This study presents updated epidemiology of substance use data in the adult population of Malaysia. To our knowledge, this is the first paper to identify the pattern of substance use using LCA among a nationally representative adult population in Malaysia. We identified three distinct classes of substance use among the adult population: those who primarily use a combination of tobacco and alcohol (Tob+Alc), those who use multi-drug (including soft and hard drugs) and those who were negligible or non-drug users. This three class solution from LCA was interpretability and relevant fit statistics model. Interestingly, the latent classes we identified are somewhat consistent with the findings of previous work that investigated the profile of substance use among men who have sex with men (MSM) in Malaysia which identified a three class solution as the best fitting model: “negligible substance use”, “soft substance use” and “amphetamine-type stimulant use” [23]. Similar to our findings, they also found that majority of participants were negligible drug users or non-drug users and detected a small group of sample using hard drugs or multi-drugs.

Our work extends the literature in that we added kratom as one of the indicators for LCA in addition with other common substance use. We found that there was high probability of co-use of kratom with other multiple substance use (refer to class 2 latent class as visually represented in Figure 1 and Figure 2). This finding was supported by previous studies in Malaysia showed that substance users in Malaysia used kratom to abstain from illicit drugs and to manage stimulant withdrawal symptoms [25,26,34]. A study on kratom consumption in the northern areas of peninsular Malaysia also reported that they used kratom to reduce their intake of more expensive opiates like heroin [25]. Although kratom has perceived therapeutic effects, several studies suggested abuse and addiction potential, synergistic toxic effects and fatal interactions with other psychoactive drugs [35,36]. Furthermore, there has been no report on the fatalities in Southeast Asia caused by the ingestion of pure kratom alone. Although the enforcement of kratom related offences in Malaysia is under the jurisdiction of the Poisons Act 1952 [37], the issue of misuse and offences must be addressed by all parties including law, policymakers and the community to improve the enforcement work done by the authorities.

In our Multinomial regression results, we chose Class 3 (non/negligible substance use) as the reference category to distinguish the pattern of polysubstance use from predominantly negligible or non-users. As reported in numerous literature, gender (male) is an important demographic characteristic associated with polysubstance use [16,19,38]. This study detected a significant association of the male gender with the Tob+Alc group but not for the multi-drug group. This may be due to a very low prevalence of multi-drug use among female respondents. In terms of age differences, the young adults (18 – 40 years) demonstrated higher prevalence and risk of being a member of polysubstance use than older adults (≥60 years), for both Tob+Alc and multi-substance classes. Our findings are in line with the study from The United States that reported the co-use of tobacco and alcohol were highest among young adults and declined with increasing age [39]. As well as for other polysubstance the risk of use, abuse/dependence and use of other forms of illicit drugs after the use of marijuana/cannabis declined with increasing age [15]. Another LCA study among the Great Britain population also reported that being young was associated with an increased likelihood of membership of the polydrug use class’s wide range and moderate range [38]. There are several explanations of this trend for polysubstance use to decline with increasing age. This includes socio maturity where increasing age will increase the social maturity on the individual’s ability to resist other illicit drug use [40]. Other explanations might be due to a recruited effect or age of onset of first drug use such as marijuana, those who take up marijuana later would be less predisposed to other illicit drug use than those who take it up early.

Our findings of polysubstance use varying among racial/ethnic groups are consistent with the results from other studies [16,19,39,41]. For example, a study in the United States found that Whites were more likely to drink while Natives American/Alaskan Natives were most likely to smoke and co-use alcohol and tobacco [39]. Our results showed that the co-use of alcohol and tobacco were more likely to occur among Indigenous Sabah & Sarawak, followed by Indian and Chinese ethnicity as compared to Malay ethnicity. However, according to the LCA model, the Malays ethnicity demonstrated the highest prevalence of multi-substance use (which had low probabilities of taking alcohol) compared to other ethnic groups. Our study was in line with the findings noted in another epidemiological study in Malaysia reporting similar results, although that study involved adolescents [42]. These findings suggest that the culture and practices of respective ethnic groups might play an important role in predicting substance use. In Malaysia, Malay adults had low prevalence of alcohol use but high prevalence of tobacco use, while Indigenous Sabah & Sarawak showed high prevalence for both alcohol and tobacco consumption [2,43]. In this country, Muslims (the majority are of the Malay ethnicity) are forbidden from consuming alcohol, and it is also illegal to sell any alcoholic beverages to them. In contrast, drinking alcohol is a culture and social obligation among other ethnic groups.For example, indigenous Sabah & Sarawak consume alcohol during their festivals and social gatherings [44].

The present findings also suggested that adults with low education levels (no formal education, primary education and secondary education) were more likely to use polysubstance of Tob+Alc group compared to higher education levels (tertiary education). However, no significant association was found between education levels and polysubstance of multi-substance group. Our findings is consistent with the latent class analysis of polysubstance from European studies which reported higher education was associated with a reduced risk of moderate but not high polysubstance class [16,38]. They found that there was a reduced risk of moderate polysubstance group membership for participants with an education level beyond secondary education (General Certificate of Secondary Education, GCSE).

Our study also revealed that working adults is a significant predictor for polysubstance use in the Tob+Alc class. This is consistent with the epidemiological data in Malaysia, which reported that the prevalence of current smokers and drinkers were higher among working groups, especially self-employed and private employees compared to those not working [2,43]. Working conditions may influence addictive behaviours such as exposure to psychological job strain (due to high work demands and low coping skills) that may lead to an increased risk of substance use especially alcohol and tobacco [45-47]. The misuse of alcohol by workers represents an important social policy issue because it has the potential to influence the employee’s job performance. For instance, even moderate daily alcohol consumption can lead to absenteeism and smoke breaks among employees that can be considered disruptive as it takes time away from work.

Taken together, the results from the current study suggest that prevention and treatment strategies for polysubstance use especially for the co-use of alcohol and tobacco should place a special emphasis on young adults and other high-risk groups such as people from the Indigenous Sabah & Sarawak ethnicity and those who are working. Whilst, the prevention strategies for polysubstance use of multi-drugs including illicit drugs, should focus on young adults and people from the Malay ethnicity. In addition, kratom use should not be reasonably expected to be safe, especially for co-use with other drugs, and it poses a public health threat. Thus, stronger measures to control of this drug are warranted and the use of kratom in Malaysia should be further investigated, including the effects of co-use with other drugs.

This study needs to be interpreted with caution in light of several limitations. The primary limitation of this study is the relatively low prevalence rate of multi-drug uses such as opioids, marijuana and amphethamine type stimulants (ATS). Due to the small number of cases of illicit drug use, our study may have been underpowered to identify additional classes and differentiate the pattern of polysubstance classes between hard drugs and soft drugs in the multiple-drug group (class 2) and other potential variables associated with multi-drug use. Our study also used self-reported substance use where the prevalence rates could be misrepresented. However, past researchers suggested suitable reliability and validity of studies with self-reported substance use [48,49]. The limitation of this study was the sample recruited from a cross-sectional design of a population-based survey, and it only investigated the socio-demographic predictor factors related to polysubstance use. Although other studies [50-52] have found the association between polysubstance use and psychological distress, as our study extracted the data from a large population-based survey which is almost similar to other existing population-based studies mainly cross-sectional, we have no capacity to consider prospectively early life predictors and to describe the association with psychological disorder. Future studies should replicate and extend these findings with a larger sample using a cohort design and to measure the association of polysubstance use with other early life predictors and mental health disorders.