Caffeine Abstinence During The Menstrual Cycle:An Evaluation Of Mood, Somatic, Cognitive, And Psychomotor Effects Of Withdrawal
H Vo, B Smith
H Vo, B Smith. Caffeine Abstinence During The Menstrual Cycle:An Evaluation Of Mood, Somatic, Cognitive, And Psychomotor Effects Of Withdrawal. The Internet Journal of Pharmacology. 2009 Volume 9 Number 1.
Women frequently report the occurrence of various symptoms that occur during the phases of the menstrual cycle. The factors affecting this symptomatology are not yet fully understood, but there is reason to believe that one widely consumed drug, caffeine, may impact these symptoms. In the present study, the effects of caffeine withdrawal across the menstrual cycle on several aspects of menstrual symptomatology and cognitive function were assessed in healthy young women. A repeated-measures design, examining caffeine intake and withdrawal symtomatology in 48 college age women used daily diaries of caffeine intake, premenstrual symptoms, and caffeine withdrawal effects, and each participant also did a daily LH surge test. In the laboratory, computerized cognitive tasks were administered in order to assess the cognitive effects of caffeine abstinence during the follicular and luteal phases of the menstrual cycle. Data analyses focused on withdrawal symptom ratings and performance on the cognitive tests one day during the follicular phase and one day during the luteal phase of the cycle after 24 hours of caffeine abstinence. Caffeine withdrawal significantly impacted self-report symptom ratings during select days of abstinence, as compared with days of non-abstinence, during both follicular and luteal phases. Withdrawal did not, however, impact cognitive performance across the menstrual cycle.
Regular caffeine users often experience symptoms of withdrawal when they are abstinent (Juliano & Griffiths, 2004; Smith, Gupta, & Gupta, 2007), and evidence indicates that withdrawal symptoms are both more frequent and more severe in women than in men (Dews et al., 1999). Moreover, the effects of this powerful stimulant and its withdrawal vary with the phase of the menstrual cycle (Terner & Wit, 2006). Since caffeine is the most widely consumed drug in the world, with tens of millions of ton ingested each year (Smith, White, & Shapiro, 2007), any effect of the drug on women is potentially both major and widespread. Taken together, these observations suggest that the symptoms associated with caffeine abstinence in regular users should vary systematically as a function of menstrual cycle phase. The purpose of the present study was to assess the presence and phase-differentiation of cognitive, motor, mood and somatic withdrawal symptoms of caffeine abstinence in regular consumers of the drug.
Despite widespread interest in the interaction of phase-related menstrual physiology with caffeine consumption and withdrawal, there is a very limited body of relevant literature. Prior studies that have examined the effects of caffeine have typically employed mixed-sex samples (Rogers & Dernoncourt, 1998; Christopher et al., 2005; Warburton et al., 2001; Rogers et al., 2005). These studies obviously do not take into account the interaction of caffeine with menstrual physiology (Stafford et al., 2007). In fact, only three studies have collected data across the menstrual cycle and addressed related caffeine effects (Rossignol et al., 1991; Pomerleau et al., 1994; Vo, Smith, & Rubinow, 2009 in press). All three investigations lacked hormonal verification of the cycle phase, and results in these studies were mixed, indicating a need for further investigation of the caffeine-menstrual cycle relationship.
Caffeine metabolism is modulated by the presence of reproductive hormones that vary with cycle phase (Fenster et al., 1999; Lucero et al., 2001; Stafford et al., 2007), hypothetically impacting both mood and psychomotor performance (Arnold et al., 1987). Briefly, the menstrual cycle consists of two primary phases. The follicular phase is characterized by the increasing levels of estradiol, while progesterone increases primarily during the luteal phase (Vo et al., 2007). It is believed that the presence of estradiol during the follicular phase, in most women, is responsible for feelings of wellbeing, while significant increases in the level of progesterone during the luteal phase are associated with negative mood states, impaired judgment, biased memory recognition (Rubinow et al., 2007), increased stress response (Farag et al., 2006), and decreased psychomotor performance (Lieberman et al., 2002).
The interaction of caffeine with reproductive hormone substrates is based on the nature of the drug and its impact on physiology. Caffeine is an adenosine antagonist that binds to adenosine receptors, reducing the inhibitory effects of neurotransmitters and increasing the excitation of the affected neural systems (Institute of Medicine, 2001). Since the central nervous system controls the production of steroid hormones, including estradiol and progesterone, the depressant action of adenosine may be influenced by increases and decreases in the production of these hormones (Kaminori et al., 1999; Mandal et al., 2007).
