Abstract
Background & Objectives:
Kratom (Mitragyna speciosa) use is associated with polysubstance use (PSU) and use disorders. However, additional research on PSU heterogeneity in populations using this novel psychoactive substance is necessary. The authors investigated patterns of past 12-month PSU among U.S. adults reporting past 12-month use of kratom and at least one additional substance.
Methods:
Latent class models were fit using 2019 National Survey on Drug Use and Health (NSDUH) data which was collected from 412 U.S. adults reporting past 12-month use of kratom and at least one of 11 additional substances.
Results:
Three distinct profiles were identified: “marijuana/alcohol/tobacco (MAT)” (63.3%), “marijuana/alcohol/tobacco + psychedelics (MAT+P)” (19.3%), and “marijuana/alcohol/tobacco + psychedelics/heroin/prescriptions (MAT+PHPR)” (17.4%).
Conclusions & Scientific Significance:
This is the first epidemiological study in which a latent class analysis was used to identify unique PSU profiles among US adults using kratom and other substances. Understanding the profiles of people using kratom in relation to the use of other drugs might help guide screening interventions, treatment needs, and policy.
Keywords: Kratom, Mitragyna speciosa, Polysubstance use, Latent class analysis, National Survey on Drug Use and Health (NSDUH)
Introduction
Kratom (Mitragyna speciosa) is a psychoactive herb that some people report using to manage psychiatric symptoms, treat pain, or reduce the use of other substances.1,2 In the United States (U.S.), past-year prevalence of kratom use is estimated to be 0.7-4.1%.3 Most people using kratom do not experience negative health effects,1 a growing amount of clinical and/or forensic literature elucidates adverse outcomes – including mortality – when kratom was used with other substances.2,4,5 For instance, Corkery et al. found that polysubstance use (PSU) was present in 87.2% of known kratom-associated fatalities.5 This is concerning given positive associations between kratom and other substance use or use disorders.6,7 However, the addiction field currently lacks reliable, prevalence estimates of PSU profiles in people using kratom.
This study described the heterogeneity of past 12-month PSU among adults who reported past 12-month use of kratom and one or more of 11 additional substances. Sociodemographic differences between these profiles were also investigated. Findings from these secondary analyses could provide crucial insight into variations in kratom and PSU profiles in the U.S. population. Such insights could help clinicians identify adults using kratom who may need interventions to prevent or reduce PSU-associated, adverse health outcomes.
Methods
Data from the 2019 National Survey on Drug Use and Health (NSDUH) – a national, cross-sectional survey – was utilized.8 Additional information on NSDUH and its methodology has been previously reported.8 From 56,136 responses, the authors excluded 13,397 NSDUH participants younger than 18 years old. NSDUH respondents who did not report past 12-month kratom use (n = 41,980), those who were identified as having “bad” data (n = 3), or participants who did not respond to this question (n = 308) were eliminated. The authors also excluded 36 respondents who indicated that they either 1) had not used any of the 11 additional substances in the past 12 months or 2) had missing responses to these questions. This yielded a final sample size of 412 U.S. adults who reported past 12-month use of kratom and at least one additional substance.
Measures
All NSDUH participants were asked about their lifetime and past 12-month use of substances, including kratom, tobacco (cigars, cigarettes, smokeless tobacco, pipe), alcohol, marijuana, cocaine (including crack), heroin, psychedelics, methamphetamines, and inhalants as well as non-medical use (NMU) of prescription pain relievers, stimulants, and benzodiazepines. Participants also reported how long it had been since they last used these substances, when applicable.9 Responses for each substance were re-categorized as binary measures to indicate past 12-month use (yes/no).
