Pharmacological risks of khat–oral antidiabetic drug interactions among patients at Gondar university referral hospital

Abstract

Background

Khat (Catha edulis), a psychoactive plant commonly chewed in Ethiopia, is known to influence drug metabolism through its active compound, cathinone. Among patients with T2DM, concurrent khat chewing and OAD use may result in pharmacokinetic and pharmacodynamic interactions, particularly in polypharmacy contexts. Despite these concerns, the clinical relevance of khat–OAD interactions remains poorly understood in high-prevalence, low-resource settings.

Methods

This study assessed the prevalence, predictors, and perceptions of potential khat–OAD interactions among T2DM patients. A convergent parallel mixed-methods design was employed from July 1 to December 30, 2024, at Gondar University Hospital, Ethiopia. A total of 422 adult T2DM patients on OAD therapy who reported khat use were systematically sampled. Quantitative data on demographics, clinical profiles, medication use, and khat chewing behaviors were collected through structured interviews and verified via medical records. Potential interactions were identified using drug interaction databases, and logistic regression was used to determine independent predictors. Additionally, in-depth interviews with 20 patients and healthcare providers explored awareness, perceptions, and clinical experiences related to khat use and diabetes care.

Results

Among the 422 participants, 63.3% (n = 267) had at least one potential khat–OAD interaction, and 23.2% (n = 98) experienced a major interaction. The most frequently implicated drugs were glibenclamide and sitagliptin. Significant predictors of interaction included female sex (AOR = 1.38; 95% CI: 1.00–1.89), polypharmacy with ≥ 7 medications (AOR = 2.43; 95% CI: 1.61–3.67), daily khat use (AOR = 2.20; 95% CI: 1.52–3.18), and khat use for ≥ 13 years (AOR = 1.09; 95% CI: 1.04–1.15). Qualitative findings highlighted the widespread cultural normalization of khat chewing, low patient awareness of potential interactions, and gaps in provider counseling.

Conclusions

Harmful khat–OAD interactions are common among T2DM patients in Northwest Ethiopia, primarily driven by behavioral and treatment-related factors, highlighting the need for culturally sensitive pharmacovigilance and patient education.

Clinical trial number

Not applicable.

Introduction

Type 2 diabetes mellitus (T2DM) poses a significant and escalating public health challenge in Ethiopia, driven by factors such as urbanization, sedentary lifestyles, and dietary transitions¹. Recent estimates indicate that the national prevalence of T2DM in Ethiopia ranges from 3.2% to 5.0%, with higher rates reported in urban areas. In Gondar, the burden is particularly notable due to urbanization and limited lifestyle interventions².

Effective management of T2DM typically involves oral antidiabetic drugs (OADs), including metformin, sulfonylureas, and Dipeptidyl Peptidase-4 Inhibitors (DPP-4 Is). However, the high prevalence of polypharmacy among diabetic patients—often necessitated by coexisting chronic conditions—raises concerns about potential drug–drug and drug–herb interactions, which may compromise treatment outcomes and patient safety^3,4. Polypharmacy, commonly defined as the concurrent use of five or more medications, is especially prevalent among patients managing comorbidities such as hypertension, dyslipidemia, or cardiovascular disease, further increasing the likelihood of such interactions⁵.

In Northwest Ethiopia, particularly in Gondar and surrounding areas, the habitual chewing of khat (Catha edulis) is culturally ingrained and remains widespread⁶. Khat contains cathinone, an amphetamine-like alkaloid with sympathomimetic effects that can influence the pharmacokinetics and pharmacodynamics of several drugs, including those used in diabetes management^7,8. Cathinone may stimulate hepatic enzymes such as CYP2D6 and CYP3A4, potentially altering the metabolism of sulfonylureas or DPP-4 inhibitors⁹. Furthermore, its appetite-suppressant and glucose-elevating effects may counteract the glycemic-lowering effects of OADs. Several experimental and observational studies suggest that khat may exacerbate glycemic variability by potentiating hypoglycemia or reducing the efficacy of OADs through mechanisms such as altered drug absorption or hepatic metabolism^10,11.

Despite these concerns, research investigating the potential interactions between khat and antidiabetic medications—particularly within polypharmacy contexts—remains scarce. Most existing literature focuses on the general physiological and neuropsychiatric effects of khat, with limited empirical data examining its implications for diabetic pharmacotherapy^12,13. his research gap is particularly concerning in high-prevalence settings such as Gondar, where T2DM, polypharmacy, and khat use converge, increasing the risk of clinically significant drug–herb interactions. Globally, drug–herb interactions are increasingly recognized as a challenge to chronic disease management, especially in regions where traditional medicine use is common, underscoring the need for context-specific evidence to inform pharmacovigilance and clinical practice¹⁴.

To the best of our knowledge, this is the first study in Ethiopia to systematically investigate the clinical and pharmacological implications of khat–OAD interactions among polypharmacy patients. Using a convergent mixed-methods approach, this study integrates clinical data with patient and provider perspectives, offering a multidimensional understanding of khat’s impact on diabetes care. The findings are expected to inform safer prescribing practices, improve patient education, and support the development of culturally sensitive treatment guidelines that accommodate prevalent traditional practices such as khat chewing.

