Influences of Fasting on Stress Response and Withdrawal Symptoms in Habitual Khat Users

Khat is a natural stimulant plant that is available in parts of East Africa and the southern part of the Arabian Peninsula [1]. Cathinone, a phenylethylamine alkaloid, is a major constituent of khat that mediates khat’s psychostimulant properties by mainly activating dopaminergic and noradrenergic pathways, though it also activates the serotonergic pathway to a lesser extent [2-5]. Cathinone chemically resembles amphetamine [2]. Historically, khat sessions are held as a means of socialization [1]. However, khat has been increasingly used for recreational purposes. It is estimated that between 30 and 50% of Ethiopians are current khat users [6], and between 70–82% of males and 30–43% of females in Yemen use khat [7, 8]. Users typically get together in the afternoon and use khat for several hours [9]. Evidence indicates that chronic khat use is associated with myocardial infarction [10], worsening of ongoing mental illness [11], and impaired neurocognitive functions [12]. Chronic khat use is also related to increased negative affect and attenuated psychobiological responses to stress [13]. Causal directions of these relationships are yet to be determined.

Healthy adults in Yemen who follow the Islamic faith abstain from consumption of food, drinks, substances, as well as sexual intercourse from sunrise to sunset during Ramadan, the ninth month of the lunar calendar. Evidence indicates that khat use is associated with alterations in psychobiological stress responses and subjective mood [13]. There is also evidence that acute fasting is associated with reduced physiological responses to laboratory stress [14], although the impacts of fasting on stress responses and withdrawal experienced by habitual khat users or other drug users are largely unknown. Khat users report depressive and anxious symptoms as well as insomnia after termination of use [15]. These symptoms are commonly reported among users of other substances, such as tobacco [16]. Withdrawal symptoms induce stress, which is a known factor in maintenance of, and relapse to, drug use [17]. The question of whether fasting and khat use status moderate stress responses and withdrawal symptoms is important in addressing the role of stress in khat use.

Here, we report results from a study examining the effects of fasting on stress responses and mood in chronic khat users and nonusers. We used an established protocol to assess subjective and objective markers of stress. Based on previous studies [13, 14], it was predicted that religious fasting (a practice that occurs during Ramadan) would be associated with enhanced negative affect and withdrawal symptoms, as well as attenuated psychophysiological responses to acute laboratory stress among khat users relative to nonusers during the fasting periods.

This study was part of a larger study that examined biobehavioral mechanisms of long-term khat use [12, 13, 18-20]. Recruitment of participants was completed through flyers posted around Sana’a University in Yemen and surrounding communities. Inclusion criteria were as follows: no current nor recent history of major physical diseases, no psychiatric disorders, and no current use of medications. Individuals who reported khat use for the past 2 years were classified as khat users and those who reported that they had never used khat were classified as nonusers. Previous studies from the same project [12, 13, 18-20] did not include individuals who completed study sessions during the month of Ramadan. This procedure was used to eliminate potential confounding effects of fasting. We used gender and khat use status of individuals who completed sessions while fasting to match them with those who completed sessions while not fasting. As a result, this study included 40 individuals (12 females and 28 males; 25 khat users and 15 nonusers) in each fasting group. The study was approved by the Institutional Review Board of Sana’a University. All participants received monetary incentives for their participation in the study.

The Subjective State Measure (SSS; [21]) is an 18-item self-report questionnaire that assesses mood states. An Arabic version was developed and adapted in our previous studies [13]. Information about these measures and their use with similar populations has been published elsewhere [13, 20, 22]. Due to some commonalities in symptomatology between khat withdrawal [15] and tobacco withdrawal [16], we measured the severity of khat withdrawal using items that assessed irritability, anger, anxiety, difficulty concentration, restlessness, sadness, and hunger. These items were adapted from the Minnesota Nicotine Withdrawal Scale [16, 23]. Systolic blood pressure (SBP), diastolic blood pressure (DBP), and heart rate (HR) were collected by a MicroLife Automatic Blood Pressure Monitor BP 3AC1. The oscillometric device has been validated by the European Society of Hypertension’s International protocol. Measurement of expired carbon monoxide was completed using MicroCO monitors (Micro Direct, Auburn, ME, USA).

