Weighted Family History Density of Substance Use: Influence on Participant Substance Use Onset, Duration, and Escalation

Introduction

Substance use disorders (SUDs) pose significant public health concerns, affecting millions worldwide and imposing substantial individual, societal, and economic burdens. Defined by physiological, behavioral, and cognitive symptoms resulting from ongoing substance use despite substance-related problems, distress, and/or impairment,¹ SUDs impacted 16.5% of Americans aged 12+ (46.3 million individuals) in 2021.² SUDs contribute extensively to the global disease burden, with 28.2% of nonelderly United States (US) adults experiencing alcohol or drug use disorders in their lifetime.³ The annual economic impact exceeds $400 billion,⁴ highlighting the urgent need to identify risk factors for effective prevention and intervention.

While genetics, environment, and social factors contribute to SUD vulnerability (Figure 1),^5–9 family history (FH), which refers to the presence of SUDs in an individual’s biological relatives, is a potent risk factor, with positive FH individuals exhibiting a two- to eight-fold higher risk of developing a SUD themselves.¹⁰ The influence of familial SUDs is striking: individuals with alcoholism in second/third-degree relatives have a 45% increased risk for alcohol dependence, rising to 86% with alcoholism in first-degree relatives, and 167% if both categories are affected.¹¹ Similar patterns have been observed for other SUDs.^8,12,13

Figure 1 Influence of WFHD on substance use risk factors and outcomes.

The heightened risk is well-documented,^9,14–17 with the impact varying based on participant sex, race/ethnicity, and the density of familial SUDs,¹⁸ as well as by specific substance types.^14,19 Children of high family history density (FHD) backgrounds (4+ first- and second-degree relatives) exhibit an earlier onset of regular drinking,²⁰ are more likely to develop substance abuse/dependence, and tend to do so at an earlier age¹⁷ compared to children of low-FHD backgrounds. These trends extend to other substances, like cannabis, where higher FHD predicts greater consumption and prolonged use.²¹ Moreover, greater FHD has been linked to a faster escalation to peak use—highlighting its potential as an early indicator of heightened risk.²²

Beyond general risk elevation, the familial transmission of SUDs may follow substance-specific pathways. Relatives of individuals with specific SUDs, such as opioid use disorder, are more likely to develop the same disorder, indicating specificity in inheritance.¹⁹ Hill et al²³ demonstrated this substance-specific pattern by showing that alcoholism and opiate abuse each clustered significantly within families but transmitted independently from each other. Even when individuals had both disorders, their relatives’ substance use patterns aligned with whichever disorder developed first in the proband, supporting the notion that different substances may involve distinct familial pathways.

Meller et al²⁴ further supported this pattern of substance-specific familial clustering, reporting that alcoholism and opiate addiction tended to aggregate independently within families, with concordance most apparent among same-substance-using male siblings. Notably, the presence of both disorders in a proband did not correspond to increased cross-substance transmission in relatives, reinforcing the conclusion that the familial transmission of alcoholism and opiate addiction reflects largely distinct pathways. These findings directly inform our examination of generational transmission of substance preferences, where we expect family alcohol use problems (AUPs) to be more strongly associated with participant alcohol use disorder (AUD) than drug use disorder (DUD), while family drug use problems (DUPs) will show stronger associations with participant DUD than AUD.

Despite these insights, much of this research is outdated, focusing on specific substances, primarily white male participants, and immediate family members, while relying on simplistic “positive/negative” FH measures. Critically, traditional binary measures fail to distinguish between minimal family history, such as having a single distant uncle affected, and extensive family history, exemplified by having both parents, all siblings, all grandparents, and multiple maternal and paternal aunts and uncles affected. By employing weighted family history density (WFHD), our approach captures these significant differences in familial risk exposure with greater precision. This study addresses these limitations by leveraging the National Epidemiological Survey on Alcohol and Related Conditions-III (NESARC-III) dataset, a large, nationally representative sample enabling a more in-depth analysis of familial influences on SUDs. We move beyond dichotomous measures by employing measures of WFHD, which has proven more robust in both sexes, enhancing diagnostic accuracy over simpler FH classifications.¹⁸