This study is particularly significant given the limited awareness among patients and healthcare providers regarding the metabolic and clinical risks of khat–OAD interactions. Such interactions can result in therapeutic failure, glycemic instability, or an increased risk of hypoglycemia—issues with profound public health implications in low-resource settings like Ethiopia.

A conceptual framework was developed to guide the study, summarizing the hypothesized pathways through which khat may interact with OADs—via enzyme induction, appetite suppression, and behavioral non-adherence (Fig. 1). The figure was carefully designed to avoid redundancies and clearly depict the mechanisms involved.

Fig. 1.

Hypothesized pathways of khat–induced interaction with OADs.

Therefore, the aim of this study was to assess the prevalence, clinical significance, and predictors of potential khat–OAD interactions among polypharmacy patients with T2DM at Gondar University Hospital. Additionally, the study sought to explore patient and provider perspectives on khat use and its perceived impact on diabetes management.

Methods

Study design and setting

A cross-sectional convergent parallel mixed-methods study was conducted from July 1, 2024, to December 30, 2024, at Gondar University Hospital, a major tertiary referral and teaching hospital in Northwest Ethiopia. In this design, both quantitative and qualitative data were collected and analyzed simultaneously, with findings integrated during interpretation to provide a comprehensive understanding of the research problem.

Gondar University Hospital serves a large catchment population and offers extensive chronic care services, including diabetes management and pharmacotherapy follow-ups. The hospital primarily serves patients from the public sector but also manages referrals from private health facilities requiring specialized or advanced care. This setting was purposefully selected due to the increasing burden of T2DM, the widespread practice of polypharmacy, and the high prevalence of khat chewing in the region, all of which may contribute to potential drug interaction risks.

The mixed-methods approach allowed triangulation of statistical associations with patient narratives. Integration of findings was performed using a joint display matrix that aligned qualitative themes with quantitative predictors of interaction risk, enhancing interpretability and contextual depth.

Ethical approval was obtained from the University of Gondar, School of Pharmacy. All methods were performed in accordance with the relevant guidelines and regulations.

Study population and sampling

The study population included adult patients (aged 18 years and above) diagnosed with T2DM who were on OAD therapy and concurrently taking three or more medications. Inclusion criteria required patients to report khat chewing at least once per week during the past month and to provide informed consent. Exclusion criteria were insulin-only therapy, severe illness requiring hospitalization, and cognitive or communication impairments.

Sample size calculation

The required sample size was calculated using Cochran’s formula:

Where:

• n = initial sample size.

• Z = Z-score for 95% confidence level (Z = 1.96).

• p = estimated proportion of the population (set at 0.5 to maximize sample size).

• d = desired precision (margin of error = 5% = 0.05).

Substituting values:

Accounting for a 10% non-response rate:

Thus, the final sample size was 422.

Sampling technique

Participants were selected using a systematic random sampling technique from the diabetic clinic registry. Approximately 2,000 eligible diabetic patients were expected to visit the clinic during the study period. The sampling interval (k) was calculated as:

.

Thus, every fifth eligible patient on the registry list was selected after a random starting point between 1 and 5 was chosen to minimize selection bias. Sampling was proportionately distributed across the six-month data collection period to account for seasonal or temporal variations in khat use and clinic attendance.

Data collection procedures

Quantitative component

The questionnaire was initially developed in English and translated into Amharic, followed by back-translation to ensure linguistic and conceptual equivalence. Items were adapted from previously validated instruments used in studies on polypharmacy, khat use, and drug–herb interactions, with modifications made to suit the local sociocultural and clinical context^3,15–18. The internal consistency of the Amharic version was evaluated using Cronbach’s alpha (α = 0.81), confirming reliability.

Data were collected through structured face-to-face interviews conducted by trained clinical pharmacists using a pre-tested, standardized questionnaire. The tool gathered information on sociodemographic characteristics (age, sex, education, occupation), medical history (duration of diabetes, presence of comorbidities), and a comprehensive medication profile. It also included questions on khat chewing behaviors, such as frequency, duration, and quantity of use, as well as patient knowledge and perceptions regarding potential drug–khat interactions.

For the quantity of khat consumed, chewing amounts were categorized as small (≤ 1 bundle per session), medium (2 bundles per session), and large (≥ 3 bundles per session)^19,20. Medication histories were cross-validated through review of medical records, prescriptions, and medication containers whenever available.

Qualitative component

A subset of participants (n = 20), comprising 10 patients and key informants—4 physicians, 3 pharmacists, and 3 nurses—were purposively selected for in-depth interviews (IDIs) to provide deeper insights into the clinical and behavioral aspects of khat use among patients on OAD therapy. Selection was guided by participants’ experience with diabetes care or personal experience with khat use, and recruitment continued until data saturation was reached, defined as the point at which no new themes or insights emerged during consecutive interviews.

An interview guide was employed to explore participants’ awareness, beliefs, and personal or clinical experiences related to khat–drug interactions, as well as the challenges of managing diabetic patients who regularly chew khat^21–24. The guide was developed based on themes identified in previous qualitative studies on substance use and chronic disease management and was adapted to reflect the local context with a specific focus on khat–OAD interactions.