Sociodemographic variables included age cohort (18-25, 26-34, 35-49, 50+), sex (male, female), race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic, other), health status (excellent/very good/good, fair/poor), past 12-month insurance coverage (yes, no), educational attainment (<high school, high school, some college, college), and past 12-month serious psychological distress (SPD) – assessed by NSDUH-derived Kessler K6 scores.9
Data analyses
Latent class analyses were conducted to identify profiles with distinct patterns of past 12-month PSU. The authors estimated model fit for a series of latent class analysis models starting with a one class model and continuing until the authors observed that the model’s parameters were no longer improving, and profile sample sizes would prohibit additional analyses (1-6 profiles). The optimal number of profiles was selected based on the clinical & public health implications of the identified profiles;10 the model with lower value for the Akaike information criterion (AIC), Bayesian information criterion (BIC), and sample-size adjusted Bayesian information criterion (SABIC), as well as larger entropies and log-likelihood values, were preferred.10
After the optimal model was chosen, chi-square tests of independence and logistic regressions were used to assess significant sociodemographic differences among participants in each of the identified latent profiles. All chi-square tests of independence and logistic regression models were conducted using SAS statistical software, version 9.4 (SAS®, Cary NC). Latent class analyses were conducted using Mplus, version 8.4 (Los Angeles, CA). The authors accounted for the NSDUH’s complex sampling design in the analyses by following Substance Abuse and Mental Health Data Archive (SAMHDA) guidelines.8
Results
Among 412 U.S. adults reporting past 12-month kratom and PSU, more than a third (39.8%) reported past 30-day kratom use. Most were male (59.7%), Non-Hispanic White (83.0%), were insured (84.2%) and had completed some college (40.5%). Approximately one third (36.3%) of respondents had experienced past 12-month SPD. Not including kratom, participants used an average of 3.4 (SE=0.1, range: 1-11) additional substances within the past 12 months. Indeed, 90.0% reported past 12-month alcohol use, 69.8% reported past 12-month tobacco use, and 67.8% reported past 12-month marijuana use.
Latent Class Model
Although models with 3 and 4 profiles were statistically optimal, the clinical interpretability was optimized using 3 profiles. In this model, nearly two-thirds (63.3%) of participants were included in the “marijuana/alcohol/tobacco (MAT)” profile which was characterized by high rates of marijuana, alcohol, and/or tobacco use and relatively low rates of other substance use. Approximately a fifth (19.3%) were grouped in the “marijuana/alcohol/tobacco + psychedelics (MAT+P)” profile. Lastly, 17.4% reported more expansive PSU, including high rates of past 12-month use of marijuana, alcohol, tobacco, psychedelics, heroin, and/or NMU of prescription medications; this third profile was named the “marijuana/alcohol/tobacco + psychedelics/heroin/prescription drugs (MAT+PHPR)” profile. Figure 1 details the predicted probabilities of each of the 11 types of substances by class.
Figure 1. Predicted probabilities of past 12-month polysubstance use by latent profile (n = 412).
a MAT PSU = marijuana/alcohol/tobacco polysubstance use profile; MAT+P PSU = marijuana/alcohol/tobacco + psychedelics polysubstance use profile; MAT+PHPR PSU = marijuana/alcohol/tobacco + psychedelics/heroin/non-medical use of prescription drugs polysubstance use profile
* indicates significance (p < 0.05)
Latent Class Sociodemographic Characteristics
In comparison to those 50+, those 18-25 years of age had significantly higher odds of being in the MAT+P or MAT+PHPR profiles, compared to the MAT profile (MAT+P: adjusted odds ratio [AOR]=13.5, 95% CI: 2.2, 83.6; MAT+PHPR: AOR=11.4, 95% CI: 1.6, 79.6). Compared to those with no past 12-month SPD, those with SPD had a 2.2 (95% CI: 1.1, 4.5) increased odds of being in the MAT+PHPR profile rather than in the MAT profile. Additional findings on the sociodemographic differences between the three profiles are depicted in Table 1.
Table 1.