Building on prior research, we examine the impact of WFHD of substance use problems (SUPs) on three key outcomes: age at first use, escalation to peak use, and duration of use, across substances like alcohol, nicotine, and drugs. While age at onset has been commonly examined in familial risk studies, our inclusion of escalation and duration represents a more detailed approach to substance use trajectories, capturing clinically meaningful patterns of progression and chronicity that have received less attention in large-scale epidemiological research. We anticipate that higher WFHD will be associated with an earlier onset, a shorter escalation period, and a longer duration of use, including onset before age 18 and durations exceeding half of one’s current age, less than 5 years, and more than 10 years.

We also investigate sex differences in familial transmission patterns, expecting correlations between affected family members and participants to vary by sex. Although meta-analytic findings by Verhulst et al¹⁶ indicate that the overall magnitude of genetic influences on AUD is similar across sexes, other studies suggest sex-specific transmission effects. For instance, a Swedish national study on the transmission of alcohol use disorder across three generations found that familial risk is substantially stronger in same-sex pairs compared to opposite-sex pairs.²⁵ Similarly, Prescott et al²⁶ reported that same-sex twins exhibit considerably higher concordance for AUD than opposite-sex twins, suggesting that genetic and environmental liabilities are more consistently shared within each sex. Together, these findings imply that while both sexes possess a significant genetic risk for AUD, the manner in which this risk is transmitted within families differs by sex, supporting our hypothesis that familial correlations will vary accordingly.

Methods

Study Design and Population

At the time of study conception, NESARC-III was the most comprehensive dataset available for investigating familial patterns of substance use, offering an unmatched breadth and depth of family history variables—including parents, siblings, aunts, uncles, and grandparents. The survey includes a nationally representative sample of 36,309 noninstitutionalized adult civilians in the US, including those in select group quarters. Data for this cross-sectional analysis were collected between April 2012 and June 2013.²⁷ NESARC-III applied random sampling, selecting counties or county groups (primary), Census-defined blocks (secondary), and households (tertiary).²⁷ To enhance representation, the study prioritized selecting two respondents from households with ≥4 minority individuals and oversampled Hispanic, Black, and Asian participants. Imputation addressed missing or inconsistent data for variables like age, marital status, and income, with adjustments needed for <1% of cases during weighting.²⁸ The final sample included 36,309 individuals, achieving an overall response rate of 60.1%, comparable to other recent US national surveys.²⁷

Ethics Approval and Data Access

This study involved secondary analysis of de-identified, limited-access data from the National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), accessed through a Data Use Agreement with the National Institute on Alcohol Abuse and Alcoholism (NIAAA). All participants in the original NESARC-III study provided informed consent under protocols approved by NIAAA.

The data were handled per federal human subjects and privacy protections, including the Privacy Act of 1974 (5 U.S.C. §552a), the Health Insurance Portability and Accountability Act of 1996 (HIPAA) or equivalent privacy regulations, 45 C.F.R. Part 46, 21 C.F.R. Parts 50 and 56, and FDA Good Clinical Practice Guidelines (ICH E6, 62 FR 25692). Protections were further supported by a Certificate of Confidentiality issued by the NIH per 42 U.S.C. 241(d) of the Public Health Service Act.

Data were maintained securely with access limited to approved research personnel, and no attempts were made to identify individual participants. The University of Kentucky Institutional Review Board reviewed this project and determined that it was exempt from further review (IRB Protocol #94948), based on the use of fully de-identified, secondary data. All analyses complied with the confidentiality and security terms outlined in the NIAAA Data Use Agreement.

Primary Exposure of Interest: WFHD of SUPs

Self-reported FH of SUPs, operationalized as WFHD, was derived from survey questions assessing SUPs among first- and second-degree relatives. In this study, first-degree relatives are defined as biological parents and siblings, while second-degree relatives are defined as biological grandparents, aunts, and uncles. Family members who were under 10 years old or who passed away before reaching age 10, as well as participants’ children and adoptive parents, were excluded; however, family members over 10 years old were included even if deceased at the time of the interview. For siblings, aunts, and uncles, the survey also quantified the number of affected relatives for a more accurate assessment of WFHD.