Each interview lasted between 30 and 60 min and was conducted in a quiet, private room within the hospital to ensure confidentiality and participant comfort. Interviews were facilitated by two trained qualitative researchers with backgrounds in clinical pharmacy and public health, both fluent in the local language and experienced in qualitative interviewing.

All interviews were audio-recorded, transcribed verbatim, and translated into English where necessary. The qualitative component was designed to complement and contextualize the quantitative findings by providing deeper insights into the behavioral and clinical perspectives influencing drug–khat interaction risks.

Assessment of Khat–OAD interactions

Each participant’s medication regimen was systematically reviewed to identify potential interactions between khat and OADs, with particular attention to commonly prescribed agents such as metformin, sulfonylureas (e.g., glibenclamide, glipizide), and DPP-4 inhibitors (e.g., sitagliptin). The assessment was conducted using internationally recognized drug interaction databases—Lexicomp^®, Micromedex^®, and Drug Bank—in combination with a comprehensive review of peer-reviewed literature^25,26, with a particular focus on studies examining the pharmacological effects of khat’s active constituent, cathinone.

All identified interactions were classified according to three dimensions:

• Mechanism: pharmacokinetic (e.g., altered absorption or metabolism) or pharmacodynamic (e.g., additive effects on blood glucose).

• Severity: minor, moderate, or major.

• Clinical significance: including risks such as hypoglycemia, reduced therapeutic efficacy, or increased toxicity^27–29.

To enhance reliability, all potential interactions were independently evaluated by two clinical pharmacists, each with more than five years of experience in chronic disease pharmacotherapy and drug interaction analysis. Any discrepancies were resolved through consensus discussions, ensuring a rigorous and consistent classification process.

Data management and analysis

Quantitative analysis

All quantitative data were double-entered into EpiData version 3.1 for accuracy and subsequently exported to SPSS version 26 for analysis. Descriptive statistics, including means, standard deviations, frequencies, and percentages, were used to summarize participant characteristics and the prevalence of potential khat–OAD interactions.

Binary logistic regression was applied to identify independent predictors of clinically significant interactions, adjusting for key covariates such as age, sex, number of medications, frequency and duration of khat use, duration of diabetes, and type of OADs prescribed. Model adequacy was evaluated using the Hosmer–Lemeshow goodness-of-fit test (p = 0.62), which indicated good model fit. Multicollinearity was assessed using variance inflation factors (VIFs), which ranged from 1.12 to 2.03, confirming the absence of significant multicollinearity. A p-value < 0.05 was considered statistically significant.

Qualitative analysis

A thematic analysis approach was employed to analyze the qualitative data. A hybrid coding strategy was used, beginning with open coding to capture initial concepts directly from transcripts, followed by axial coding to organize related codes into broader categories and overarching themes. Two independent researchers coded the data using NVivo version 12, with discrepancies resolved through discussion and consensus to ensure consistency and reliability.

Themes were identified inductively, allowing new insights to emerge directly from participant narratives. To enhance the trustworthiness and credibility of the findings:

• Triangulation was applied by comparing responses across patients and healthcare providers (physicians, pharmacists, and nurses).

• Member checking was conducted with a subset of participants to validate the accuracy of interpretations.

• Peer debriefing sessions with qualitative research experts were held to refine thematic categories and ensure methodological rigor.

Results

Sociodemographic characteristics of participants

A total of 422 patients participated in the study. The mean age of participants was 52.4 years (± 11.3 SD), indicating that most individuals were middle-aged or older. When categorized into age groups, 11.7% (n = 49) were young adults aged 18–39 years. The majority, 58.3% (n = 246), were middle-aged adults aged 40–59 years, while 25.1% (n = 106) were aged 60 years and above, representing the older adult group.

In terms of sex distribution, 58.1% (n = 245) were male, and 41.9% (n = 177) were female.

Regarding marital status, the largest proportion of participants were married (63.3%, n = 267), followed by single individuals (22.7%, n = 96). Smaller percentages were divorced (8.1%, n = 34) or widowed (5.9%, n = 25).

The educational background of respondents varied: 25.1% (n = 106) had no formal education, 31.8% (n = 134) had completed primary education, 25.8% (n = 109) had attained secondary education, and 17.3% (n = 73) had tertiary-level education.

In terms of occupational status, the sample included farmers (28.4%, n = 120), government employees (21.8%, n = 92), merchants (20.9%, n = 88), and those who were unemployed or retired (28.9%, n = 122).

Finally, the majority of participants were urban residents (63.5%, n = 268), while the remaining 36.5% (n = 154) lived in rural areas (Table 1).

Table 1.

Sociodemographic characteristics of participants (n = 422).