A recent study showed that long-term administration of caffeine in female mice decreased the production of progesterone (Mandal et al., 2007). Studies in rats have shown that the administration of progesterone suppresses the spontaneous firing of cerebral cortical neurons and prolongs the mean duration of adenosine-evoked inhibitors by about 56% (Phillis, 1985). Progesterone potentiates the action of adenosine (Phillis, 1986) by inhibiting its reuptake into neurons and glial cells. This action creates higher concentrations of adenosine in extracellular space, allowing for greater experiences of fatigue and depression associated with the luteal phase of the cycle for most women (Lucero, 2001). Because caffeine administration produces physiological reactions that reduce the effects of progesterone during the luteal phase, caffeine abstinence during this phase may then yield the opposite effect, likely impairing psychomotor performance and increasing withdrawal symptoms. It is important to note that most women experience mood symptoms associated with increasing levels of progesterone during the luteal phase, with or without caffeine. However, it is likely that caffeine abstinence in habitual caffeine consumers exacerbates these mood symptoms as a result of withdrawal effects.
Unlike progesterone, estradiol increases the firing of relatively few neurons (approximately 9% compared to 42% attributable to progesterone; Phillis et al., 1985), resulting in only a weak ability to inhibit adenosine transport (Phillis, 1991). In effect, estradiol is much like caffeine, in that it feebly antagonizes the inhibitory actions of adenosine (Fenster et al., 1999). Phillis (1991) suggests that although estradiol concentration in the extracellular fluid of the brain is unlikely to reach levels at which the activity of the adenosine transporter is compromised, it may be sufficient to antagonize the actions of endogenously released adenosine, exerting caffeine-like stimulating action on the central nervous system (CNS). Such action could, at least partially, account for the positive emotions attributed to the follicular phase of the cycle.
In sum, a moderate dose of caffeine enhances mood and psychomotor performance by inhibiting adenosine receptors (James, 1997; Smith et al., 2007), and the presence of estradiol should facilitate the stimulating effect of caffeine (Phillis et al., 1991); therefore, the interaction of the two substances should enhance psychomotor performance and alleviate withdrawal symptoms. On the other hand, progesterone inhibits the reuptake of adenosine, leading to an increased concentration of adenosine particles in the extracellular space that then manifests in more fatigue, depression, and relative sedation (Lieberman et al., 2002; Rubinow et al., 2007). As demonstrated, the action of caffeine is facilitated by the presence of estradiol and inhibited by progesterone. As such, the fluctuation of progesterone and estradiol during the menstrual cycle also implicates differential physiological processes and withdrawal manifestations as it interacts with the absence of caffeine in habitual caffeine users. The present study sought to systematically assess the effects of caffeine withdrawal across the menstrual cycle. The designed was strengthened by the use of daily symptom diaries and daily ovulation tests across two complete menstrual cycles in each participant.
Materials and Methods
Definition of Cycle Phase
Definitions of cycle phases are adopted from Allen and colleagues (1996) and are similar to those used in many other studies (e.g., Allen et al., 1996; Hellstrom-Lindberg, 2000; Soderberg et al., 2006). The first day of bleeding was defined as Day 1 of the menstrual cycle. The two cycle phases were defined as follows: the first day of the follicular phase as Day 5 of the menstrual cycle; the first day of the luteal phase as 4 days after the luteinizing hormone (LH) surge, as determined by hormonal verification using midstream ovulation test strips (see below).
The Physician Health Questionnaire (PHQ; Spitzer, Kroenke, & Williams 1999) screens for common mental disorders, functional impairment, recent stressors, menstruation, and pregnancy. The measure has reliability and validity (Kroenke et al., 2001) and is particularly useful in research settings (Spitzer et al., 1999).
The Caffeine Intake Form was used in the study to record daily caffeine consumption. It assesses the amount of consumption for various products containing caffeine. This form was included in the daily diaries.
The Self-Rating Scale for Premenstrual Tension Syndrome (PMTS-SR; Steiner et al., 1980) was used as part of the daily diary in order to check for PMS symptoms. The widely used PMTS-SR has substantial reliability and validity (Bergant et al., 2004) and was the basis for diagnosing PMS in the present study.
The Caffeine Withdrawal Questionnaire (Griffiths et al., 1990a; Subjective-Effects Questionnaire) is a checklist that asks participants to rate the extent to which they experience withdrawal symptoms. The inventory assesses a total of 26 mood (e.g. irritability, depression) and somatic (e.g. lightheadedness, upset stomach) symptoms. Studies assessing these symptoms show that 100-300-mg/day of caffeine results in a reliable caffeine withdrawal syndrome when a placebo is substituted (Griffiths et al., 1990a; 1990b; Evans & Griffiths, 1992; 1999; Schuh & Griffiths, 1997).
