a placebo control.

The purpose of the current review is to summarize and evaluate the results of experimental studies measuring the effects of 1st - and/or 2nd -generation H ₁ -antagonists on behavioral and cognitive performance skills relevant for driving. Measures of subjective sedation also are evaluated but only if they were part of a study primarily investigating behavioral or cognitive effects. That is, this review did not include clinical trials which were limited only to reported adverse effects or subjective ratings. Alcohol's effects on driving-related performance have been studied extensively and can be used as benchmark to evaluate the traffic safety profile of medicinal drugs. Thus, for consistency and comparison, the current review organized the performance measures generally within the same behavioral categories as employed in the first author's prior reviews on alcohol's driving-related effects (Moskowitz & Robinson, 1988; Moskowitz and Fiorentino, 2000). Finally, studies investigating acute and chronic doses were considered for this review, whereas studies of drug-alcohol (and drug-drug) interactions were not included since such studies were more limited in number.