Alcohol Hangover: Underlying Biochemical, Inflammatory and Neurochemical Mechanisms

Abstract

Aim

To review current alcohol hangover research in animals and humans and evaluate key evidence for contributing biological factors.

Method

Narrative review with alcohol hangover defined as the state the day after a single episode of heavy drinking, when the alcohol concentration in the blood approaches zero.

Results

Many of the human studies of hangover are not well controlled, with subjects consuming different concentrations of alcohol over variable time periods and evaluation not blinded. Also, studies have measured different symptoms and use varying methods of measurement. Animal studies show variations with respect to the route of administration (intragastric or intraperitoneal), the behavioural tests utilised and discrepancy in the timepoint used for hangover onset. Human studies have the advantage over animal models of being able to assess subjective hangover severity and its correlation with specific behaviours and/or biochemical markers. However, animal models provide valuable insight into the neural mechanisms of hangover. Despite such limitations, several hangover models have identified pathological changes which correlate with the hangover state. We review studies examining the contribution of alcohol’s metabolites, neurotransmitter changes with particular reference to glutamate, neuroinflammation and ingested congeners to hangover severity.

Conclusion

Alcohol metabolites, neurotransmitter alterations, inflammatory factors and mitochondrial dysfunction are the most likely factors in hangover pathology. Future research should aim to investigate the relationship between these factors and their causal role.

INTRODUCTION

Alcohol hangover is defined as the experience of various unpleasant physiological and psychological effects that follow the consumption of high quantities of alcohol. The symptoms occur several hours after alcohol consumption, ∼10h, (typically after a blood alcohol content, BAC, is >0.08%) (Verster et al., 2014). Hangovers can last for several hours or even more than 24 h. Over 47 symptoms of hangover were identified (Penning et al., 2012), the most common symptoms being tiredness, headache, nausea and impaired attention (van Schrojenstein Lantman et al., 2017a, 2017b, 2018). Several of these symptoms also occur during alcohol withdrawal in dependent drinkers, when they are more severe and of longer duration, in addition to other symptoms such as severe anxiety and tension, negative emotional state, sweating and seizures (De Witte et al., 2003). Hangover also cause several neurocognitive impairments related to executive function (impairment of attention, memory and psychomotor skills) (Gunn et al., 2018) as well as everyday tasks such as driving (Mackus et al., 2016). These cognitive impairments can reduce overall performance resulting in increased accidents (Cherpitel et al., 1998), workplace absenteeism (Verster et al., 2010) and reduced productivity (Gjerde et al., 2010). Reduced productivity has a huge overall impact on society; in 2010, the US Centre of Disease Control and Prevention estimated that alcohol hangovers cost ~$179 billion (Sacks et al., 2015).

Despite having a large health and economic impact and the potential for scientific understanding and remediation, little is understood about the biochemical and neurochemical changes that occur (Verster et al., 2010). The symptoms of hangover begin several hour after the drinking session concludes, possibly when BAC falls to very low residual levels (Ylikahri et al., 1974a; Kim et al., 2003a; 2003b; Penning et al., 2010a, 2010b) or even approaches zero (van Schrojenstein Lantman et al., 2016). However, a range of biochemical and neurochemical parameters will have altered prior to the occurrence of hangover symptoms, the intensity of such alterations could reflect the severity of the hangover.

