This study on the risks of SUDs in adult children with severe parental AUD found that the risks of specific SUDs (e.g., opioid use disorder, cocaine use disorder) and of composite SUD outcomes (any, one, or two or more SUDs) were elevated in adult children with severe parental AUD. The highest excess risks of specific SUDs were for other stimulant, volatile solvent, opioid, and multiple drug use disorders. After adjusting for sociodemographic factors, parental SUD, and parental psychiatric disorder in the final model, risks of specific SUDs declined, and HRs of the diagnoses converged to approximately twice as high in the adult children of parents with than without severe AUD. This convergence suggests that excess risk of SUDs is similar across different SUD diagnoses. Factors that stood out as particularly influential in specific outcomes included sociodemographic factors in the increased risk of stimulant use disorder, as well as parental SUDs other than AUD in the increased risk of multiple SUDs. The findings on composite SUD outcomes showed that drug combinations that included alcohol played a noteworthy role in the excess risk of multiple SUDs and, by extension, any SUD. It is worth highlighting that despite their elevated risk, the great majority of adult children with severe parental AUD remained free of diagnoses of even the most common SUDs, including AUD.

With regard to the current study’s findings on incidence, in keeping with prior research [15], AUD was the most common SUD diagnosis in the study population. Multiple drug use disorder (ICD-10 code F19) was also common in the cohort, as was the composite variable two or more SUDs. The frequency of multiple SUDs in the study population is consistent with findings that multiple substance use is common in people treated for SUDs [16, 17].

Turning to the findings on hazard ratios, the elevated risk of SUDs observed in the adult children with severe parental AUD was in line with the results of previous research [4–7], and the finding that risks of specific SUDs converged in the final model was consistent with the hypothesis that in families with AUD, the risk for non-alcohol SUDs is general rather than substance specific. Most prior research on this topic comes from studies that examined family risk factors and genetics. For example, one study of 803 individuals and their families found that while family AUD predicted AUD in individuals, it was also a consistent and non-specific predictor of other SUDs [10]. A slightly different result was observed in a family genetic study from Sweden [9]. That study found minor specific genetic risk for certain SUDs, but for the most part, SUDs, including AUD, were associated with the risk of other SUDs in a non-specific way. The hypothesis that SUD risk has an important general component is also supported by genome-wide association studies that have observed evidence of a general addiction risk factor for alcohol, tobacco, cannabis, and opioids [18, 19] possibly related to dopamine regulation [18]. Researchers have suggested that a general genetic liability may express itself as use of specific substances based on the availability and acceptability of those substances [20], and the findings of the current study are consistent with that hypothesis.

Other stimulant use disorder was the disorder for which risk was most elevated in those whose parents had severe AUD. Amphetamines and methamphetamines have been used in Sweden since the late 1930s, and use reached a peak in the 1950s [21]. The drugs remained common into the 2000s, more so than in many other European countries [22]. The present study results are consistent with those of a 2009 investigation of Swedish Prison and Probation Service clients who were primary users of amphetamines, heroin, and cocaine [22]. That study found that amphetamines were the most common of the three and that parental alcohol problems were more characteristic of the amphetamine users. In the current study, adjustment for sociodemographic factors had an important impact on the risk of other stimulant use disorder, which is consistent with the history of amphetamine use in Sweden and with research from the United States that found that illicit stimulant use is associated with low socioeconomic status [23].

Another robust finding was the elevated risk of opioid use disorder in the adult children with severe parental AUD. This result is broadly consistent with those in the limited literature on the connection between parental AUD and risk of opioid use disorder in the next generation. For example, a study of 87 treatment-resistant opioid users found that parental use of alcohol was associated with childhood maltreatment, which in turn was associated with starting to inject opioids at a younger age [24]. Furthermore, an analysis of U.S. national survey data connected parental alcohol use with adolescent children’s nonmedical use of prescription opioids, an association that disappeared after adjustment for potential confounders [25]. Finally, another study based on U.S. national survey data found that a history of parental alcoholism significantly increased the odds of prescription drug misuse and disorder in a subpopulation of people who self-identified as Latino [26]. The prescription drug outcome was broad and covered opioids, sedatives, and tranquilizers.

The elevated risk of multiple SUDs, observed for all proxy measures used in the study, was consistent with the increased risk of SUDs in general in people with parental AUD [4–7]. Few studies have examined parental AUD and risk of multiple SUDs in adult children. However, latent-class analysis of data from a U.S. national survey found that parental drinking increased the risk of being in the class with the highest probability of polysubstance use [27].