The survey inquired about lifetime SUPs for each family member, with separate sets of questions for drug- and alcohol-related problems. Although the survey did not assess formal SUD diagnoses or the specific drugs used, the questions provided clear definitions of what constituted a “problem” with drugs or alcohol by referencing clinical criteria from the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), such as physical or emotional difficulties, relationship conflicts, work or school problems, driving under the influence, and significant time spent using or recovering from substances.^29,30

In contrast, NESARC-III participants were diagnosed with SUDs based on the Alcohol Use Disorder and Associated Disabilities Interview Schedule-5 (AUDADIS-5),³¹ a structured, computer-assisted interview administered by trained lay interviewers. The AUDADIS-5 systematically assessed alcohol and drug use behaviors and experiences across multiple timeframes (past-year, prior-to-past-year, and lifetime), using built-in skip patterns and DSM-5-based criteria. For the present study, we used the lifetime SUD diagnosis variable, which indicates whether a participant met the diagnostic criteria—defined as endorsing two or more symptoms within a single 12-month period—at any point in their life. Diagnostic variables in the NESARC-III dataset were generated using validated scoring algorithms developed by the NESARC-III team,³² mapped individual item responses to DSM-5 criteria and are publicly documented to support transparency and reproducibility. Thus, participant diagnoses in the current study reflect formal, criteria-based determinations aligned with DSM-5 standards.

WFHD was calculated using a method inspired by Zucker, Ellis and Fitzgerald’s³³ family expression of alcoholism formula. Due to limitations of the NESARC-III dataset, adjustments for affected-to-total family ratios were not possible. Instead, WFHD was calculated by assigning weights of 0.5 to affected first-degree relatives and 0.25 to affected second-degree relatives, yielding values from 0 to 18.25. This method reflects genetic proximity and its influence on familial substance use patterns.

Where Substance Use Problems include both Drug Use and Alcohol Use Problems.

Primary Outcomes of Interest: Age at Onset and Duration of Substance Use

Our primary outcomes were (1) age at onset and (2) duration of substance use. Age at onset was calculated via self-reported responses to questions regarding participants’ age at first use of any substance, including sedatives/tranquilizers, painkillers, cannabis, cocaine/crack, stimulants, club drugs, hallucinogens, inhalants/solvents, heroin, other drugs, cigarettes, cigars, pipes, snuff/chewing tobacco, e-cigarettes, and alcohol.

The duration of substance use was calculated in two steps. First, the age at most recent use was identified using variables measuring time since last use across substance categories. The time scale varied: alcohol was measured in days or months, nicotine in hours, and other drugs in days. The total duration of any substance use was then calculated by subtracting the earliest age at onset from the age at the most recent substance use.

Separate analyses were conducted for alcohol, nicotine, and other drugs. Three sets of models were run: (1) WFHD of any substance use problems (SUPs) predicting age at onset and duration of alcohol, drug, and nicotine use; (2) WFHD of alcohol use problems (AUPs) predicting these same outcomes; and (3) WFHD of drug use problems (DUPs) predicting these outcomes. Each outcome—alcohol, drug, and nicotine age at onset and duration—was modeled separately.

While NESARC-III also included questions about the age at which problematic use became more regular, we focused on age at first use because it reflects the earliest point of exposure—an important marker of risk. Numerous studies have demonstrated that early initiation is associated with increased vulnerability to substance use problems and the later development of substance use disorders, regardless of whether use continued immediately thereafter.^34–36

To further explore whether these associations differed among individuals with and without a history of SUD, we re-ran all models restricting the sample to participants with a lifetime SUD diagnosis. Additionally, we tested interactions between WFHD and lifetime SUD status to formally assess whether associations between WFHD and substance use outcomes differed based on whether a participant had ever met criteria for a SUD (n = 15,104).