Common factors which may contribute to hangover severity are outlined in Table 1A. However, only a few have been correlated with hangover severity (Penning et al., 2010a, 2010b). Furthermore, Penning et al. (2010a, 2010b) suggested that immune factors might correlate with hangover severity since this was reduced by the inhibition of prostaglandin synthesis. Factors that have been identified as contributing to hangover severity but not being the sole mechanism of action are outlined in Table 1A. Symptoms of dehydration and sleep disturbances are commonly reported in subjects experiencing hangover, when changes in various hormones, e.g. vasopressin and cortisol, electrolytes and glucose content may be involved. Meanwhile genetic factors, e.g. individual difference in the propensity to experience hangover and of being resistant to hangover, are also important. Between 5% (Kruisselbrink et al., 2017) and 23% (Howland et al., 2008) of the population are reported to be hangover resistant and may therefore be worthy of further investigation to identify biochemical and neurochemical differences between such populations after alcohol ingestion. Some studies have indicated that urinary ethanol concentrations significantly correlates with a variety of hangover symptoms, including, nausea, concentration problems, sleepiness, weakness, apathy, sweating, stomach pain, thirst, heart racing, anxiety and sleep problems in hangover-sensitive individuals (Van de Loo et al., 2017). Ethyl glucuronide and ethyl sulphate, non-oxidative minor metabolites of ethanol metabolism (>0.1%) correlated with urinary ethanol concentration but not with overall hangover severity (Mackus et al., 2017b). Others found no correlation between breath alcohol content and hangover severity, in either hangover sensitive or hangover insensitive individuals (Mackus et al., 2018). It therefore remains to be investigated whether changes in ethanol elimination rates may be responsible for reducing hangover severity.

Other compounds within alcoholic beverages, specifically congeners, are also known to contribute to hangover severity Table 1B. Congeners are substances that are produced during distillation and fermentation, and may contribute to the symptoms of hangover induced by ethanol. High concentrations of congeners are present in red wine and dark spirits, e.g. brandy, and lowest in clear spirits such as vodka. A high content of congeners possibly results in a more severe hangover although the effect of ethanol itself will have a considerably stronger influence on hangover severity than congener content (Rohsenow and Howland, 2010; Rohsenow et al., 2010). The list of congeners includes amines, amides, acetones, acetaldehyde, polyphenol, methanol, histamines, fusel oil, esters and tannins although the contribution of each compound to alcohol hangover is unknown. Methanol is considered to be a major contributor to the symptoms of hangover (Bendtsen et al., 1998). Alcohol dehydrogenase, ADH, will metabolise methanol at a slower rate than ethanol, to form formaldehyde and formic acid both of which are highly toxic and may contribute to hangover. Young-Sup et al. (2005) assayed blood methanol in humans 13 h after ingestion of 1.5 g/kg ethanol and identified a positive correlation between methanol concentration and a subjective hangover scale. A highly significant correlation was found between the presence of headache, nausea and vertigo and urinary methanol concentration in subjects 6–11 h after ingestion of 50–80 g ethanol (Bendtsen et al., 1998), but not in other studies (Mackus et al., 2017a). Further studies on the role played by methanol on hangover severity are clearly warranted.

BIOCHEMICAL AND NEUROCHEMICAL FACTORS INVOLVED IN HANGOVER

Ethanol and its metabolites

Alcohol is initially metabolised by ADH to acetaldehyde which is rapidly metabolised to acetate by aldehyde dehydrogenase (ALDH). Acetate is a precursor of acetyl CoA which can be converted to carbon dioxide and water in the Krebs cycle. Although acetaldehyde is highly toxic, its rapid oxidation by mitochondrial ALDH may preclude any toxicity, although its rate of elimination in the brain is unknown. Approximately 36% of East Asian subjects experience an alcohol-induced flush reaction (Brooks et al., 2009) after ingestion of a small amount of alcohol. It presents as flushing on the face, neck and shoulders (Wolff, 1972), as well as nausea, tachycardia and headaches (symptoms often experienced in hangovers) which is caused by high levels of acetaldehyde due to a deficient gene variant of the mitochondrial ALDH (Brooks et al., 2009). However, other populations, namely Native Americans, Australian, Irish and British also exhibit alcohol-induced adverse effects (tachycardia, flushing and headache) after a small amount of alcohol, although this is not caused by an increase in circulating levels of acetaldehyde (Ward et al., 1994; Slutske et al., 1995; Gill et al., 1999), but may be caused by changes in the metabolism of vasoactive amines including histamine and catecholamines (Ward et al., 1994).