Variable		Frequency (%) / Mean ± SD
Age (years)		52.4 ± 11.3
Age category	18–39	49 (11.7%)
	40–59	246 (58.3%)
	≥ 60	106 (25.1%)
Sex	Male	245 (58.1%)
	Female	177 (41.9%)
Marital Status	Single	96 (22.7%)
	Married	267(63.3%)
	Divorced	34 (8.1%)
	Widowed	25 (5.9%)
Education Level	No formal education	106 (25.1%)
	Primary	134 (31.8%)
	Secondary	109 (25.8%)
	Tertiary	73 (17.3%)
Occupation	Farmer	120 (28.4%)
	Government employee	92 (21.8%)
	Merchant	88 (20.9%)
	Unemployed/Retired	122 (28.9%)
Residence	Urban	268 (63.5%)
	Rural	154 (36.5%)

Medical history and medication use

The clinical profile of participants revealed diverse patterns in Type 2 Diabetes Mellitus duration, medication use, and comorbidity status. A total of 17.5% (n = 74) of participants had been living with T2DM for three years or less, indicating a relatively recent diagnosis. The majority, 58.8% (n = 248), had a disease duration of four to nine years, suggesting an established stage of T2DM. Meanwhile, 23.7% (n = 100) had a long-standing diagnosis of ten years or more.

Comorbid conditions were highly prevalent in the study population, with 71.6% (n = 302) of participants reporting at least one additional chronic illness. This highlights the complexity of managing T2DM in the context of multimorbidity and underscores the potential for drug–drug and drug–herb interactions.

In terms of OAD medication burden, 30.8% (n = 130) of participants were taking three to four medications, 39.1% (n = 165) were on five to six medications, and 30.1% (n = 127) were taking seven or more medications, indicating a significant level of polypharmacy. Additionally, 17.5% (n = 74) of participants were on concurrent insulin therapy alongside OADs.

Regarding the types of OADs used, the majority of participants (64.2%, n = 271) were on combination therapy that included metformin with either sulfonylureas or DPP-4 inhibitors. Monotherapy with metformin was reported in 10.7% (n = 45) of participants, while 16.4% (n = 69) were using sulfonylurea-based regimens. A smaller proportion, 8.8% (n = 37), were prescribed sitagliptin-based therapy, either alone or in combination with metformin. This distribution reflects current prescribing trends in resource-limited settings, where metformin remains the cornerstone of diabetes management.

When exploring medication access, the hospital pharmacy was the primary source for most participants (65.9%, n = 278). Private pharmacies served 24.6% (n = 104) of participants, while the remaining 9.5% (n = 40) obtained medications from other sources, including community health centers or informal vendors (Table 2).

Table 2.

Medical history and medication use (n = 422).

Variable			Frequency (%)
Duration of diabetes (years)		≤ 3 years	74 (17.5%)
		4–9 years	248 (58.8%)
		≥ 10 years	100 (23.7%)
Presence of comorbidities			302 (71.6)
Number of OAD medications used	3–4 medications		130 (30.8%)
	5–6 medications		165 (39.1%)
	≥ 7 medications		127 (30.1%)
OAD Type	Metformin only		45 (10.7%)
	Sulfonylurea-based (Glibenclamide, etc.)		69 (16.4%)
	Sitagliptin-based (± Metformin)		37 (8.8%)
	Metformin + Sulfonylurea/DPP-4		271 (64.2%)
Concurrent insulin use			74 (17.5)
Source of medications		Hospital pharmacy	278 (65.9)
		Private pharmacy	104 (24.6)
		Other	40 (9.5)

Khat chewing behaviors

The frequency of khat chewing varied among respondents: 44.1% (n = 186) chewed 1–2 times per week, 36.5% (n = 154) reported chewing 3–4 times per week, while 19.4% (n = 82) engaged in daily khat use.

The duration of each chewing session also differed, with the majority (52.8%, n = 223) reporting sessions lasting between 2 and 4 h. About 24.6% (n = 104) reported sessions exceeding 4 h, and 22.5% (n = 95) indicated shorter sessions of less than 2 h. Among those chewing khat for more than 4 h daily, 62.5% reported episodes of dizziness, fatigue, or irregular blood glucose readings, potentially indicative of glycemic disruption.

The average duration of khat use among study participants was 8.9 years (± 5.7 SD), indicating a long-standing habit in this population. When categorized by duration, 14.9% (n = 63) of participants had used khat for three years or less, representing short-term users. The majority, 51.1% (n = 216), reported using khat for four to twelve years, falling within the moderate-term use category. Meanwhile, 23.5% (n = 99) had a history of khat use extending thirteen years or more, reflecting long-term habitual use.

Regarding the amount of khat consumed per session, nearly half (49.5%, n = 200) described their intake as medium, 30.1% (n = 127) reported consuming a large quantity, and 20.4% (n = 86) used a small amount per session.

When exploring the timing of medication intake relative to khat chewing, participants were almost evenly split: 38.4% (n = 162) took their medications before chewing, 38.2% (n = 161) took them after, and 23.5% (n = 99) took their medications during khat chewing sessions.

Importantly, 39.8% (n = 168) of respondents reported experiencing side effects following khat use, raising concerns about potential adverse interactions between khat and medications, particularly oral antidiabetic drugs (Table 3).

Table 3.

Khat chewing behaviors (n = 422).