Hormonal Verification for Cycle Phase
The ovulation test is a midstream test that detects the “LH Surge,” the dramatic increase in luteinizing hormone that takes place before ovulation. The test has sensitivity to 20mlU/ml/lh and is 99.9% accurate in anticipating ovulation, with 99.8% specificity (www.earlypregnancytest.com).
All potential participants signed informed consent during an orientation session and underwent the intake interview. The interview consisted of the Participant Screening Form and the PHQ. Qualified volunteers were given instructional handouts and procedural demonstrations for (1) testing for ovulation using the provided kit, (2) accessing their daily diaries online, and (3) practicing computerized tasks. The latter familiarized participants with the CDR computerized tasks described above. The Cognitive Drug Research battery has standardized practice sessions that run approximately 30 minutes long.
Participants proceeded with daily entries online at the same time every morning for two complete menstrual cycles (approximately 56-80 days, depending on individual cycle length), and researchers checked each morning to see that all had provided their data. Daily dairies were completed at the same time each morning to account for possible confounding variables, such as time-of-day-effects on mood and caffeine withdrawal. Participants were called after one missed day. They were allowed one missed entry the following day, but more than one miss excluded the participant from inclusion in the data analysis unless it was demonstrated that ovulation did not take place during the missing day(s).
During one of the two menstrual cycles, participants were asked to eliminate, insofar as possible, their intake of caffeine by avoiding certain foods and drinks (e.g. coffee, tea, alcohol, coconut products, shellfish, and all foods containing chocolate products; Jones, et al., 2000) during the follicular phase, beginning at noon of one day and continuing until after completing the laboratory session the next day. During the other menstrual cycle, they restricted caffeine intake during the luteal phase. The cycles were counterbalanced as to phase of food restriction. To enhance compliance with the restriction requirement, each participant was given a handout containing a complete list of foods and drinks to avoid that also appeared on the daily diary website.
During each of the laboratory sessions, the participant’s inner cheek was swabbed, and participants were told that the saliva sample was collected in order to verify caffeine abstinence. They were given an opportunity to reschedule the laboratory session if they reported consuming any caffeinated foods during the restricted timeframe. The saliva samples were not analyzed (Aguinis, et al., 1993). Participants were also asked to complete their daily diaries online following the computerized behavioral tasks. Both CDR tasks and questionnaires were completed on a computer.
For purposes of the data analysis, scores for each relevant variable were first derived.
Mean caffeine intake was computed for each participant for each phase of the cycle (caffeine values adopted from Juliano & Griffiths, 2005). Mean caffeine intake was the average across two months of daily dairies, excluding the two days of abstinence.
Premenstrual Syndrome Symptoms
Scores for PMS were based on the PMTS-SR. In order to meet criteria for PMS and to be excluded from the final sample, participants had to endorse at least a 30% increase in symptom severity during the luteal phase, as compared to the follicular phase, for 2 consecutive months.
Caffeine Withdrawal Symptoms
Withdrawal scores based on the self-report questionnaire were obtained for each day of abstinence (one for each phase) and also for one randomly selected non-abstinence day in each phase. Ratings (on a 4-point Likert scale) of each individual symptom from the scale were summed to achieve a total withdrawal score. Withdrawal scores were examined further using frequency of individual withdrawal symptom scores during abstinence and non-abstinence days. The most common symptom was also selected for additional comparisons, with the rated withdrawal symptom during abstinence and scores during non-abstinence serving as the dependent variable.
Cognitive Drug Research Scores
The tasks included factor scores of
A series of analyses addressed phase differences in caffeine consumption, self-reported symptoms, cognitive performance, and caffeine consumption.