In one study of hangover in human subjects no correlation was evident between blood acetaldehyde levels and hangover severity (Ylikahri et al., 1974b) thereby suggesting that acetaldehyde might only play a minor role in the adverse symptoms experienced during hangover. Another study showed that serum and urine acetaldehyde levels were relatively low and changed minimally following acute ethanol intake (Tsukamoto et al., 1989). Interestingly, in this study serum acetate levels were significantly elevated for at least six hours after ethanol ingestion (Lundquist, 1962; Tsukamoto et al., 1989) which could indicate an increased acetate metabolism in the brain (Volkow et al., 2013). Serum levels of acetate are reported to be in the mM range in contrast to acetaldehyde levels which are in the uM range. Since glucose metabolism may be altered after ethanol ingestion it has been suggested that acetate could be used as an alternative brain energy source. In dialysis patients it has been shown that increased blood acetate levels were associated with the presence of headache (Diamond and Henrich, 1987).

Animal studies have suggested a role for acetaldehyde in hangover. Social interaction normalised after co-administration of 4 g/kg ethanol with rosiglitazone (a peroxisome proliferator-activated receptor (PPAR)-g agonist) which increased ALDH2 activity (Jung et al., 2006), although acetaldehyde levels were not assayed. Inhibiting ALDH2 activity worsened hangover intensity in a chronic headache model in rats (Maxwell et al., 2010). Ethanol-induced effects on the central nervous system may be caused by acetate increasing adenosine content in many tissues, including the brain (Carmichael et al., 1991; Campisi et al., 1997) and inducing reduced motor coordination and locomotor activity (Carmichael et al., 1991) and analgesic effects. These effects were reversed by administration of an adenosine receptor blocker, suggesting that adenosine metabolically produced from ethanol may contribute to its central effects. Additional research using positron-emission tomography (PET)-CT imaging techniques has recently found that colonic acetate can cross the blood brain barrier and be taken up by the brain (Frost et al., 2014). It has also been demonstrated to play a role in appetite suppression which is one of the symptoms which can occur during hangover (Penning et al., 2012). Further research is required to highlight how these metabolites contribute to alcohol hangover.

Neurotransmitters and receptors

Acute ethanol consumption affects a variety of neurotransmitter systems which include: GABA, glutamate, dopamine, serotonin and endocannabinoid systems. The stimulating effects occurring in the initial alcohol intoxication relate to changes in dopamine and BDNF (Bosse et al., 2012), the latter promoting activation of the TrkB receptor and the downstream signalling pathways (Jeanblanc et al., 2009; Logrip et al., 2009). The balance between inhibitory GABAergic and excitatory glutamatergic neurotransmission will also be altered, with reduced levels of GABA and GABA receptor insensitivity (Crews et al., 2005) and increased glutamate and glutamate receptor suppression. Meta-analysis data from published datasets of various regions of the rat brain, where the effect of acute ethanol administration on glutamate and GABA levels had been determined, showed that extracellular levels of glutamate were decreased in the nucleus accumbens, while extracellular levels of GABA and glutamate were elevated in other regions (Fliegel et al., 2013). This was shown to correlate with intensity of withdrawal symptoms. The toxicity of such increased glutamate release plays an important role during the initial stages of alcohol withdrawal after chronic alcoholisation, in the striatum as well as ‘binge drinking’ (Ward et al., 2009). Future research should investigate the role of these neurotransmitter changes in hangover.

