The Alcohol HangoverFREE

We searched the MEDLINE database (1965 to 1999) for epidemiologic, physiologic, and economic studies about alcohol hangover, using the terms hangover, alcohol withdrawal, and alcohol intoxication. Bibliographies of selected articles and personal communication with selected authors were used to extend the search. Articles were screened for their relevance to the specific topic of alcohol hangover on the basis of the title and abstract. All studies that specifically referred to the alcohol hangover were included in the data analysis. A formal meta-analysis was not performed because the few clinical trials were not similar in design or hypotheses.

There is no consensus definition of veisalgia (“alcohol hangover,” from the Norwegian kveis, or “uneasiness following debauchery,” and the Greek algia, or “pain”). Most descriptive (7, 9) and experimental studies (10-12) have identified a set of common symptoms: headache, diarrhea, anorexia, tremulousness, fatigue, and nausea (Table 1). Objective criteria have focused on decreased occupational, cognitive, or visual–spatial skill performance or on alterations in hemodynamic and hormonal measurements. Although tachycardia, tremor, orthostasis, cognitive impairment, and visual–spatial impairment are frequently observed (12, 14), they do not capture the overall experience for the patient. This remains subjective, varying from person to person and from episode to episode (15). To permit a uniform discussion, we defined hangover as the presence of at least two of the symptoms in Table 1, occurring after the consumption and full metabolism of alcohol with sufficient severity to disrupt the performance of daily tasks and responsibilities.

Although hangover might be considered trivial—just deserts for the overindulgent—it has substantial economic consequences. A recent British study noted that alcohol use accounted for 2 billion pounds ($3.3 billion U.S.) in lost wages each year, most of which resulted from work missed because of hangover (16). Alcohol costs in Canada amount to $7.5 billion each year, and $1.4 billion is lost each year because of decreased occupational productivity caused by hangover-like symptoms (17). Studies in other countries have yielded similar per capita estimates for the annual cost of alcohol ingestion: Australia, $3.8 billion (18); New Zealand, $331 million (19); and the United States, $148 billion (20). Greenfield recently estimated the average annual opportunity cost due to hangover as $2000 per working adult (20).

In the workplace, the greatest cost incurred by alcohol is the decreased productivity of affected employees as a result of hangover-related absenteeism and poor job performance (20, 21). In Finland, which has a population of 5 million persons, more than 1 million workdays are lost each year because of hangover (22). Light-to-moderate users of alcohol (0 to 3 drinks per day for men and 0 to 1 drink per day for women) (17) account for most of the lost-work costs because they make up most of the work force (23). Fifty-four percent of all alcohol-related problems in the workplace are caused by light drinkers, and 87% are caused by light-to-moderate drinkers. The primary morbidity that affects light-to-moderate drinkers is the hangover, not the long-term consequences of alcohol abuse, such as cirrhosis and cardiomyopathy (20). Chronic alcoholism is responsible for only a small proportion of the total societal cost of alcohol use (24).

Perhaps the most alarming feature of veisalgia is its high prevalence. In a study of college students, 25% of students reported experiencing a hangover in the previous week and 29% reported losing school time for hangover recovery (5). More than 75% of men and women who have consumed alcohol report that they have experienced hangover at least once (25), and 15% experience hangovers at least monthly. Ten percent of British men reported hangover-related problems at work at least monthly (16). Paradoxically, hangover is much more common in light-to-moderate drinkers (70%) (16) than in heavier drinkers (13, 17, 26, 27).

Although hangover may be interpreted as merely uncomfortable, the patient with hangover is at increased risk for injury and poor job performance. Patients with hangover have diminished visual–spatial skills and dexterity, even after alcohol can no longer be detected in the blood. Impairment from hangover has been experimentally demonstrated in pilots (12, 28), people who drive (29, 30), and skiers (31). Managerial skills and task completion are also adversely affected (32, 33). Frequent hangovers have also been shown to be associated with increased cardiac death in patients not known to have coronary artery disease (relative risk, 2.4 [95% CI, 1.0 to 5.5]); however, after adjustment for risk factors and alcohol consumption, this association was no longer statistically significant (relative risk, 1.8 [CI, 0.7 to 4.3]) (6).