A result that stood out was the substantial decline in excess risk of all measures of multiple SUDs after adjustment for parental SUD in addition to AUD. Declines were similar regardless of whether the multiple SUD outcome included AUD. The finding indicates that a parent’s dependence on alcohol and at least one other substance is a warning sign that their children are vulnerable to SUDs. The relatively small decline after additional adjustment for parental psychiatric disorder may be explained by high psychiatric comorbidity in people with AUD and other SUDs [1].

The risk of multiple SUDs was meaningfully higher than that of single SUDs only when AUD was included in the outcomes. Thus, alcohol was important in substance combinations in those with severe parental AUD. The finding also suggests that combinations of alcohol and other substances played an important role in the raised risk of two or more SUDs, and by extension, in the overarching category of any SUD. These results are consistent with existing evidence that the offspring of parents with AUD have an elevated risk of AUD [4], and that individuals with one SUD have an increased risk of developing additional SUDs [28].

Cannabis is the most-used illicit drug in Sweden [29] and worldwide [30]. Numerous studies have found that parental alcohol use is associated with cannabis use-related outcomes in the next generation [31–33]. Fewer studies have investigated cannabis-related disorders as outcomes, but researchers who followed nearly 2500 children born in the early 1980s in Brisbane, Australia, for over 20 years found that those whose mothers drank >1 glass of alcohol a day had higher odds of cannabis use disorder than those whose mothers drank ≤ 1 glass a day [34]. Additionally, a cross-sectional study of more than 3000 college students in northeastern France observed a connection between parental alcohol dependence and young adults’ cannabis dependence [35]. However, in an interview study of 719 young adults and their parents in the U.S. state of Oregon, no association was seen between parental AUD and cannabis use disorder [36].

There are few prior studies on the risk of sedative or hypnotic use disorder in the adult children of parents with AUD, so the information provided here is novel. Consistent with the current findings of elevated risk in the adult children with severe parental AUD, a U.S. national survey study found that a history of parental alcoholism significantly increased the odds of prescription drug misuse and prescription drug use disorder, a composite measure that covered sedatives, tranquilizers, and opioids [26].

The risk of cocaine use disorder was elevated in the adult children with severe parental AUD but was not among the most elevated risks in the study. Because patterns of cocaine use differ in Sweden and elsewhere, caution is needed in generalizing the findings to other contexts. Historically, cocaine use has been less common in Sweden than in the United States and elsewhere in Europe, and cocaine has primarily been a club drug used by people under the age of 40 [37].

There are few specific studies on the association between parental AUD and cocaine use disorder in the next generation. In line with the current findings, however, a 1990s study of more than 400 male adolescents incarcerated in a large U.S. city suggested a relationship between parental alcohol problems and cocaine use disorder in the participants [38].

This cohort study, like others, provides information about risk, but causality cannot be inferred. The strengths and limitations common to national register studies are also found in the current study. Individual-level, linkable information was available on the national population, including health outcomes and sociodemographic factors. Additionally, the accuracy of the diagnoses in the National Patient Register is generally high [39, 40]. Inpatient care coverage is nearly complete, although data are missing for a few of Sweden’s counties for specific years in the mid-1980s [41].

As in other national register studies, however, the data were not collected specifically for the study, so some information of interest is missing. For example, although parental mortality data were available to the researchers, information on separation from parents for other reasons or data on the children’s living situations was not. A further limitation of the registers used in the study is that they do not include information from social services or self-reported data such as adverse childhood experiences or protective factors in the children’s lives. Additionally, data on children born outside Sweden were not available to the researchers.

Several limitations were specific to the use of inpatient data from the National Patient Register to define AUD and other psychiatric disorders. Most parental AUD identified in the register was probably on the severe end of the spectrum. This is the case even though both primary and secondary diagnoses of AUD-related disorders were used in the study, so AUD was not necessarily the main reason for every hospital visit that resulted in a diagnosis. Moreover, the researchers did not have access to data on psychiatric outpatient or primary care diagnoses, convictions, prescriptions, or participation in psychosocial treatment (e.g., 12-step programs). The study therefore underestimated the total number of people with parental AUD, and the results are less generalizable to and may overestimate SUD risk in the many people whose parents have mild or moderate AUD. Furthermore, certain characteristics affect the likelihood that a person with AUD will appear in different Swedish national registers, possibly because they affect behaviors such as care-seeking [42]. For this reason, women, younger people, and people with higher levels of education with AUD may have been underrepresented in the data. The results may be most relevant to places that have similar sociodemographic characteristics, patterns of substance use, and social and health care systems.

The study was exploratory, and the findings can be vulnerable to the multiple comparisons problem even though the data are on the whole population. Additionally, the findings estimate the average association between severe parental AUD and SUDs during the study period. Analyses of time trends over the study period were beyond the scope of this study.