Secondary Analyses of Interest: Escalation of Substance Use, Correlation Between Affected Family Members and Participants, and Concordance of Substance Preferences

To assess whether higher WFHD correlates with a shorter escalation period to peak alcohol use, we calculated an “escalation period” variable, defined as the number of years between age at first alcohol use and the age participants reported as the start of their heaviest drinking period. We then examined correlations between participants’ DSM-5 diagnosed SUDs and SUPs among parents, siblings, and extended relatives to identify which affected family members had the strongest associations with affected participants, while also considering potential sex-specific patterns in familial SUD transmission. Lastly, we investigated how participants’ preferences for alcohol and drugs aligned with those of their family members, identifying familial trends in substance choice to provide insights into how substance use behaviors may be influenced or inherited.

Statistical Analysis

To explore the relationship between WFHD and various sociodemographic and health characteristics, we categorized participants into tertiles based on their WFHD scores, with an additional “none” group for those without a FH of SUPs. We analyzed these groups on a range of sociodemographic variables and displayed the total sample size, group sizes, and the percentage distribution of each characteristic within each group (Table 1). Specifically, we compared sex, age, race/ethnicity, sexual orientation, religious affiliation, educational attainment, income, rurality, and Census region across the WFHD groups to identify significant patterns associated with differing levels of familial SUPs. We conducted Chi-square tests to compare distributions of categorical variables across WFHD tertiles, ANOVA tests for normally distributed continuous variables, and Kruskal–Wallis tests for continuous variables that were not normally distributed.

Table 1 Descriptive Statistics Comparing NESARC-III Participant Sociodemographic and Health Characteristics Between WFHD Categories

Linear regression models examined relationships between WFHD, age at onset, and duration of substance use, controlling for potential confounding variables. Logistic regression tested WFHD’s relationship with dichotomous outcomes, including early onset (<18 years) and long-term use (>5 years, >10 years, over half of current age). A Kaplan-Meier survival analysis assessed escalation periods (time from first alcohol use to peak use). WFHD categories (None, Tertiles 1–3) were compared using survival curves and Log rank tests to evaluate whether higher WFHD predicted shorter escalation periods.

Pearson correlations (r) between affected family members and participants, stratified by participant sex, were examined to identify potential sex-specific or intergenerational patterns in SUD inheritance. Logistic regression analyses assessed the association between WFHD of alcohol and drug use problems and the likelihood of AUD and DUD development in participants. Separate models were used to assess the impact of alcohol-related WFHD on AUDs and DUDs and drug-related WFHD on AUDs and DUDs.

In additional analyses, we repeated all regression models among the subset of participants with a lifetime substance use disorder diagnosis to test whether the strength or direction of associations differed in this high-risk group. These models mirrored the structure of the primary analyses and included unadjusted and fully adjusted versions. We then conducted formal interaction analyses to test whether the relationships between WFHD and each outcome (onset and duration) significantly differed between individuals with and without a lifetime SUD diagnosis. Interaction terms between WFHD and lifetime SUD status were added to each model, and interaction p-values were evaluated for significance. Finally, we conducted logistic regression models using both the full sample and the SUD-only subset for all dichotomized outcomes (eg, onset before age 18, duration longer than 5 or 10 years, and duration exceeding half of current age) and compared odds ratios across four tiers of model adjustment (unadjusted, adjusted for demographics, adjusted for socioeconomic status, and adjusted for geography). This allowed us to assess whether restricting to individuals with a SUD history meaningfully altered model estimates or improved prediction of key substance use outcomes.

Two-tailed p-values < 0.05 were considered statistically significant. All analyses were conducted using R Studio Version 4.3.1 and were weighted to account for the complex survey design and probability of non-response in NESARC-III. While this analysis offers insights into familial patterns and substance use, it was not pre-registered and should be considered exploratory.