Caffeine Consumption. In order to ensure that our results were not affected by possible phase differences in caffeine consumption, we first
Symptomatology. Our primary goal was to determine the effects of caffeine abstinence on self-reported symptoms across the menstrual cycle. Accordingly, withdrawal effects were examined during two normal days compared with two days of abstinence during the follicular and luteal phases. The ANOVA indicated a significant effect of abstinence on mood and somatic symptoms during the follicular phase of the menstrual cycle, F(1, 47) = 6.32,
Phase-Differentiated Effects of Caffeine Abstinence: The CDR Factors
Five repeated measures ANOVAs were conducted examining each of the five factor scores from the CDR battery. Analysis suggested no significant differences between the luteal and follicular phases in
Overall, the present study did not reveal statistically significant differences for phase-differentiated effects of caffeine withdrawal on self-reported mood, somatic symptoms, or cognitive function. An additional analysis was conducted to examine possible effects of the order of laboratory visits (luteal-follicular vs. follicular-luteal). The order effects were not significant, F(1, 47) = 0.007,
The present study involved a repeated-measures design with counterbalancing that examined the effects of caffeine abstinence across the major phases of the menstrual cycle. Theoretically, the cognitive and emotional sequelae of caffeine withdrawal should vary systematically with the changing hormonal physiology that produces and accompanies the sequential phases of the menstrual cycle (Hampson & Kimura, 1988; Hampson, 1990). However, most prior caffeine abstinence studies have not attempted to take these physiological changes into account (Stafford, et al., 2007). The present study was an initial effort to address this gap in our knowledge. The primary goal of the study was, therefore, to examine phase-differentiated effects of caffeine abstinence, in habitual caffeine consumers, on objective measures of psychomotor and cognitive tasks and on self-reports of mood and somatic symptoms.
Overall Withdrawal Effects
Overall, present data show that caffeine abstinence in women who consume on average 349.38mg/day results in withdrawal affects on self-reported measures of mood and somatic variables, including, in particular, increases in caffeine craving (Griffith et al., 1990; Juliano & Griffiths, 2004), sleepiness (Hewlett & Smith, 2006; Rogers et al., 2005), and fatigue (Hewlett & Smith, 2006; Smith et al., 1999; Warburton et al., 2001).
The brain mechanisms underlying the aforementioned effects of caffeine abstinence are not entirely understood. However, the molecular target of action of caffeine is known. Caffeine antagonizes the activity of the brain neuromodulator adenosine at specific receptor subtypes (i.e.,A1 and A2A receptors). Adenosine is a nucleoside that is widely available in the cerebral cortex, and its actions are associated with a reduction in neural activity and with cerebral vasodilatation. In general, the blockage of adenosine by caffeine results in cerebrocortical activation and vasoconstriction. It is unclear which of these effects, or others, underlie which specific symptoms of withdrawal.
Such symptoms as fatigue, sleepiness, and decreased alertness have been demonstrated in previous studies (Hewlett & Smith, 2006; Rogers et al., 2005; Smith et al., 1999; Warburton et al., 2001). The underlying cause of fatigue can be extrapolated from the known mechanisms that involve the inhibitory action of caffeine on adenosine receptors. Caffeine serves to primarily inhibit adenosine receptors in the medial reticular formation (mRF). This activity of the mRF in the brain stem is gated through the reticular nucleus of the thalamus to cortical areas. One implication may be that the effects of caffeine depend on the existing level of cortical arousal, suggesting that the impact of caffeine and of its withdrawal are, in part, state-dependent (Nehlig, 1992). In fact, caffeine-induced elevations in cortical arousal manifest as increases in the frequency and amplitude of the EEG (Lorist & Snel, 1997). Since these effects depend on the combined effects of many concurrently running facilitatory and inhibitory neural processes, fatigue may follow withdrawal as a result of increasing demands on the system as a whole.
To date, research examining caffeine craving has been minimal (Griffiths et al., 1990). However, research on cravings for other drugs (e.g. cocaine) is much more extensive. That research suggests that environmental cues, such as visual, olfactory, and other sensory stimuli, are important contributors to drug use (Reuter et al., 2002; Bordnick, et al., 2004; Matto, 2005). Cues that trigger drug use have previously been associated with increases in levels of dopamine—typically associated with feelings of reward or pleasure. Studies show that the brains of individuals with drug addictions tend to have fewer of the D2 receptors through which dopamine acts to stimulate pleasurable feelings from basic activities (e.g. eating). As a result it is more difficult for these individuals to experience pleasure. In addition, environmental cues trigger memories of use that may lead to increased activation of pleasure centers in the brain, which in turn predicts the likelihood of drug-seeking behavior and compulsive behavior—such as drug usage (Lende & Smith, 2002).
Not surprisingly, withdrawal symptoms resulting from caffeine abstinence have been quite consistently demonstrated in prior studies, including those mentioned above. Caffeine withdrawal has been associated with increases in the severity of several mood and somatic symptoms, including, headache (Silverman et al., 1992; Lane, 1997), difficulty concentrating (Garrett & Griffiths 1998; Evans & Griffiths, 1999; Jones, et al., 2000), and irritability (Griffiths et al., 1990; Evans & Griffiths 1999; Silverman et al., 1992; Roger et al., 2005), in addition to those found in the present study. However, it is important to note that not all symptoms are present in all studies (Juliano & Griffiths, 2004).