Inflammation

Ethanol can profoundly impact inflammatory-related processes, both peripherally as well as in the brain in the absence and presence of an immune challenge. Furthermore, there is a complex relationship between ethanol exposure and the immune response, ethanol dampening cytokine expression in response to an antigen (D’Souza et al., 1989; Kishore et al., 2002; Bhatty et al., 2011) yet exacerbating the cytokine response to bacterial challenge in other circumstances, (Valles et al., 2003, 2004; Qin et al., 2008). Toll like receptors are activated by ethanol in both the periphery, e.g. monocytes and the brain, e.g. microglia, (Alfonso-Loeches et al. 2010), (leading to NFkappaB activation), while ethanol-induced dysbiosis in the gut (Moos et al., 2016), will induce the release of endotoxins from the ‘leaky gut’ to induce inflammation and oxidative stress in both glial and neuronal cells (Crews and Nixon, 2009; Crews and Vetrano, 2018; Karoly et al., 2018) and the release of both anti-inflammatory and pro-inflammatory cytokines in the absence of an immunological challenge.

When ethanol is administered either chronically or voluntarily to rats selective neuronal damage is induced as a result of increased oxidative-nitrosative stress and activation of the inflammatory response, elevating cytokine expression in the hippocampus and cortex (Valles et al., 2004; Tiwari et al., 2009). However, Marshall et al. (2013) reported that there was only partial microglia activation in a chronic ethanol model, suggesting that this was a consequence of alcohol-induced damage rather than the source. An inflammatory response may lead to a variety of symptoms, e.g. nausea, vomiting, headache, confusion, tremor, as well as clinical depression (Harrison et al., 2009), (inducing mood changes and cognitive impairment), and learning and memory deficits (Dantzer, 2004), many of which are evident during hangover, although to a lesser degree than chronic alcoholism and withdrawal. Alteration in specific cytokine levels observed during hangover may contribute to such adverse effects (Dantzer, 2004). Varying results are reported for changes in cytokines during hangover which may relate to the sampling timing, concentration of ethanol administered, route of administration and the different tissues used. Changes in cytokines should be evaluated prior to, during and after hangover to understand the role of each cytokine in hangover. For example, in one study ethanol, 4 g/kg, intraperitoneal, was administered to adult Sprague-Dawley rats, and IL-1 and TNFα increased in the hypothalamus, IL-6 and TNFα were elevated in the hippocampus after 3 h but there were no significant alterations of any of these cytokine in the cerebellum. Increases in each of these cytokines was detected in the liver and spleen (Doremus-Fitzwater et al., 2014) Following 4 g/kg ethanol i.p., IL-6 levels were elevated in the hippocampus, paraventricular nucleus and amygdala after 3 h (Doremus-Fitzwater et al., 2015). Administration of 6 g/kg ethanol to C57BL/6 J mice, induced an increase in IL-4 as BAC approached 0 (Walter and Crews, 2017) although such studies were of brain homogenates such than subtle changes of cytokines in specific brain regions would have been lost. In an animal model of social anxiety, central injection of several cytokines (e.g. IL-1, TNFα) augmented ethanol withdrawal-associated anxiety, possibly through CRF-related mechanisms (Knapp et al., 2011). Clearly, further studies are warranted to delineate changes in specific cytokines during intoxication and as hangover occurs.