Results

Description of Study Participants

Table 1 compares the sociodemographic and health characteristics of NESARC-III participants across different categories of WFHD, showing significant differences across all variables (p < 0.001). Males constituted 43.7% of the total sample, with male representation decreasing as WFHD increased. Younger individuals (18–29) were more prevalent in higher WFHD tertiles. Racial/ethnic distribution indicated a higher proportion of non-Hispanic Whites in lower WFHD categories, while higher tertiles contained more Hispanic participants. Sexual orientation, religion, education, income, rurality, and census region also varied significantly across WFHD categories, with higher WFHD tertiles generally associated with lower income and educational levels, and more urban residence.

Impact of WFHD of SUPs on Age at Onset and Duration of Substance Use Across Alcohol and Drug Use

Table 2 presents regression results examining the association between WFHD and the onset and duration of substance use, revealing consistent and significant associations. In unadjusted models, each unit increase in WFHD was linked to an earlier onset of substance use by 0.54 to 0.82 years (p < 0.001), with the strongest effect for those with a FH of AUPs (β = −0.82, 95% CI: −0.89, −0.76). Similarly, WFHD was associated with a longer duration of substance use, extending it by 0.48 to 0.79 years (p < 0.001), with FH of AUPs again showing the strongest association (β = 0.79, 95% CI: 0.69, 0.89). These effects remained robust after adjusting for confounds, with each unit increase in WFHD continuing to predict an earlier onset by 0.53 to 0.84 years and a longer duration by 0.55 to 0.90 years, particularly for AUDs (onset: β = −0.84, 95% CI: −0.91, −0.78; duration: β = 0.90, 95% CI: 0.80, 1.00). The adjusted models demonstrated improved explanatory power, particularly for age of onset (R² = 0.077 to 0.088) and duration (R² = 0.755 to 0.756), highlighting the strong predictive influence of WFHD on both the initiation and persistence of substance use across different substance types.

Table 2 Impact of WFHD on Substance Use Onset and Duration: Unadjusted and Adjusted Regression Models

Detailed Analysis of the Associations Between WFHD and Age at Onset and Duration Across Adjusted Models

Table 3 displays the impact of WFHD on substance use onset and duration, demonstrating consistent associations across various levels of model adjustment. For onset before age 18, each one-unit increase in WFHD corresponded to a 26% increase in odds in the unadjusted model (OR = 1.26, 95% CI: 1.24–1.28, p < 0.001), rising to 27% in the fully adjusted model (OR = 1.27, 95% CI: 1.25–1.29, p < 0.001), with model fit improving from a McFadden’s R² of 0.019 to 0.059.

Table 3 Effect of WFHD on Substance Use Onset and Duration: Odds Ratios and Model Fit Statistics by Sequentially Adjusted Models

For substance use duration exceeding half of one’s current age, a one-unit increase in WFHD raised the odds by 17% in the unadjusted model (OR = 1.17, 95% CI: 1.15–1.19, p < 0.001), reaching 19% in the fully adjusted model (OR = 1.19, 95% CI: 1.17–1.21, p < 0.001), with R² improving from 0.27 to 0.288. For substance use lasting over five years, the odds increased by 22% in the unadjusted model (OR = 1.22, 95% CI: 1.18–1.26, p < 0.001), and up to 26% in the fully adjusted model (OR = 1.26, 95% CI: 1.22–1.30, p < 0.001) for each unit increase in WFHD, with R² improving from 0.267 to 0.293. For duration exceeding ten years, a one-unit increase in WFHD raised the odds by 18% in the unadjusted model (OR = 1.18, 95% CI: 1.15–1.21, p < 0.001), and by 21% in the fully adjusted model (OR = 1.21, 95% CI: 1.18–1.24, p < 0.001), with an enhanced model fit from R² = 0.387 to R² = 0.402, explaining about 40% of the variance in substance use duration. Results remained consistent when analyses were restricted to participants with a lifetime SUD diagnosis, with nearly identical effect sizes and significance levels across models and only two of 18 interaction terms reaching significance, underscoring the robustness of WFHD as a predictor of earlier onset and longer substance use duration regardless of clinical history (Supplemental Tables 1 and 2).