Participants were asked to complete an onsite, semi-structured screening interview for assessment of eligibility, which included questions regarding medical and substance use histories. Qualified participants were asked to schedule 1 ambulatory and 1 laboratory session. A set of guidelines and instructions about diet and sleep was provided prior to those sessions. Each participant was tested individually. On the ambulatory session day, participants reported to the laboratory between 10 a.m. and noon to receive a set of mood questionnaires (e.g., SSS). The participants were instructed to complete the SSS 5 times, at: 10:00 a.m.; 7:00 p.m.; 10:00 p.m., or before going to bed; 8:00 a.m. the next day, immediately after waking up; and 9:00 a.m. This timing was kept consistent across all groups of participants. Participants returned to the laboratory the next morning. All laboratory stress sessions began between 9 a.m. and 10 a.m. Upon arrival, the participants sat in a comfortable chair. After a blood pressure cuff was attached, the participants rested for 30 min. This baseline period was followed by exposure to an acute stress challenge (10 min). The participants were asked to subtract the number 7 or 13 from a four-digit number. When the participants made a mistake, they were asked to go back to the previous correct response. This task has been used in stress reactivity studies [24]. Then, participants rested for another 30 min. Blood pressure and HR were obtained every 5 min during baseline and post-stress recovery, and they were obtained every 3 min during the stress task. SSS was administered at the end of each period.

Subjective mood data from the ambulatory sessions (SSS positive affect and negative affect) were analyzed using a 2 fasting group (fasting and no-fasting) × 2 khat use group (current user and nonuser) × 5 sampling times (10 a.m., 7 p.m., 10 p.m. or bedtime, 8 a.m. or wake-up, and 9 a.m.) multivariate analysis of variance (MANOVA) with Wilks’ lambda test. Withdrawal measures were analyzed in khat users only using a 2 fasting group × 5 sampling times MANOVA with Wilks’ lambda test. For data analysis of laboratory data, cardiovascular measures were averaged in each period. A series of 2 fasting group × 2 khat use group × 3 times (baseline, stress, and recovery) MANOVAs with Wilks’ lambda tests were conducted on cardiovascular (SBP, DBP, and HR) and mood measures. Khat withdrawal symptoms were analyzed in khat users only. If a significant interaction was found, follow-up tests were conducted using the Bonferroni correction. Slight variations in reported degrees of freedom existed due to missing data.

The mean age of the sample was 22.9 years (SD = 5.6), and the mean BMI was 20.6 (SD = 4.0). Reported average hours of nighttime sleep over the past week was 7.2 (SD = 1.8). There were no fasting nor khat use group differences in these measures (see Table 1). On average, khat users started chewing when they were 15.5 years old (SD = 4.3), and they used khat 5.3 days a week (SD = 2.1) for 6.7 years (SD = 5.3). There were no fasting group differences in khat exposure. Thirty percent of khat users (24/80) reported smoking. Smokers were equally distributed in 2 fasting groups (p = nonsignificant; see Table 1); they smoked 14.2 cigarettes per day (SD = 5.3). Smokers in the fasting group smoked for more years than those in the no-fasting group (F[1, 18] = 5.07, p = 0.04). Expired carbon monoxide was lower in smokers who were in the fasting group than in smokers in the no-fasting group (F[1, 22] = 9.08, p = 0.006), which provided support for abstinence from tobacco use in the fasting group.