The ways in which hangover affects total alcohol consumption are not clearly understood. Many persons believe that hangover is a punishment for alcohol consumption and therefore prevents subsequent alcohol use (34). Hangover has never been shown to effectively deter alcohol consumption, however, and no evidence shows that alleviation of hangover symptoms would result in further consumption (34). In contrast, the discomfort of hangover symptoms may prompt further alcohol intake (for example, the “eye opener”). One study of 178 persons found that “those who experience greater hangover may choose to drink more alcohol in order to relieve these adverse effects” (35). Therefore, successful treatment of hangover could mitigate total alcohol consumption.

Part of the mystery of hangover is the set of ill-defined physiologic characteristics that underlie the syndrome. One theory is that hangover is the first stage of alcohol withdrawal. However, the hormonal and hemodynamic changes seen in hangover are distinct from those seen in alcohol withdrawal (Table 2).

Although larger doses of alcohol lead to more severe symptoms, hangover is not solely dose-related (37). Acetaldehyde, the dehydrogenated product of alcohol metabolism (38), might be responsible for hangover symptoms (10). Congeners, the byproducts of individual alcohol preparations (which are found primarily in brandy, wine, tequila, whiskey, and other dark liquors), increase the frequency and severity of hangover (24, 39, 40). Clear liquors, such as rum, vodka, and gin, tend to cause hangover less frequently, which may explain why patients with chronic alcoholism use these liquors disproportionately. In an experimental setting, 33% of patients who consumed 1.5 g/kg of body weight of bourbon (which has high congeners) but only 3% of those who consumed the same dose of vodka (which has low congeners) experienced severe hangover (41).

The constellation of hangover symptoms (nausea, headache, diarrhea) resembles that seen in conditions related to dysregulated cytokine pathways (for example, in viral infections and after administration of interferon-α). Alcohol alters cytokine production through a thromboxane pathway. Levels of thromboxane B₂ are elevated during experimentally induced alcohol hangover (42), and the administration of tolfenamic acid, a prostaglandin inhibitor, at the time of alcohol consumption has a small prophylactic effect in reducing hangover severity (9).

Several hormonal alterations have been observed in patients with hangover (43-45) (Table 3). Hangover severity is proportional to antidiuretic hormone concentration (46). Alcohol inhibits the effect of antidiuretic hormone on the kidneys, thereby inducing diuresis that is out of proportion to the volume of fluid ingested. As blood alcohol concentration decreases and dehydration persists, the serum level of antidiuretic hormone increases, maintaining water retention in dehydrated patients with hangover. In our clinical experience, hydration attenuates but does not completely relieve hangover symptoms. Serum aldosterone and renin levels also increase with hangover, but unlike antidiuretic hormone, they do not correlate well with hangover severity (47, 48).

The effect of alcohol consumption and hangover on glucose metabolism is incompletely understood. Alcohol seems to inhibit the availability of glucose through an insulin-mediated mechanism (49). Glucagon is increased in acute alcohol intoxication, but its effect during hangover is unknown (50). Cortisol release is also suppressed during acute alcohol intoxication, but this effect does not persist in hangover (51). Levels of thyroid (44, 52) and growth hormone (52) do not change during acute alcohol intoxication or hangover. Both acute intoxication and hangover cause metabolic acidosis (53, 54).

Hemodynamic changes seen in the patient with hangover include increases in heart rate, left ventricular performance (as measured by ejection fraction), and blood pressure (14, 55). The peripheral vasodilatation seen with acute alcohol ingestion is not observed in hangover (14). The increased cardiac work with normal peripheral resistance that occurs in hangover may explain the associated increased cardiac mortality rates (6).