Behavior			Frequency (%)
Frequency of chewing		Daily	82 (19.4)
		3–4 times/week	154 (36.5)
		1–2 times/week	186 (44.1)
Duration of chewing session		< 2 h	95 (22.5)
		2–4 h	223 (52.8)
		> 4 h	104 (24.6)
Years of khat use (Years)	≤ 3 years		63 (14.9)
	4–12 years		216 (51.1)
	≥ 13 years		99 (23.5)
Quantity per session		Small	86 (20.4)
		Medium	200 (49.5)
		Large	127 (30.1)
Medication timing relative to chewing		Before	162 (38.4)
		During	99 (23.5)
		After	161(38.2)
Reported side effects post-chewing			168 (39.8)

Awareness and perceptions of Khat–Drug interactions

Participant awareness of the potential interactions between khat and medications was generally low. Only 28.7% (n = 121) of respondents reported that they had heard that khat may interact with medications, while the vast majority (71.3%, n = 301) had never heard of such interactions.

When asked specifically whether they believed khat could affect the effectiveness of diabetes medications, responses were mixed. About 33.6% (n = 142) believed that khat use does influence diabetes drug efficacy, while 41.7% (n = 176) did not believe there was an effect. An additional 24.6% (n = 104) were unsure.

Only 26.8% (n = 113) of participants had ever received counseling from a healthcare provider regarding the interaction between khat and medications. Among those who had been counseled, the majority cited physicians (54.9%, n = 62) as the source, followed by pharmacists (28.3%, n = 32) and nurses (16.8%, n = 19). These findings echoed qualitative insights in which both patients and providers reported confusion or underestimation of khat’s pharmacological impact, despite observable clinical signs.

Perceptions regarding khat’s influence on blood glucose control were similarly divided. About 38.6% (n = 163) believed khat affects blood sugar levels, while 42.4% (n = 179) believed it did not. The remaining 18.9% (n = 80) were uncertain about the effect.

Encouragingly, most participants (63.3%, n = 267) expressed a willingness to reduce or stop khat use if advised by a healthcare provider. However, 13.7% (n = 58) were unwilling to change their behavior, and 23.0% (n = 97) remained undecided (Table 4).

Table 4.

Awareness and perceptions on Khat–Drug interactions (n = 422).

Question		Yes (n, %)	No (n, %)	Not sure (n, %)
Heard khat may interact with medications		121 (28.7%)	301 (71.3%)	-
Believe khat affects diabetes drug effectiveness		142 (33.6%)	176 (41.7%)	104 (24.6%)
Ever counseled by the provider about khat and medication		113 (26.8%)	309 (73.2%)	-
Provider type (if counseled):	Doctor	62 (54.9%)	-
	Pharmacist	32 (28.3%)	-
	Nurse	19 (16.8%)	-
Believe khat affects blood sugar levels		163 (38.6%)	179 (42.4%)	80 (18.9%)
Willing to reduce/stop khat use if medically advised		267 (63.3%)	58 (13.7%)	97 (23.0%)

Medication interaction profile

To assess potential drug interactions with khat, each participant’s current medication list was reviewed and verified through multiple sources, including medical records (49.1%, n = 207), prescriptions (36.5%, n = 154), and medication containers (14.5%, n = 61). This triangulation approach ensured a more accurate assessment of prescribed therapies.

Each participant’s regimen was then systematically evaluated for possible interactions between khat and OADs, with particular focus on commonly prescribed agents such as metformin, sulfonylureas (e.g., glibenclamide, glipizide), and DPP-4 inhibitors (e.g., sitagliptin). Potential interactions were identified using internationally recognized drug interaction resources (Lexicomp^®, Micromedex^®, and Drug Bank), supplemented by a review of peer-reviewed literature. Based on these sources, interactions were categorized according to their mechanisms. Pharmacokinetic interactions (alterations in absorption, metabolism, or elimination) were identified in 42.2% (n = 178) of participants. Pharmacodynamic interactions (synergistic or antagonistic effects at receptor or physiological levels) were observed in 31.8% (n = 134). Mixed interactions (both pharmacokinetic and pharmacodynamic) were noted in 10.7% (n = 45). Overall, 63.3% (n = 267) of participants were taking at least one medication with a known or potential khat–OAD interaction.

To explore clinical relevance, patient-reported symptoms and medical record reviews were examined for evidence of clinically observed interactions (e.g., unexplained hypoglycemia, poor glycemic control despite adherence, or adverse gastrointestinal symptoms temporally associated with khat use). While no laboratory-confirmed drug–khat interactions were documented.

In terms of severity, moderate interactions were the most common, affecting 37.7% (n = 159) of participants. These interactions were considered clinically relevant but generally manageable with dose adjustment or monitoring. Major interactions, which could lead to serious adverse effects or significant therapeutic failure, were identified in 23.2% (n = 98) of participants. Glibenclamide and sitagliptin accounted for over 60% of major interaction cases, based on their narrow therapeutic index and CYP3A4 metabolism. Minor interactions, typically requiring no change in therapy, were noted in 18.0% (n = 76) of cases (Table 5).

Table 5.

Medication interaction profile (From interviewer Review, n = 422).