Impact of WFHD Type on Age at Onset and Duration of Substance Use Across Substance Types

Increases in WFHD of SUPs were linked to earlier onset and longer duration of alcohol, drug, and nicotine use. As shown in Table 4 Model 1 (overall SUPs), higher WFHD predicted significantly earlier onset (by 0.30–0.47 years) and longer duration (by 0.31–0.73 years) across all substances (p < 0.001). Table 5 Model 2 (alcohol use problems) showed the strongest effect on drug use (onset: −0.79 years; duration: +1.06 years), whereas Table 6 Model 3 (drug use problems) most strongly predicted drug use outcomes (onset: −0.54 years; duration: +1.01 years). These findings suggest that WFHD of alcohol use problems had a broader cross-substance impact—particularly on drug use—while WFHD of drug use problems demonstrated greater substance specificity. Notably, this initial analysis included all participants with available family history data—not only participants with SUDs—emphasizing the predictive value of familial SUPs across the full sample. However, this unexpected cross-substance association persisted even when restricting the sample to individuals with a lifetime SUD diagnosis (Supplemental Table 3, Models 1–3). Sample sizes for each model varied slightly depending on outcome availability and inclusion criteria, as detailed in Table 7.

Table 4 Model 1: Adjusted Analysis of WFHD of All Substance Use Problems Predicting Participant Age of Onset and Duration of Alcohol, Drug, and Nicotine Use

Table 5 Model 2: Adjusted Analysis of WFHD of Alcohol Use Problems Predicting Participant Age of Onset and Duration of Alcohol, Drug, and Nicotine Use

Table 6 Model 3: Adjusted Analysis of WFHD of Drug Use Problems Predicting Participant Age of Onset and Duration of Alcohol, Drug, and Nicotine Use

Table 7 Participant Sample Sizes per Analysis

Analysis of Alcohol Use Escalation

The Kaplan-Meier survival analysis (Figure 2) suggested that participants with higher WFHD of AUPs tended to reach peak alcohol use more quickly compared to those with lower WFHD. This observation highlights the potential influence of familial AUPs on the trajectory of participant alcohol consumption, with greater WFHD contributing to faster escalation to peak use.

Figure 2 Kaplan-Meier survival probability of peak alcohol use by WFHD level of AUP.

Differential Impact of WFHD on Alcohol and Drug Use Disorders

For alcohol-related WFHD, each unit increase corresponded to a 56.8% increase in the odds of AUD diagnosis (OR = 1.568, 95% CI: 1.530–1.607) and a 56.4% increase in the odds of DUD diagnosis (OR = 1.564, CI: 1.526–1.603). Similarly, for drug-related WFHD, each unit increase was associated with a 41.8% increase in the odds of DUD (OR = 1.418, CI: 1.383–1.455) and a 39.5% increase in the odds of AUD (OR = 1.395, CI: 1.360–1.430).

Sex-Specific Patterns of Inheritance

A heatmap analysis revealed sex-specific correlations between SUP prevalence in family members and SUDs in male and female participants (Figure 3). For males, the strongest correlations were linked to fathers (r = 0.195), followed by maternal grandfathers (r = 0.155), and mothers (r = 0.147), suggesting a greater paternal role in SUD predisposition. Conversely, for females, the correlations with mothers (r = 0.203) and fathers (r = 0.201) were nearly identical, indicating similar maternal and paternal influences. However, the mother-daughter correlation (r = 0.203) was notably stronger than the mother-son correlation (r = 0.147), highlighting a more substantial maternal impact on female participants.

Figure 3 Heatmap of sex-specific correlations between affected family AUPs and participant SUDs. Note: Pearson’s r was used to compute correlation coefficients.

Discussion

Our study highlights the strong association between WFHD and substance use behaviors, influencing when individuals start using substances, how quickly they escalate to peak use, and how long they continue. Higher WFHD was associated with earlier onset and longer duration of substance use across alcohol, drug, and nicotine outcomes, even after adjusting for demographic, socioeconomic, and geographic characteristics. In addition, individuals with higher WFHD of alcohol use problems experienced shorter periods from first use to peak use, reflecting a faster progression of substance use among those with elevated familial risk. We also observed concordance in substance preferences within families, indicating that specific substance use patterns are often replicated across generations. Finally, distinct male and female inheritance patterns emerged, suggesting the need for sex-specific prevention and treatment approaches.