Patients with hangover have a diffuse slowing on electroencephalography (36), which persists up to 16 hours after blood alcohol level becomes undetectable. Decreased auditory evoked responses and psychomotor deficits have also been noted (22). These findings suggest that hangover, which manifests as “diffuse cortical depression,” may be a different physiologic process than “alcohol withdrawal,” which is characterized by general hyperexcitability of the brain (22, 56).

The absence of a standard instrument for hangover assessment makes comparison of treatments difficult (15). Among patients with experimentally induced hangover, an alcohol dose of 1.5 to 1.75 g/kg (five to seven standard cocktails), administered orally over 4 to 6 hours, is almost always followed by hangover symptoms (36, 37, 42, 48, 49, 52). Other factors that increase the severity of hangover include lack of food consumption, decreased quality and quantity of sleep, increased physical activity while intoxicated, dehydration, and poor physical health (28, 30, 38).

Despite the ease in inducing hangover, only a few trials have studied treatment of hangover symptoms. Bogin and coworkers (10) compared propranolol with placebo but found no effect. In their crossover study, 10 healthy college students drank enough alcohol to reach a blood alcohol level of at least 0.1 g/dL. On the night the alcohol was consumed, 5 of the students received a long-acting oral preparation of 160 mg of propranolol; the remaining students received placebo. The study was repeated on a second occasion with a switch in treatment assignment. The morning after each night of alcohol consumption, researchers assessed the vital signs of each participant and calculated a score on a hangover severity scale. Objective analysis of hand tremor, blood pressure, and cardiac rate and rhythm showed no significant differences. On a scale of 1 to 5, the mean (±SD) severity of hangover was 2.9 ± 1.1 in participants who received propranolol and 2.8 ± 0.9 in those who received placebo (CI for the mean difference, −0.84 to 1.04). The authors concluded that propranolol does not produce a clinically important benefit in the symptoms of hangover. However, the study's power was only 50% for the hypothesized primary finding, at a two-sided α value of 0.05 (10).

Two trials have investigated the effect of simple carbohydrates on hangover severity. Seppala and colleagues (29) assigned 40 healthy men to receive one of six treatments. Participants received 1.75 g/kg of alcohol from 6:00 p.m. to 9:00 p.m. and an oral glucose or fructose solution (0.5 g/kg or 1 g/kg). The participants in the two control groups received alcohol or glucose but not both. All patients subjectively assessed their hangover severity and underwent psychomotor tests that determined reaction time, coordination, and attention span. Administration of glucose at the time of alcohol ingestion reduced mistakes by 50% on the choice reaction test (which was given 15 hours later) but slightly increased mistakes on the coordination test. Hangover severity was not affected. The study was limited by the small number of participants (n = 5) in each treatment group. Moreover, the statistical analysis did not account for multiple comparisons. Ylikahri and coworkers (37) used a similar design to study 109 healthy men (1.75 g/kg of alcohol administered orally over 3 hours, followed by 1 g/kg of oral glucose or fructose). Patients who received glucose or fructose had lower serum levels of free fatty acids and ketones but no change in subjective or objective hangover symptoms.

In a nonblinded study sponsored by its manufacturer, an herbal preparation called Liv.52 (Himalaya Drug Co., Bombay, India) decreased hangover symptoms more than placebo (11). Among patients who took Liv.52 at the time of alcohol consumption, the ratio of blood and urine alcohol levels to acetaldehyde increased; this was the postulated mechanism for the observed decrease in hangover symptoms. The investigators hypothesized that Liv.52 blocked the conversion of ethanol to acetaldehyde, which they believed to be the cause of hangover symptoms.

Tolfenamic acid, a prostaglandin inhibitor, was associated with a small improvement in hangover symptoms when administered prophylactically on the night of alcohol consumption. Thirty healthy volunteers participated in a double-blind crossover study, in which they received 200 mg of tolfenamic acid or placebo before sleep. The following morning, participants who received tolfenamic acid reported less headache, nausea or vomiting, irritation, and thirst. Levels of prostaglandin E₂ and thromboxane B₂ were lower in the tolfenamic acid group, and the investigators suggested that these cytokines may be associated with hangover symptoms (9). Other nonsteroidal agents are frequently used to treat hangover symptoms but have not been studied.