Variable		Frequency (%)
The medication list was verified using:	Medical record	207 (49.1)
	Prescription	154 (36.5)
	Medication container	61 (14.5)
Medications with potential interaction with khat		267 (63.3)
Interaction classification	Pharmacokinetic	178 (42.2)
	Pharmacodynamic	134 (31.8)
	Both	45 (10.7)
Severity of interactions	Minor	76 (18.0)
	Moderate	159 (37.7)
	Major	98 (23.2)

Multivariate analysis of predictors for Khat–OAD interactions in diabetic patients

Predictor variables that were statistically significant in the bivariate analysis were included in the multivariate logistic regression model using the backward elimination method to identify independent predictors of clinically significant interactions between khat use and OADs.

Sex was found to be a significant predictor, with female participants exhibiting higher odds of interaction compared to males (AOR = 1.38; 95% CI: 1.00–1.89; p = 0.048).

Polypharmacy emerged as a strong predictor of interaction. Participants taking 5–6 medications had more than twice the odds of experiencing clinically significant interactions (AOR = 2.12; 95% CI: 1.44–3.12; p = 0.001), and those taking ≥ 7 medications had even higher odds (AOR = 2.43; 95% CI: 1.61–3.67; p = 0.001), compared to participants taking 3–4 medications.

The frequency of khat use was also significantly associated with interaction risk. Daily users had significantly higher odds (AOR = 2.20; 95% CI: 1.52–3.18; p = 0.001), followed by those using khat 3–4 times per week (AOR = 1.89; 95% CI: 1.31–2.74; p = 0.001), compared to those using khat 1–2 times per week.

Regarding the duration of khat use, participants who reported 4–12 years of use had increased odds of interaction (AOR = 1.06; 95% CI: 1.02–1.10; p = 0.004), and those with ≥ 13 years of use had even higher odds (AOR = 1.09; 95% CI: 1.04–1.15; p = 0.002), compared to participants with ≤ 3 years of use.

Concerning OAD types, the use of combined metformin with either sulfonylureas or DPP-4 inhibitors was significantly associated with increased interaction risk (AOR = 1.84; 95% CI: 1.29–2.62; p = 0.001). In contrast, regimens involving glibenclamide/glipizide (± metformin) (AOR = 1.28; 95% CI: 0.91–1.80; p = 0.150) and sitagliptin (± metformin) (AOR = 1.31; 95% CI: 0.90–1.91; p = 0.142) were not statistically significant in the multivariate model (Table 6).

Table 6.

Binary logistic regression predicting clinically significant Khat–OAD interactions (n = 422).

Predictor Variable		COR (95% CI)	p-value	AOR (95% CI)	p-value
Sex	Male	1	–	1	–
	Female	1.42 (1.01–2.00)	0.045	1.38(1.00–1.89)	0.048
Number of Medications	3–4 medications	1	–	1	–
	5–6 medications	2.08 (1.42–3.05)	0.001	2.12 (1.44–3.12)	0.001
	≥ 7 medications	2.35 (1.55–3.57)	0.001	2.43 (1.61–3.67)	0.001
Frequency of khat Use	Daily	2.15 (1.50–3.07)	0.001	2.20 (1.52–3.18)	0.001
	3–4 times/week	1.77 (1.26–2.49)	0.001	1.89 (1.31–2.74)	0.001
	1–2 times/week	1	–	1	–
Duration of khat Use (years)	≤ 3 years	1	–	1	–
	4–12 years	1.55(1.14–2.11)	0.005	1.06 (1.02–1.10)	0.004
	≥ 13 years	1.68 (1.18–2.38)	0.004	1.09 (1.04–1.15)	0.002
OAD Type	metformin only	1	–	1	–
	glibenclamide, glipizide (± metformin)	1.36 (0.95–1.93)	0.091	1.28 (0.91–1.80)	0.150
	sitagliptin (± metformin)	1.42 (0.97–2.08)	0.069	1.31 (0.90–1.91)	0.142
	metformin + sulfonylurea/DPP-4	1.76 (1.26–2.46)	0.001	1.84 (1.29–2.62)	0.001

Bold values = Statistically significant (p < 0.05).

Emerging themes on khat use and diabetes management: insights from in-depth interviews

Data from in-depth interviews with 20 purposively selected participants—including 10 patients, 4 physicians, 3 pharmacists, and 3 nurses—yielded valuable insights into the behavioral, clinical, and systemic dynamics influencing the interaction between khat use and T2DM management.

A key theme that emerged was the limited awareness of potential drug interactions. Notably, 80% (8 out of 10) of patient participants reported that they had never considered the possibility that khat could interfere with their OADs.

Another consistently reported theme was the cultural normalization of khat chewing. All interviewees (100%) across the various stakeholder groups acknowledged that khat use is deeply embedded in local customs and daily life.

In addition to behavioral insights, the interviews revealed critical clinical concerns related to glycemic control. All physicians interviewed (100%, 4 out of 4) reported observing erratic blood glucose levels in patients who regularly chewed khat.

Despite these risks, the perceived benefits of khat were frequently mentioned by patients. Half of the patient participants (5 out of 10) stated that khat helped them feel more focused, productive, or emotionally balanced.