These associations remained robust in models restricted to individuals with a lifetime history of SUDs, with nearly identical effect sizes and directions of association, emphasizing the clinical relevance of these patterns among affected populations. All main findings—across onset and duration of alcohol, drug, and nicotine use—remained significant in the SUD-only models, with results consistent with those observed in the full sample. Interaction models further indicated that associations between WFHD and substance use outcomes were largely similar across participants with and without SUDs, with only two of eighteen interaction terms reaching statistical significance. This consistency supports the inclusion of all participants in our primary analyses. Models examining alcohol-specific and drug-specific WFHD revealed the same patterns, with alcohol-related WFHD predicting broader, cross-substance effects and drug-related history showing greater substance specificity. This consistency supports the inclusion of all participants in our primary analyses and emphasizes the value of assessing WFHD not only for early identification of risk but also for understanding severity, chronicity, and progression among those already experiencing substance use problems.

These results are consistent with a robust body of evidence identifying family history as a central factor in the development and progression of substance use problems.^{9,12,13,16–21,37} For example, Hill and Yuan²⁰ reported that adolescents from high-density alcohol-affected families initiated regular drinking at significantly younger ages than those from low-density families, while Hill et al¹⁷ found that children with elevated familial alcohol risk exhibited both earlier onset and higher risk of developing substance abuse problems. Similarly, Dawson et al¹¹ demonstrated a dose-response effect, showing that risk for alcohol dependence increased in proportion to the number of affected first- and second-degree relatives. Our findings extend this work by using a continuous, WFHD measure to predict not only onset but also escalation and duration, with associations persisting across substances even after adjusting for sociodemographic and geographic factors.

Further supporting familial specificity, our findings reveal that WFHD of AUPs more strongly predicted AUD than DUD, while WFHD of DUPs more strongly predicted DUD than AUD. This observation aligns with Bierut et al,¹⁴ who reported that siblings of alcohol-dependent individuals were more likely to develop alcohol dependence, while siblings of cannabis- and cocaine-dependent individuals were at higher risk for those specific dependencies. Similarly, Merikangas et al¹⁹ and Tyrfingsson et al³⁷ identified strong familial concordance for opioid, cannabis, alcohol, and stimulant use disorders, reinforcing the idea of substance-specific inheritance.

Despite this substance-specific diagnostic concordance, we observed that WFHD of AUPs was unexpectedly associated with earlier onset and longer duration of drug use compared to alcohol use. This finding suggests that familial AUPs represent a broader, transdiagnostic risk factor, potentially due to shared genetic and environmental vulnerabilities, cross-substance escalation, and social influences, with familial traits such as impulsivity, sensation seeking, or altered stress responses potentially heightening susceptibility to earlier onset and prolonged drug use even when the primary SUD risk remains alcohol-specific. These associations remained when the analysis was restricted to individuals with a lifetime history of SUD, with nearly identical direction and magnitude of effects, reinforcing the robustness and clinical relevance of these subtype-specific patterns. Taken together, the results suggest that WFHD is not only a meaningful marker of individual risk—but one that consistently shapes substance use trajectories across both general and affected populations. Supporting this, Grucza et al³⁸ observed significant associations between familial alcoholism risk and obesity, particularly among women, suggesting that familial AUDs may confer a transdiagnostic vulnerability—potentially through shared neurobehavioral pathways such as reward processing dysregulation and heightened impulsivity—with sex-specific biological or environmental factors amplifying this risk in women. Mennella et al³⁹ similarly reported heightened sweet preferences in children with a family history of alcoholism, further supporting the concept of shared reward-processing pathways underlying broader susceptibilities.