Increased levels of cytokine are evident in the blood, mononuclear cells and saliva during hangover. IL8 levels were significantly increased in healthy non-alcoholic volunteers 36 h after an acute alcohol challenge (60 g) (Gonzalez-Quintela et al., 2000) while elevated IL-1Ra and decreased MCP-1 were identified in 20 healthy males during intoxication following a peak BAC of 0.12%. However, when BAC = 0 at 12 h, only MCP-1 was increased (Neupane et al., 2016). Ingestion of 1.5 g/kg of ethanol-induced significant elevation of IL-10, IL-12 and IFN-gamma from the phytohemagglutinin stimulated mononuclear cells during the hangover state, after 13 h, but no significant changes were evident for Il-1b, IL-4, IL-6 and TNFα by comparison to the pre-ethanol value (Kim et al., 2003a). Furthermore, IFNγ and IL-12 concentrations were positively correlated with total hangover scale values. Saliva and urine collected from 36 healthy participants after a night of heavy drinking ~11.6 alcoholic drinks (undefined in grams), showed an increase in saliva content of IL2, IL-4, IL5, IL-6, IL-10, IFNγ and TNFα nine hours after ethanol consumption while increases in urinary levels of IL-4 and IL-6, IL8 were assayed (Raasveld et al., 2015). Interestingly urinary cytokine levels were less pronounced than saliva. BACs were not assayed, and it was assumed that hangover occurred at 9 h. Dividing the subjects according to whether they were alcohol sensitive or alcohol insensitive did not identify any differences in the altered cytokine levels during the hangover period. Assessment of immune status by Immune Function Questionnaire which catalogues the occurrence of sore throat, headache, flu, runny nose, coughing, cold sores, mild fever, warts, pneumonia, bronchitis, sinusitis during the past year (van de Loo et al., 2018a) identified that drinkers with hangovers had significantly higher self-reported overall immune function scores when compared with hangover-resistant drinkers (mean ± SD = 10.5 ± 3.6 vs. 13.1 ± 4.9, P = 0.0001), indicating a poorer immune status. A further study of 341 hangover-sensitive drinkers showed a positive association between mental resilience and perceived immune functioning evaluated after their most recent hangover; however, neither of these parameters correlated with hangover severity. The study concluded that immune status can influence the presence of hangover but not the severity (van de Loo et al., 2018b).

The interpretation as to whether cytokines are altered in these subjects is unclear as it is unknown whether a ‘good’ immune response is associated with higher or lower circulating cytokine levels.

Glial cells play an important role in alcohol-induced neuroinflammation and may contribute to the alcohol hangover. Microglia are the primary mediators of the innate immune system within the brain (Lenz and Nelson, 2018). Increases in microglia numbers, as assayed by the presence of the microglia markers Iba1 and CD68 were quantitated 18 h after ethanol administration when BAC approached 0. Depletion of microglia in male C57BL/6 J mice using colony stimulating factor 1 receptor (CSF1R) inhibitor PLX5622 (Walter and Crews, 2017) resulted in a blunted response in pro-inflammatory cytokines (TNFα and Ccl2) after 6 g/kg ethanol. However, depleting microglia did not reduce gene expression of all neuroinflammatory factors or change the behavioural response (motor impairment measured using a rotarod). The role of the microglia in ethanol-induced cytokine changes was also investigated using minocycline (a microglia inhibitor) in Adult Sprague-Dawley rats (350-450 g) given 4 g/kg ethanol intraperitoneally (Doremus-Fitzwater et al., 2014). The results showed that the application of minocycline failed to reverse ethanol-related changes in cytokine expression. These findings highlight that the microglia may not express cytokines in response to an acute ethanol challenge and further investigation is required to identify the mechanism of this neural inflammation.

Indirect evidence for inflammation contributing to alcohol hangover was indicated by the fact that tolfenamic acid a non-steroidal anti-inflammatory drug (400 mg) (Jayawardena et al., 2017) could reduce the severity of hangover symptoms (nausea, vomiting, thirst, dry mouth, tremor and irritation) and hangover score when administered prior to and after ethanol consumption (Kaivola et al., 1983). Plasma levels of prostaglandin E₂ and thromboxane B were also reduced.