Khat users had lower positive affect than nonusers, as found by a main effect of the khat use group (F[1, 70] = 9.65, p = 0.003). This main effect was further qualified by a khat use group × time interaction (F[4, 67] = 2.50, p = 0.05). Follow-up MANOVAs conducted in each khat use group (adjusted p value = 0.05/2 = 0.025) indicated that levels at 10 p.m./bedtime were lower than those at 10 a.m. of the same day in the khat users (p < 0.04); however, this was not found in nonusers (see Fig. 1a). Also, a significant group main effect in negative affect indicated greater levels in khat users than in nonusers (F[1, 70] = 12.8, p < 0.001). Negative affect at 8 a.m./waking was lower than that at 10 a.m. the day prior (time main effect: F[4, 67] = 2.77, p = 0.03; see Fig. 1b). There was a fasting group × time interaction in withdrawal symptoms (F[4, 42] = 4.48, p = 0.004). A follow-up MANOVA revealed greater symptoms during the evening hours (7 and 10 p.m.) than during the morning hours the next day (9 a.m.) in the no-fasting group (p ≤ 0.005); however, this time effect was not found in the fasting group (see Fig. 1).

Khat users had lower positive affect than nonusers (khat use group main effect: F[1, 70] = 4.57, p = 0.04; see Table 2). A significant main effect of time in positive affect (F[2, 69] = 5.46, p = 0.006) was qualified by a fasting group × time interaction (F[2, 69] = 3.99, p = 0.02). This interaction was due to a stress-related decrease in the no-fasting group (post hoc: p = 0.002); positive affect remained low and flattened in the fasting group (Fig. 2a). With respect to negative affect, khat users had greater negative affect than nonusers (F[1, 70] = 9.88, p = 0.002). A fasting group × time interaction (F[2, 69] = 3.07, p = 0.05) found a stress-related increase in negative affect in the no-fasting group (post hoc: p = 0.009) that was not found in the fasting group (see Fig. 2b). Withdrawal symptoms increased in response to stress in all groups (time effect: F[2, 44] = 3.29, p = 0.047), but this change was not associated with fasting nor khat use.

There were significant main effects of time in blood pressure (SBP: F[2, 70] = 16.6, p < 0.001; DBP: F[2, 70] = 10.5, p < 0.001) that were further qualified by khat use group × time interactions (SBP: F[2, 70] = 5.37, p = 0.007; DBP: F[2, 70] = 9.20, p < 0.001; see Table 2). These effects indicated stress-related increases in nonusers (SBP: p < 0.001; DBP: p = 0.001) but not in khat users (see Fig. 3a, b). For HR, a significant time effect (F[2, 65] = 7.91, p = 0.001) indicated a stress-related increase in all groups (p = 0.03; see Fig. 3c). There were no differences related to fasting in cardiovascular measures.

This study shows that fasting and habitual khat use independently influence subjective and physiological indices related to stress. Khat users in a no-fasting group reported increased withdrawal symptoms during the evening hours compared to those in the fasting group. The group difference peaked in the late evening and diminished the next morning. This pattern is consistent with previous research that examined withdrawal symptomatology of khat use. Khat users typically get together with peers in the early afternoon and start chewing khat [9], and they usually smoke tobacco during this period [9, 25]. Users report negative affect and insomnia after khat sessions [15]. The current finding of increased withdrawal symptoms during the evening hours among khat users who were not fasting may reflect elevation in post-use symptoms. In contrast, during Ramadan, adults abstain from eating, drinking, and substance use from sunrise to sunset. It is possible that lower withdrawal symptoms in the fasting group reflected full abstinence from substance use, use of substance(s) after sunset, or a mixture of these conditions. Asking about substance use at each time point might improve results; however, we note the possibility that people may not disclose drug use during Ramadan due to social desirability.

Fasting and no-fasting groups showed different response profiles to acute stress. The no-fasting group reported increased negative affect and decreased positive affect in response to stress. These results are consistent with patterns observed in healthy adults without diet restrictions [24]. The causes of the flattened responses in the fasting group are not known at this time, although several factors might have contributed to this observation. First, one purpose of Ramadan is supposedly to enhance awareness of people who are poor and not fortunate through prayer, charity, and fasting. Fasting is also considered important to gain patience and self-control. It is possible that Ramadan modifies psychological processes in response to stress. The other possibility is that fasting changes metabolic functions [26] and leads to changes in central and peripheral systems associated with stress and coping. Fasting did not influence cardiovascular responses to stress; exposure to stress resulted in the expected changes. Future research should elucidate associations between fasting during Ramadan and stress mechanisms through multilevel measures, including stress and appetite hormones.