Prophylactic vitamin B₆ (pyritinol) reduced the number of hangover symptoms by approximately 50% in one study (57). Seventeen men and women attended two parties and were asked to drink the alcoholic beverage of their choice until intoxicated. Fifty percent of participants received 1200 mg of vitamin B₆ (400 mg at the beginning of the party, 400 mg 3 hours later, and 400 mg at the party's conclusion), and 50% received placebo. At the second party, participants received the alternate tablet. A survey instrument of 20 symptoms was used to assess hangover severity. On a scale of 1 to 10, the mean symptom score (±SD) was 3.2 ± 2.8 with pyritinol and 6.8 ± 3.8 with placebo (P < 0.01). The mechanism of this effect is unknown.

Therapeutic administration of 1 g of chlormethiazole (a psychotropic sedative) reduced symptoms of hangover when compared with placebo. However, it also diminished cognitive performance and had important side effects, such as sedation (58).

Despite the hazards of hangover, the question remains whether treatment of hangover symptoms improves physical health. Although the physiologic and symptomatic impairments seem to be correlated, symptomatic therapy may not improve physiologic outcomes. If a successful treatment were found, patients might experience less discomfort with hangover but retain its visual–spatial, cognitive, and cardiovascular detriments.

However, interviewing patients about their hangover experiences may have clinical relevance. First, hangover is common and underdiagnosed and can have serious physical, psychiatric, and occupational consequences. Therefore, it is a potentially modifiable cause of morbidity and offers an opportunity for substantial symptom relief (59). Particular attention should be paid to patients with cardiac risk factors or coronary artery disease, who may be particularly at risk because of the physiologic stress induced by veisalgia.

Standard alcohol screening, which focuses on detecting alcohol dependency, rarely addresses the frequency and severity of hangovers. Screening for hangover severity and frequency may be used to augment current strategies for early detection of alcohol dependency (3). For example, sons of alcoholic parents, who are at greater risk for alcoholism, have more frequent hangovers (60). The Short Michigan Alcoholism Screening Test has adopted this consideration, noting “hangover severity” as one of the predictive criteria for alcoholism. Newlin and Pretorious (60) have noted that patients with frequent hangovers seem to be at risk for further alcohol dependence. Finally, depression and other psychological disorders are more common in patients with hangover (61).

For the patient with occasional hangovers, recommended interventions include the discussion of potential therapies for alcohol hangover (adequate hydration) and a reminder that 1.5 g/kg of alcohol (approximately 5 to 6 drinks for an 80-kg man and 3 to 5 drinks for a 60-kg woman) will almost always lead to hangover. Patients should also be educated about the physiologic effects of hangover. Most patients understand the impairment inherent in acute alcohol intoxication, but few appreciate the cognitive and visual–spatial impairment that may accompany alcohol hangover (7, 12). Patients who work with heavy equipment or are involved in transportation should be warned of the potential hazards.

Hangover's historical past may predict its future. Homer provided one of the first descriptions of the disorder. A companion of Odysseus, Elpenor, awoke from a drunken sleep, sprang up, and jumped off a roof, falling to his death (62, 63). Of interest, Elpenor “returned from the dead, begging Odysseus to bury his body,” a sentiment we have often heard echoed by patients with hangover. The most extreme form of hangover, a psychiatric dissociation characterized by irrational behavior, has since become known as the Elpenor syndrome.

As a condition warranting physician recognition and treatment, veisalgia, with its cardiac, neurologic, and psychiatric effects, is far more than a mere nuisance. Perhaps the spirit of Elpenor will inspire physicians to prevent and treat hangover, thereby preventing patients from falling to their deaths.