Unadjusted and adjusted multivariate logistic regression models between demographic and substance-related characteristics and latent class profiles (n = 411)
| Unadjusted Models | Adjusted models | |||
|---|---|---|---|---|
| “MAT+P” PSUa OR (95% CI) |
“MAT+PHPR” PSU OR (95% CI) |
“MAT+P” PSU* AOR (95% CI) |
“MAT+PHPR” PSU AOR (95% CI) |
|
| Age cohort | ||||
| 18 – 25 | 9.9 (1.8, 55.3) | 11.1 (1.4, 86.8) | 13.5 (2.2, 83.6) | 11.4 (1.6, 79.6) |
| 26 – 34 | 4.1 (0.7, 25.2) | 9.0 (1.1, 77.2) | 4.6 (0.6, 35.4) | 8.8 (1.0, 75.2) |
| 35 - 49 | 1.7 (0.2, 12.3) | 1.3 (0.1, 12.9) | 1.6 (0.2, 12.0) | 1.0 (0.1, 11.4) |
| 50 and older | -- | -- | -- | -- |
| Sex | ||||
| male | -- | -- | -- | -- |
| female | 0.7 (0.4, 1.2) | 0.6 (0.2, 1.4) | 0.6 (0.3, 1.2) | 0.4 (0.2, 1.0) |
| Race/ethnicity | ||||
| Non-Hispanic White | -- | -- | -- | -- |
| Non-Hispanic Black | 0.4 (0.1, 2.5) | 0.4 (0.1, 2.7) | 0.2 (0.0, 1.5) | 0.3 (0.0, 2.0) |
| Hispanic | 0.9 (0.3, 2.8) | 0.7 (0.2, 2.5) | 0.3 (0.1, 1.2) | 0.5 (0.1, 1.8) |
| other | 0.5 (0.1, 1.8) | 0.1 (0.0, 0.5) | 1.0 (0.4, 2.5) | 0.1 (0.0. 0.5) |
| Health status | ||||
| good/very | ||||
| good/excellent | -- | -- | -- | -- |
| fair/poor | 0.7 (0.4, 1.3) | 0.7 (0.3, 1.4) | 1.1 (0.6, 2.1) | 0.9 (0.4, 2.0) |
| Educational attainment | ||||
| less than high school | -- | -- | -- | -- |
| high school grad | 1.8 (0.3, 12.0) | 1.0 (0.3, 3.5) | 2.0 (0.3, 15.4) | 0.9 (0.3, 3.2) |
| some college | 3.2 (0.7, 14.6) | 1.7 (0.6, 5.1) | 5.1 (1.0, 25.7) | 2.4 (0.6, 9.2) |
| college grad | 1.7 (0.4, 7.4) | 0.7 (0.2, 2.7) | 3.8 (0.9, 16.7) | 1.2 (0.3, 4.8) |
| Past 12-month health insurance | ||||
| no | -- | -- | -- | -- |
| yes | 0.7 (0.3, 1.6) | 0.6 (0.3, 1.2) | 0.8 (0.3, 2.1) | 0.9 (0.4, 2.0) |
| Past 30-day serious psychological distress | ||||
| no | -- | -- | -- | -- |
| yes | 1.4 (0.6, 2.9) | 2.1 (1.0, 4.1) | 1.3 (0.5, 2.9) | 2.2 (1.1, 4.5) |
Referent group for both unadjusted and adjusted models is the MAT (marijuana/alcohol/tobacco) PSU profile
MAT+P PSU = marijuana/alcohol/tobacco + psychedelics polysubstance use profile
MAT+PHPR PSU = marijuana/alcohol/tobacco + psychedelics/heroin/non-medical use of prescription drugs polysubstance use profile
Discussion
Among a sample of U.S. adults who reported past 12-month use of kratom and at least one additional substance, a latent class analysis was used to identify three distinct profiles (MAT, MAT+P, and MAT+PHPR) to describe the variability in past 12-month PSU. Most participants were classified in the MAT profile. Significant differences in age cohorts were observed among the profiles. Consistent with prior research, substance use, including psychedelics, cocaine, and inhalants, tend to be more prevalent among U.S. adults between 18-25 years of age, compared to those 26+ years of age.11 Clinicians - especially those treating younger patients - should consider routinely screening patients for emerging psychoactive substances, including herbal products and other legal botanicals. Given the variability in kratom-based products and use patterns, health providers should specifically inquire about a patient’s use of kratom, including the type of product used and dosage.12 Eliciting information related to patients’ motives for kratom use may also be beneficial since some individuals report using kratom to reduce or quit their use of other drugs, including opioids.1 Patients reporting kratom and polysubstance use should be offered timely and appropriate preventative services and, if necessary, intervention.
Several limitations restrict the interpretation of these analyses. First, the authors defined PSU as using two or more substances within the past 12 months, rather than simultaneously or sequentially. Second, the selected model had an entropy value of 0.7, which is lower than the recommended level of 0.8; however, the authors felt it was clinically superior to the models with an entropy of 0.8.10 Third, this dataset did not permit the authors to examine the frequency, quantity, dose, motives, and routes of administration for kratom use. Despite these limitations, this study provides preliminary information to advance the understanding of PSU profiles among people using this psychoactive herb and other drugs.
Conclusions
To the authors’ knowledge, this is the first study examining the heterogeneity of PSU among a sample of U.S. adults who reported past 12-month use of kratom and one or more additional substances. To further the clinical utility of this work, future research should identify motives for PSU and elucidate specific drug-herb interactions that may increase the risk for kratom-associated, adverse health outcomes. This may guide health providers in tailoring medical guidance, treatment modalities and needs, and provision of appropriate resources and support services for patients reporting the use of kratom and other substances.