Chronic khat use was associated with lower positive affect and enhanced negative affect during ambulatory and laboratory sessions, yet khat users exhibited attenuated cardiovascular responses to stress. These findings are consistent with those of prior studies among khat users [13], tobacco users [27], and concurrent khat and tobacco users [20]. Animal studies have also shown that chronic administration of cathinone, one of the main constituents of khat, is related to dose- and time-dependent reductions in serum cortisol. There is a growing body of literature suggesting that fasting induces functional changes in central nervous and endocrine systems [28], both of which play essential roles in stress regulation. To date, there are only 2 human studies that examined the effects of short-term fasting on stress indices, such as cortisol, in controlled environments. One study showed that low levels of glucose after fasting (≥8 h) were associated with attenuated cortisol stress response [14]. Another pilot investigation indicated reduced basal adrenocortical activity among fasting women during the luteal menstrual phase [29]. Results of this study extend existing literature by confirming blunted stress responses among khat users, regardless of fasting. The finding that fasting did not moderate the link between khat use status and stress adds support to the robustness of khat-related dysregulation of the stress response.

Due to a small sample size, results of this study should be considered preliminary. Also, our findings are limited by our use of a cross-sectional approach. A within-subject design (each person completes 1 session while fasting and 1 session while not fasting) may improve results. This study lacked objective measures of fasting and khat use. We did not know whether participants in the fasting group were truly abstinent from food and khat use. However, low CO in the fasting group lends some support for drug abstinence in this group. There is no gold standard questionnaire to assess khat withdrawal. While self-reported symptoms after khat use are similar to those after smoking or drinking, research examining the psychobiological impacts of khat withdrawal is needed. In this study, 1 research staff member administered a mental arithmetic task to measure psychophysiological responses to stress. This approach was based on local vetting, refinement of the protocol, and available human resources. However, future research should consider using a combination of mental math and other challenging tasks, such as public speaking or anger recall interviews, to reliably assess psychobiological stress responses. Biomarkers of stress, including adrenocorticotropic hormone, cortisol, and HR variability, should be included in future research. It should be noted that we did collect saliva samples for the measurement of stress responses; however, we encountered some inconsistencies in labeling samples, and we were not able to carefully analyze them. Nevertheless, the use of a larger sample with sufficient statistical power, multiple stress tasks, and stress measures should improve current results. Strengths of this study include our use of established protocol to assess stress and our use of a matched sample to mitigate the effects of confounding factors on observed findings.

In conclusion, this study is the first to report different patterns of stress responses and withdrawal symptoms between khat users who were fasting and those who were not fasting during laboratory and ambulatory periods. The study also found flattened emotional responses to acute stress in fasting individuals. Results show complex relationships among fasting, khat use, and stress mechanisms. An emerging literature suggests potential protective effects of fasting against chronic diseases [30]. Findings from this study call for the need of researchers to examine the potential effectiveness of fasting, not necessarily in the context of religious fasting, in reducing stimulant use.

We thank the following for their help with this project: Najat Saem Khalil, Molham Al Habori for helping with the data collection; Amal Alanisi, Tawfeek Alharazi, Ashrak alAwdari, Bakeer Dahmash, and Essa Oumairi for assisting with data collection; and Abdul Kareem Nasher for help with data entry.

All study procedures were reviewed and approved by the Institutional Review Board of the University of Minnesota, and the study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.

The authors of this manuscript do not have any conflict of interested related to this work.

M.A. designed the study and developed the protocol. M.N. conducted literature review on relevant studies and data management and analysis. M.N. and M.A. interpreted results of the analysis. Both authors contributed to the development of the manuscript.

This research was supported by a FIRCA grant from the National Institutes of Health/Fogarty International Center (R03TW007219), an R21 National Institute for Drug Abuse grant (DA024626), and a grant from the Office of International Programs at the University of Minnesota. These grants had no role in the study design, collection, analysis or interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.