This study reported statistically significant relationships between ED and all ABs. Problems and severity were the indicators that showed the strongest effect sizes in relation to ED. Sex and continental region were effective moderators of the relationship between ED and gambling. Moreover, there were stronger relationships in older participants than in younger participants (i.e., alcohol and gambling) and in clinical compared to community samples (i.e., cannabis). Minimal impacts of publication bias were found overall.

There were differences in the relationships between specific ED strategies and ABs. Both lack of strategies and difficulty in controlling impulses in the face of negative emotions were the only ED dimensions related to all substance and non-substance ABs. This maps well with the evidence supporting impulsivity conceptually overlapping the impulse dimension of ED (Willie et al., 2022) and means that difficulties in ER are in part due to impulse control behaviors, which are common in people with substance use disorders or presenting behavioral addictions (Cyders and Smith, 2008; Di Pierro et al., 2015). From this point of view, impulsivity may confer high vulnerability and prevent people envisioning possible negative consequences stemming from ABs, making it harder to access adaptive emotional regulation strategies (López-Torres et al., 2021).

Gaming and gambling were more strongly related than the other ABs to overestimation of negative emotional experiences (catastrophizing) and difficulties in impulse control (impulse). Catastrophizing of life events may confer vulnerability to gambling/gaming as a way of coping given these behaviors may serve to anesthetize disturbing emotional experiences (Melodia et al., 2022; Wang et al., 2020). On a similar note, difficulties in ER are particularly intertwined with two conditions that are prevalent in both gamblers and gamers, anxiety, and depression (Bridges-Curry and Newton, 2022; Daros et al., 2021), which also lead to risk of increased gambling/gaming by affecting how negative emotions are managed (Gerdner and Håkansson, 2022; Neophytou et al., 2023). In gamblers specifically, the prominence of these strategies may be accounted for by the many legal consequences (money lost, seizure of properties) of gambling and its negative impact on interpersonal relationships (e.g., loss of partners) (Fong, 2005). In addition, the significant relationship between self-blame strategies and gambling may be related to more severe emotional distress caused by feelings of guilt and the blame from relatives about the onset of this behavior and any ongoing relapses (Fan, 2020).

For alcohol, lack of emotional control and modulate were the ED strategies that showed the largest magnitude of effects. This suggests that alcohol users may have difficulties in reducing the intensity and the length of time they experience negative emotions. These behaviors have been related to impaired executive functions. In fact, there are several empirical studies suggesting that alcohol use hinders inhibitory control skills, which leads to difficulties in inhibiting alcohol use in the presence of alcohol-related stimuli (e.g., negative affect) (Fleming & Bartholow, 2014; López-Caneda et al., 2014).

For tobacco, lack of awareness, non-acceptance, and impulse were the ED strategies that showed the largest magnitude of effects, which suggests tobacco users would benefit from interventions providing psychoeducation on emotions. Awareness is associated with faster stress recovery following stressors (Borges, 2020). For this reason, training strategies aimed at increasing awareness of affective responses may be particularly important in that they can serve to buffer the negative symptoms associated with nicotine's pharmacological effects (e.g., increases in anxiety because of cravings). In addition, acceptance strategies to help managing urges to smoke are highly encouraged. This may include increased attention towards (and reduced avoidance of) negative symptoms through mindfulness-based and Acceptance and Commitment Therapy (ACT) (Barlow & Eustis, 2022).

With regards to moderating variables, it is worth noting that, except for gambling, no association was found between sex, ED, and ABs, and this suggests ED is a risk factor for substance use and gaming regardless of sex. Age and type of population (clinical vs. community) were found to moderate several of the tested associations. The relationship between ED and alcohol was stronger in older populations, which is in line with existing research (Aldao et al., 2010). Although ER competence improves during life (Riediger and Bellingtier, 2021), it notably decreases in people with high levels of emotional activation, such as those struggling with ABs and emotional disorders (Lincoln et al., 2022).

Our findings showed a significant impact of continental region on the association between ED and gambling. Oceania, followed by Asia, Europe, and North America, exhibited the largest effect sizes. This finding is in line with research concluding that people's emotional regulation processes are shaped by culture (Ip et al., 2021; Su et al., 2015). Generally, gambling has less stigma than substance use, as its social status is judged to be higher (Dabrowska et al., 2020), which may in part account for the moderating effects seen in gambling but no other ABs. This has particular significance for Asian cultures that tend to value control overt behavior and use suppression and moderation of intense emotional experiences (Su et al., 2015). Moreover, Asian culture emphasizes the ‘interdependent self’, which suggest more dysfunctional strategies (e.g., suppression) may be deemed as culturally adaptive for the sake of harmonious social relationships (Ip et al., 2021). For western cultures, both the expression of and focus on emotional experiences may account for the larger observed estimates, as the process of enculturation in these countries emphasizes the need to foster positive emotions that include a positive sense of self and avoidance of negative emotions (e.g., sadness, depression) (Jung et al., 2009).

Interestingly, studies including clinical samples of cannabis users showed stronger associations. This is consistent with the fact that cannabis is overrepresented in people undergoing substance use treatment (Andersson et al., 2021; Pinto et al., 2019). What these results suggest is the importance of screening for ED in cannabis users and delivering ED-focused interventions to avoid falling into more severe patterns of use.

This study is not without limitations. Firstly, conducting multiple statistical tests in separate meta-analyses increases the risk of Type I error (false positives). While the random-effects approach provides some evidence for our hypothesis, it does not actually provide a truly comprehensive summary of the variables as if a single analytical model had been run. Secondly, because it examined cross-sectional relationships, no conclusions can be drawn about causality. Research on this topic is still in the early stages, and there is a need for more longitudinal studies on the association between ED and ABs before conducting a meta-analysis. Thirdly, we were unable to conduct a meta-analysis on cocaine as there was only one study that provided overall effect sizes between ED (total score) and cocaine. Other variables (e.g., ED dimensions, impulsivity, mental health disorders) were not considered as potential moderators. Virtually all studies provided data on the DERS-36 (149/189) and there was not sufficient variability in the ED measures to conduct a moderation analysis. In addition, none of the included studies provided detailed information on the presence of other mental health disorders and levels of impulsivity, and those providing such data were highly varied in terms of the variables and questionnaires they used, meaning it was not possible to analyze its moderating role. This poses an intrinsic limitation of the reviewed literature and there is need for future studies to look at individual moderators of the association between ED and ABs.