Mitochondrial dysfunction

Ethanol-induced damage to mitochondrial DNA, particularly in the liver has been extensively studied. Since brain cells are reliant on mitochondria, as they operate at the limits of energy availability, even slight damage to mitochondrial may result in elevated free radical production to induce toxicity in a number of brain region, e.g. cerebellum, prefrontal cortex and hippocampus. In various studies, the onset of hangover was found to correlate with: mitochondrial dysfunction as exemplified by alterations in ROS and RNS in the cortex of mouse models (Bustamante et al., 2012) which also correlated with reduced motor performance (Bustamante et al., 2012). Mitochondrial dysfunction was also present in other brain regions (cerebellum) as a result of alcohol hangover (Karadayian et al., 2015). Administration of a free radical scavenger to male mice, e.g. melatonin, protected against mitochondrial dysregulation and the motor impairments that result from alcohol hangover (Karadayian et al., 2014). However, oestrogen blocked the protective action of melatonin (Karadayian et al., 2012). Studies of free radical production and antioxidant status in isolated mitochondria and synaptosomes from male Swiss mice concluded that the synaptic terminals were mainly affected, suggesting bioenergetic dysregulation may contribute to the pathology of alcohol hangover. (Karadayian et al., 2017). Such results indicate redox imbalance may contribute to the motor related symptoms of alcohol hangover in mice. Neuronal mitochondria play an important role in neuronal development and synaptic plasticity, such that their dysfunction can produce apoptosis and ROS both of which have been linked to cognitive impairment (Da Silva et al., 2018).

Auditory evoked response

Brainstem auditory evoked potentials are responses to an auditory stimulus, which reflect neuronal activity in the auditory nerve, cochlear nucleus, superior olive and inferior colliculus of the brainstem. Studies of subjects during hangover, compared with tired controls and normal subjects, identified a reduced auditory threshold during hangover compared with the tired controls which resembles alcohol intoxication (Jarvilehto et al., 1975).

MRI spectroscopy

Advances in neuroimaging have made it feasible to closely characterise changes in brain structure and function and therefore pinpoint the circuitry and regions that may subserve the neuropsychological deficits observed during hangover. Techniques such as functional magnetic resonance imaging (fMRI), magnetic resonance spectroscopy, electroencephalogram, magnetoencephalography and PET offer unparalleled ability to explore regional changes in brain function and metabolism. This has revolutionised the study of many normal and abnormal brain functions and have recently been used to study the effects of alcohol. One fMRI study showed that when a steady state concentration of blood alcohol of 0.08% was achieved, there was increased brain blood flow in thalamus but no other brain regions (Quelch et al., 2017). During the hangover period, 8 h after ethanol ingestion, fMRI identified an increase in activation in the orbital frontal, dorsolateral temporal and hippocampus regions. Participants were able to perform the psychomotor vigilance task correctly which was attributed to compensatory recruitment of prefrontal and temporal structures (Howland et al., 2010). This is an area of great potential and we plan to exploit this technique in our further studies.

CONCLUSIONS

The hangover state is a multifactorial event involving a variety of biochemical, neurochemical events as well as genetic factors which lead to the appearance of hangover symptoms. However, one of the major drawbacks has been establishing the exact time that hangover commences. A variety of animal and human studies have been conducted with a wide variety of ethanol doses administered by a variety of different routes. Further limitations in this field include: reliance on recall, uncontrolled variables, difficulty blinding and the extent of symptoms encountered and their measurements. Human alcohol hangover studies have the significant advantage over animal models of easily being able to assess hangover severity and demonstrate that this correlates with specific behaviours or biochemical markers. However, animal models provide valuable insight into the neuronal mechanisms of hangover.

Despite multiple limitations, several hangover models have identified pathological changes which correlate with the hangover state. Of these, inflammatory factors, alterations in neurotransmitters and receptors, mitochondrial dysfunction and alcohol metabolites are the most likely to be involved in hangover pathology. Upcoming investigations should focus on the contribution of different factors to specific hangover symptoms or the hangover state in general. It is likely that multiple factors contribute to the production of the alcohol hangover, but individual factors may be relevant to a few of the symptoms. It is therefore important that ongoing research continues to assess alcohol hangover as an individual concept but also as a collection of symptoms. Future research should aim to investigate the relationship between these factors and how they influence alcohol hangover as well as the consideration of other confounding factors that may contribute to the severity of alcohol hangover, such as the type and quantity of alcohol consumed, sleep, food intake, genetics, blood glucose levels, dehydration and congeners.

CONFLICT OF INTEREST STATEMENT

None declared.

REFERENCES