Neuroimaging research by Lees et al⁴⁰ also offers valuable insights into sex-specific neurobiological mechanisms underlying this broader vulnerability. They observed that children with a parental history of AUD exhibited sex-specific neural differences during response inhibition tasks; notably, all youth with paternal AUD history showed increased activation in the medial orbital frontal cortex during successful inhibition, while only female youth with maternal AUD history displayed heightened neural activation in the cerebellum during failed inhibition. These neural findings indicate distinct sex-specific neurodevelopmental vulnerabilities affecting cognitive control and impulse regulation, potentially explaining why familial AUP risk broadly influences susceptibility to drug use behaviors beyond alcohol alone, particularly regarding early onset and prolonged duration among women with affected mothers.

This interpretation aligns with our observed sex-specific patterns: Males showed stronger correlations with paternal substance use, whereas females exhibited stronger correlations than males with nearly all affected family members—especially with affected mothers (Figure 3). Christensen and Bilenberg⁴¹ noted that daughters are often more affected by parental alcoholism than sons, possibly due to greater emotional involvement and engagement with family matters.^37,42 Dawson et al¹¹ also found that women with multiple alcoholic relatives were significantly more likely than men to begin drinking before age 15, with elevated risk persisting into adulthood compared to men.

McGue et al⁴³ provide further insight into these sex-specific familial pathways, showing that early alcohol use is highly heritable in boys—primarily due to genetic liability for disinhibitory psychopathology—whereas in girls, early alcohol use is more strongly linked to shared environmental influences, particularly maternal drinking. Together, these findings suggest that while familial alcohol problems may confer substance-specific risk, their broader and more diffuse impact—particularly among women—may reflect environmentally mediated pathways that extend beyond alcohol use, underscoring the need for prevention efforts that account for both the substance type and the sex-specific nature of familial transmission.

Limitations

Our study contributes strong evidence that WFHD consistently influences substance use behaviors, influencing when individuals start using substances, how quickly they escalate to peak use, and how long they continue. We observed concordance in substance preferences within families, indicating that specific substance use patterns are often replicated across generations. Distinct male and female inheritance patterns further suggest the need for sex-specific prevention and treatment approaches.

While our study boasts a large sample size and comprehensive control of potential confounds, its cross-sectional design limits causal inferences. Additionally, reliance on self-reported survey data may introduce recall and social desirability biases, potentially affecting the accuracy of WFHD variables and reported substance use behaviors. These concerns may be especially relevant given the wide age range of the sample—approximately 20% of whom were aged 60 or older—raising the possibility that older participants may have less reliable recall or reduced contact with affected relatives. Conversely, individuals in the highest WFHD tertile were disproportionately younger (about 50% under age 40), which may reflect greater proximity to living, substance-affected relatives or enhanced recall, potentially introducing bias.

Although our findings strongly suggest that greater WFHD influences earlier onset, longer duration, and faster escalation to peak substance use, the retrospective design requires cautious interpretation of these associations. The absence of clinical severity and comorbidity data in our current analytic models also limits our ability to fully characterize these familial risk pathways. While the NESARC-III dataset includes DSM-5-based severity gradings and information on psychiatric comorbidities, our analytic focus was limited to lifetime SUD status to maintain consistency across substance categories and avoid overcomplicating model interpretations. Future research could build on these findings by examining how familial risk interacts with disorder severity, comorbidity, and diagnostic timing. Longitudinal data could also address these limitations by tracking substance use behaviors over time to establish causation and reduce recall bias. Investigating how WFHD interacts with environmental stressors, epigenetic changes, and genetic markers could further clarify substance use etiology.

Conclusion

Recognizing the importance of FH and inheritance patterns is essential as we refine approaches to substance use prevention and treatment. Incorporating FH assessments into prevention strategies can enable early identification of individuals with a high familial burden of SUDs, facilitating targeted efforts to delay onset, reduce duration, and address rapid escalation of substance use. Moreover, sustained interventions are critical, not only to address initiation, continuation, and escalation of substance use, but also to equip individuals in recovery with the skills to navigate relationships with family members who are actively using substances. Future causal longitudinal studies should extend these findings across diverse populations, examining genetic predispositions, environmental factors, and interactions by substance type and participant sex. By considering familial patterns and integrating broader variables, we can create more personalized prevention and treatment strategies, ultimately advancing public health and clinical solutions for managing SUDs.