Communication gaps between providers and patients also emerged as a significant issue. Two-thirds (66.7%, 6 out of 9) of healthcare providers—including nurses, pharmacists, and physicians—reported challenges in obtaining accurate information from patients about their khat use. Nurses specifically highlighted that patients often withheld such information unless directly asked, making it difficult to assess potential risks comprehensively.

Finally, the interviews revealed a strong need for integrated counseling and clinical guidelines. All three pharmacists interviewed (100%) expressed concern about the lack of formal institutional support, standardized counseling tools, or clinical protocols tailored to managing drug–khat interactions.

Discussion

This study provides critical insights into the intersection of traditional substance use and modern pharmacotherapy. It highlights the significant prevalence and clinical implications of khat–OAD interactions among polypharmacy patients with T2DM in Northwest Ethiopia.

The demographic distribution, with a mean age of 52.4 years and a high proportion of middle-aged adults, mirrors population profiles of T2DM across similar low- and middle-income countries (LMICs). In these regions, middle-aged adults are increasingly affected by chronic non-communicable diseases (NCDs) due to urbanization and lifestyle changes¹. The high prevalence of comorbidities (71.6%) and polypharmacy (69.2% using ≥ 5 medications) is consistent with global findings, which emphasize the complex therapeutic needs of diabetic populations⁵.

The use of combination OAD regimens—particularly metformin with sulfonylureas or DPP-4 inhibitors—aligns with treatment guidelines in LMICs, where cost-effective pharmacotherapy is prioritized⁴. However, polypharmacy significantly increases the risk of clinically relevant drug–herb interactions, as demonstrated in this study and supported by international reviews³.

Over 60% of participants were exposed to medications with known or potential interactions with khat, and major interactions were observed in nearly a quarter (23.2%) of cases. Cathinone, the psychoactive component of khat, influences hepatic enzymes such as CYP2D6 and CYP3A4, which are critical in the metabolism of drugs like glibenclamide and sitagliptin¹⁴. The finding that these two agents accounted for over 60% of major interaction cases is therefore pharmacologically plausible.

Although the observational design of this study does not allow for definitive causal inference, the reported effects can be reasonably attributed to the combination of khat chewing and OAD use based on several lines of evidence. First, a temporal association was observed, as participants reported concurrent khat use and OAD intake, with adverse effects such as glycemic instability occurring during periods of khat chewing. Second, pharmacological plausibility supports this link, given cathinone’s established effects on CYP enzymes and autonomic function, which provide a mechanistic basis for altered drug metabolism and glycemic variability. Third, triangulation of data sources—combining quantitative findings on interaction prevalence with qualitative reports from patients and clinicians—strengthens the likelihood that the observed effects are attributable to the combination rather than to OADs alone. Finally, a dose-dependent relationship was evident, as higher frequency and longer duration of khat use were associated with increased interaction risk, further supporting a direct association.

Cathinone’s appetite-suppressing and sympathomimetic effects may also contribute to glycemic instability. This was confirmed by reports from both patients and clinicians in the qualitative data. Prior studies have similarly shown that khat use can potentiate hypoglycemia (via appetite suppression) or reduce antidiabetic efficacy (via enzyme induction or behavioral non-adherence)^30,31. These findings align with evidence that cathinone modulates autonomic function, indirectly affecting glycemic variability in patients with chronic metabolic diseases³².

The entrenched cultural acceptance of khat, reported by 100% of interviewees, presents a major challenge in clinical management. Previous ethnographic and public health studies from Ethiopia and Yemen have documented similar normalization, highlighting the difficulty of modifying such behaviors through conventional health education approaches^13,33.

Low awareness of khat–drug interactions (28.7% of patients) and limited provider counseling (26.8%) further reflect systemic gaps in pharmacovigilance and patient education. Similar trends are reported in other African and Middle Eastern contexts, where healthcare provider training often lacks herb–drug interaction content¹⁴. This underscores the importance of integrating herbal interaction modules into continuing medical education (CME), especially in regions where traditional medicine is widely practiced³⁴.

The communication gap was further emphasized by qualitative findings. Provider participants noted that patients were often reluctant to disclose khat use, perceiving it as benign or beneficial. This mirrors earlier qualitative research in Ethiopia, which showed that patients are less likely to report khat use during consultations¹².

The multivariate analysis identified several strong predictors of clinically significant khat–OAD interactions. Female sex was significantly associated with higher odds of interaction (AOR = 1.38; 95% CI: 1.00–1.89; p = 0.048). This may reflect sex-based differences in pharmacokinetics, healthcare engagement, or behavioral patterns.

Research indicates that cathinone inhibits cytochrome CYP450 enzymes, notably CYP2D6 and CYP3A4²³, which metabolize various drugs, including some antidiabetic medications³⁵. Although studies on sex-based differences in khat-induced CYP450 inhibition are limited, broader pharmacological evidence shows that women may display different enzyme activity levels due to hormonal variations. These differences could affect drug metabolism and interaction profiles^36–38.

Although khat chewing has traditionally been more common among men, recent studies indicate a rising trend among Ethiopian women³⁹. This shift may increase exposure to potential drug interactions in women. Additionally, differences in healthcare-seeking behavior and adherence to medical advice may further influence sex-specific risks.