Acknowledgments
Research reported in this publication was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under University of Florida and Florida State University Clinical and Translational Science Awards TL1TR001428 (CCH: August 2022-Present, PI: Wayne T. McCormack, University of Florida, Gainesville, FL) and UL1TR001427 (PI: Duane A. Mitchell, University of Florida, Gainesville, FL). This work was partially supported by the University of Florida’s Department of Epidemiology (AMF: May 2022-Present; CCH: May 2022-August 2022) and the UF Substance Abuse Training Center in Public Health from the National Institute of Drug Abuse (NIDA) of the National Institutes of Health under award number T32DA035167 (PI: Linda B. Cottler, University of Florida, Gainesville, FL) (AMF: May 2019-May 2022; CCH: May 2020-May 2022). In addition, this work is supported by award number K01DA046715 (PI: Lopez-Quintero, Gainesville, FL). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Declaration of Interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.
Data Statement
The data used in these analyses are readily available by the Substance Abuse and Mental Health Services Administration. Data can be accessed by going to https://www.samhsa.gov/data/release/2019-national-survey-drug-use-and-health-nsduh-releases.
References
- 1.Garcia-Romeu A, Cox DJ, Smith KE, Dunn KE, Griffiths RR. Kratom (Mitragyna speciosa): User demographics, use patterns, and implications for the opioid epidemic. Drug and Alcohol Dependence. 2020;208:107849. doi: 10.1016/j.drugalcdep.2020.107849 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Striley CW, Hoeflich CC, Viegas AT, et al. Health Effects Associated With Kratom (Mitragyna speciosa) and Polysubstance Use: A Narrative Review. Subst Abuse. 2022;16:11782218221095872. doi: 10.1177/11782218221095873 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Henningfield JE, Grundmann O, Garcia-Romeu A, Swogger MT. We Need Better Estimates of Kratom Use Prevalence. Am J Prev Med. 2022;62(1):132–133. doi: 10.1016/j.amepre.2021.07.022 [DOI] [PubMed] [Google Scholar]
- 4.Post S, Spiller HA, Chounthirath T, Smith GA. Kratom exposures reported to United States poison control centers: 2011–2017. Clinical Toxicology. 2019;57(10):847–854. doi: 10.1080/15563650.2019.1569236 [DOI] [PubMed] [Google Scholar]
- 5.Corkery JM, Streete P, Claridge H, et al. Characteristics of deaths associated with kratom use. J Psychopharmacol. 2019;33(9):1102–1123. doi: 10.1177/0269881119862530 [DOI] [PubMed] [Google Scholar]
- 6.Palamar JJ. Past-Year Kratom Use in the U.S.: Estimates From a Nationally Representative Sample. American Journal of Preventive Medicine. 2021;61(2):240–245. doi: 10.1016/j.amepre.2021.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Xu KY, Mintz CM, Borodovsky JT, Glaser PEA, Bierut LJ, Grucza RA. Prevalence of Kratom Use and Co-Occurring Substance Use Disorders in the United States. Prim Care Companion CNS Disord. 2021;23(4):21br02930. doi: 10.4088/PCC.21br02930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Center for Behavioral Health Statistics and Quality. 2019 National Survey on Drug Use and Health: Methodological Summary and Definitions. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2020. https://www.samhsa.gov/data/report/2019-methodological-summary-and-definitions [Google Scholar]
- 9.Center for Behavioral Health Statistics and Quality. 2019 National Survey on Drug Use and Health Public Use File Codebook. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2020. https://www.datafiles.samhsa.gov/sites/default/files/field-uploads-protected/studies/NSDUH-2019/NSDUH-2019-datasets/NSDUH-2019-DS0001/NSDUH-2019-DS0001-info/NSDUH-2019-DS0001-info-codebook.pdf [Google Scholar]
- 10.Weller BE, Bowen NK, Faubert SJ. Latent Class Analysis: A Guide to Best Practice. Journal of Black Psychology. 2020;46(4):287–311. doi: 10.1177/0095798420930932 [DOI] [Google Scholar]
- 11.Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2019 National Survey on Drug Use and Health (HHS Publication No. PEP20-07-01-001, NSDUH Series H-55). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration; 2020. Retrieved from https://www.samhsa.gov/data/ [Google Scholar]
- 12.Smith KE, Dunn KE, Epstein DH, et al. Need for clarity and context in case reports on kratom use, assessment, and intervention. Subst Abus. 2022;43(1):1221–1224. doi: 10.1080/08897077.2022.2074608 [DOI] [PubMed] [Google Scholar]