Though it was hypothesized that mood and somatic symptoms that result from withdrawal may negatively impact performance on cognitive, attentional, and psychomotor tasks (Phillips-Bute & Lane, 1998; Rogers et al., 2005; Hewlett & Smith, 2006; Stafford et al., 2007), impairment on cognitive function was not detected across the menstrual cycle. One possible conclusion is that caffeine abstinence simply does not, in fact, impact cognitive task performance. However, two factors may, instead, account, at least in part, for the nonsignificant findings. First, women in the present study may not have experienced withdrawal symptoms potent enough for phase differences to be detectable, in part because they consumed relatively low doses of caffeine (M = 349.33mg/day), roughly equivalent to 3.5 average cups of coffee. Although other studies have shown that caffeine withdrawal can develop in low caffeine consumers (<300 mg/day; Evans & Griffiths, 1999; Nehlig, 1992, it is likely that the low level of consumption in this college student sample is nevertheless a factor in the failure to confirm the hypothesis. This factor may have also contributed to the relatively low severity of withdrawal symptoms experienced, thereby reducing the likelihood of detectable differences during the follicular and luteal phases.
The second factor may be the “menstrual normality” of the women in the sample. Women with PMS, PMDD, and irregular cycles were systematically screened out of this study. It may be less likely that external factors, such as caffeine consumption, will affect phase-differentiated symptomatology in women with normal menstrual cycles and little menstrual distress than might be the case in a more heterogeneous sample (Vo et al., 2007). Although no studies have directly addressed this issue with regard to caffeine, some research has shown that women with normal cycles and little or no distress are less likely to report differences between the luteal and follicular phases in mood or somatic symptoms (Backstrom et al., 1983; Lucero et al., 2001)
Although the hypothesis that caffeine withdrawal effects should vary with menstrual phase was not supported in the present study, it is quite possible that design modifications will later demonstrate that the hypothesis is viable under certain circumstances. First, studies involving the experimental manipulation of caffeine intake are needed. Such investigations would involve the laboratory administration of three or more doses of caffeine during the follicular and luteal phases of the menstrual cycle. Allowing time for the caffeine to take effect, immediate differential effects of the drug on cognitive, psychomotor, mood, and somatic variables in women of different ages, in those with normal and abnormal cycles, and in women who are low, moderate, and high habitual consumers could be investigated. Across many of these and other studies, the diary technique developed by the authors could be applied for purposes of both data collection and screening. Withdrawal effects could also be examined within the context of these designs by testing their frequencies and magnitudes in groups of women of varied age, habitual caffeine intake, and menstrual history.
A second body of research would be based on the observation that there is individual variability in both quality and intensity of withdrawal symptoms (Juliano & Griffiths, 2004; 2005; Nehlig, 1992). Perhaps, women with previous experiences of intense caffeine withdrawal effects may be more likely to show phase differentiation. A third possibility for future research would be to take into account relevant environmental variables (e.g. academic and interpersonal stressors) that could impact self-reports of mood and also performance on cognitive tasks, thereby reducing extraneous variance. Fourth, future research might seek to control or systematically take into account the use of oral contraceptives, which may have affected present results, as well as those of earlier studies. Fifth, in withdrawal studies, level of habitual consumptoin should be examined, taking into account both daily quantity of caffeine consumed and duration of use. Some withdrawal symptoms, such as headache, appear to be more likely with longer-term use, which can produce characteristic changes in blood flow velocities (Nehlig, 1992). Other possibilities would be to include biological verification of abstinence, assessment of sleep during abstinence, and the use of samples from older populations, thereby increasing generalizability.
The goal of the present study was to examine the effect of caffeine abstinence across the menstrual cycle as it impacts cognitive, psychomotor, mood, and somatic variables. This study is one of the first to examine the effects of caffeine abstinence in women in relation to the menstrual cycle. Consistent with some of the previous studies in the literature, present data provided no evidence of differential memory effects as a result of caffeine abstinence. Furthermore, there were no differences on memory tasks between the follicular and luteal phases of the menstrual cycle. However, mood and somatic withdrawal effects were detected during abstinence in women who habitually consumed 300 mg of caffeine daily. Thus, we can conclude that caffeine withdrawal does impact both affective and physical systems, but not, in the present study, cognitive responses.