The association between female sex and increased risk of khat–OAD interactions underscores the need for healthcare providers to consider sex-specific factors in T2DM management. Tailored patient education and monitoring strategies may be warranted to mitigate risks. Further research is also needed to clarify the mechanisms underlying these sex-based differences and to develop targeted interventions.

Polypharmacy was a strong predictor of clinically significant khat–OAD interactions, consistent with evidence that a higher number of medications increases the risk of drug–drug and drug–herb interactions^40,41. This highlights the clinical challenge in managing diabetic patients in Ethiopia, where comorbidities and multiple prescriptions are common, and emphasizes the need for careful monitoring and individualized treatment plans.

The frequency and duration of khat use were associated with higher interaction risks, indicating a dose-dependent relationship. Additionally, complex treatment regimens, such as combined metformin with sulfonylureas or DPP-4 inhibitors, further increased the likelihood of clinically significant interactions. This underscores the risk associated with complex diabetic regimens in the context of herbal substance use. These findings are consistent with prior studies showing that both exposure intensity (frequency and chronicity of herb use) and treatment complexity (polypharmacy) are key risk factors for adverse interactions^3,7. These results should guide clinicians to proactively evaluate drug regimens in patients with disclosed herbal practices⁴². Clinicians should also be alerted to inquire about herbal product use, particularly in the setting of polypharmacy.

Interestingly, while sitagliptin and glibenclamide were not individually significant predictors in multivariate models, their clinical importance remains high. Both have narrow therapeutic indices and undergo metabolism via CYP450 pathways, reinforcing the need for close monitoring rather than statistical exclusion.

Overall, these findings call for the urgent incorporation of culturally informed pharmacovigilance into diabetic care protocols in Ethiopia and similar settings. The high burden of interaction risk, combined with low awareness and provider training gaps, poses a serious threat to treatment outcomes. Localized clinical guidelines, herbal use risk assessments within electronic health records, and patient decision aids may help bridge these gaps.

Integrating local cultural norms into pharmacotherapeutic strategies is also critical. Partnerships with community and religious leaders could improve adherence to medical advice. Furthermore, pharmacists—given their accessibility and expertise—can play a key role in delivering targeted education on drug–herb interactions.

Strengths and limitations

This study’s primary strength lies in its convergent mixed-methods design, which allowed for triangulation across clinical, behavioral, and systemic domains. By integrating quantitative associations with qualitative narratives, the study provides a multidimensional perspective on khat–OAD interactions and enhances the validity and contextual richness of the findings. The systematic sampling strategy and the use of standardized tools, alongside rigorous thematic analysis procedures, further strengthen the credibility of the results.

Nevertheless, several limitations should be acknowledged. First, the cross-sectional design limits the ability to establish temporal or causal relationships between khat chewing and OAD interaction outcomes. Longitudinal or experimental studies are needed to clarify directionality and causality. Second, reliance on self-reported khat use introduces potential recall bias and social desirability bias, which may have led to under- or overestimation of khat consumption patterns. Although interviewer-administered questionnaires and triangulation with qualitative data helped mitigate this, biochemical validation of khat exposure and drug levels was not feasible in this setting.

Third, the single-center design at Gondar University Hospital may restrict the generalizability of findings to other Ethiopian regions or comparable low-resource settings with differing khat use norms or prescribing patterns. Fourth, although systematic random sampling reduced selection bias, patients who irregularly attend follow-up clinics were likely underrepresented, potentially skewing results toward individuals with better healthcare engagement.

Fifth, the drug interaction assessment relied primarily on secondary databases and published literature. While these are widely used and evidence-based, they may not fully capture the pharmacokinetic and pharmacodynamic complexity of khat constituents in real-world Ethiopian populations. Sixth, the study did not measure clinical endpoints such as HbA1c variability, hospitalization rates, or adverse drug events, which would have strengthened the clinical relevance of the findings.

Finally, the sample size of the qualitative arm (n = 20), though sufficient for thematic saturation, may not have captured the full diversity of perspectives across different sociodemographic groups. Despite these limitations, the study offers important initial insights and provides a foundation for future longitudinal, multicenter, and biomarker-based research to better establish the clinical consequences of khat–OAD interactions.

Conclusion

In Ethiopia, where traditional and modern medical practices intersect, the high prevalence of khat–OAD interactions represents a serious and immediate clinical concern. Addressing this challenge requires coordinated clinical, community, and policy responses to safeguard diabetes care and improve patient outcomes.

Author contributions

A.K.M., H.S.A., TAT., T.D.H, M.A.A, and S.T.G. wrote the main manuscript text and A.T.B. prepared Figs. 1 and 2. All authors reviewed the manuscript.

Funding

This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data are provided within the manuscript.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

The study protocol was reviewed and approved by the University of Gondar, School of Pharmacy (Approval Number: SOP/127/2024). All participants were fully informed of the study’s purpose, procedures, risks, and benefits, and provided written informed consent before data collection. Participation was entirely voluntary, and participants were assured that their decision would not affect their clinical care. To maintain confidentiality and anonymity, unique identification codes were used, and access to the data was restricted to the research team only.

Consent for publication

Not applicable.