Opioid use disorder mechanism

Opioids are natural or synthetic chemicals that reduce pain by interacting with opioid receptors on the cell surface. Opioids include the illicit drug heroin, lab-made opioids such as fentanyl and legally available painkillers such as morphine, oxycodone, hydrocodone and codeine (Ref. 12).

Opioids exert their effect by travelling through the blood to the brain and binding to opioid receptors on the cell surfaces of opiate-sensitive neurons (Ref. Reference Kosten and George13). There are three types of opioid receptors: the mu opioid receptor (MOR), the delta opioid receptor (DOR) and the kappa opioid receptor (KOR). Enkephalin and β-endorphin are the two main endogenous ligands for the MOR. Dynorphins and enkephalins preferentially bind to KOR and DOR, respectively (Refs Reference Minami, Hosoi, Toya, Katao, Maekawa, Katsumata, Yabuuchi, Onogi and Satoh14, Reference Reed, Butelman, Yuferov, Randesi and Kreek15). When opiates bind to receptors, the brain’s reward system is activated, resulting in feelings of pleasure while engaging in brain-promoting activities such as eating (Ref. Reference Haile, Kosten and Kosten16).

Opioids (morphine, codeine, opium) are frequently prescribed for the treatment of moderate to severe pain, which controls pain, relieves cough and diarrhoea and, due to prolonged use, also induces euphoria and stimulates the brain’s reward system (Ref. Reference Schuckit17). The patient continues to use opioids despite the development of serious health issues, which results in OUD. A patient with drug addiction, which is a type of SUD, may become dependent on higher doses of drugs to achieve the desired feeling of pleasure, which can result in fatal conditions such as respiratory distress.

Opioids simply stimulate the mesolimbic (midbrain) reward system. This system generates signals in the ventral tegmental area (VTA) that cause the nucleus accumbens to release the neurotransmitter dopamine. This dopamine stimulation underlies the effect on pleasure sensations. Opioids prevent dopamine cells from inhibiting dopamine release. This allows dopamine to enter the nucleus accumbens and prefrontal cortex of the forebrain, which plays a crucial role in drug dependence (Ref. Reference Kosten and George13). Moreover, drug addiction has been linked to a decrease in GABA communication between the VTA and the nucleus accumbens (Ref. Reference Reid and Lingford-Hughes18). GABA is an inhibitory amino acid neurotransmitter that modulates the release of dopamine, serotonin, noradrenaline and glutamate, which can affect neurotransmitter systems (Ref. Reference Fatemi, Folsom and Thuras19). These medications can cause the striatum to release more dopamine and glutamate while decreasing GABA production in the thalamus (Ref. Reference Mark, Soghomonian and Yamamoto20). Other parts of the brain create a long-term memory that links these positive emotions to their circumstances and surroundings. This frequently results in a drug craving (Ref. Reference Kosten and George13) (Figure 1).

Figure 1. Distribution of dopamine (Dopaminergic system) - Left, opioid receptors, GABA, and mechanism of OUD in the Human Brain – Right.

The diagram represents the key pathways of the brain involved in development of Euphoria induced by opioids and the development of OUD. Left: Dopaminergic neurons are represented in blue in the ventral tegmental area (VTA) and projects to the nucleus accumbens (NAc) in the mesolimbic reward pathway. Right: Opioids (e.g., heroin, morphine) bind to the mu-opioid receptors (MOR, depicted in red) on GABAergic interneurons in the VTA. This binding inhibits the release of GABA because the activaton of MOR causes the hyperpolarization of GABA neurons. GABA (depicted in green) is an inhibitory neurotransmitter that normally suppresses the firing of VTA dopamine neurons. By inhibiting GABAergic neurons, opioids disinhibit the dopaminergic cells The outcome is an enhanced release of dopamine in the nucleus accumbens (NAc) and the prefrontal cortex. The rise in dopamine levels (represented in blue) within these areas leads to the strong sensations of pleasure or euphoria linked to opioid consumption, reinforcing the behavior of taking drugs.To sum up, the illustration indicates that opioids promote increased dopamine release by inhibiting GABA’s inhibitory action – an essential process in the reward pathway. Important brain regions (VTA, NAc, prefrontal cortex) and neurotransmitters are identified in the diagram. (Note: To enhance clarity, the anatomical labels in the original figure have been magnified.)

Genetics and OUD

Many studies have examined the genetic causes of SUDs and have investigated the genes associated with OUD. However, over the past decade, the studies have concentrated on the single-nucleotide variants of genes reported to be associated with OUD. Numerous groups have hypothesised that addictions are complex disorders that may be influenced by the polymorphisms of many genes. Certain studies have also investigated whether a general set of genetic variables increases or decreases susceptibility to addiction (Ref. Reference Kreek, Nielsen and LaForge21).

There are diverse types of evidence to establish the genetic link between drug dependence and its aetiology. Family, adoption and twin studies were carried out in the early phase to examine substance dependence and abuse, supporting the theory that there is a substantial genetic component to drug addiction (Refs Reference Tsuang, Lyons, Meyer, Doyle, Eisen, Goldberg, True, Lin, Toomey and Eaves10, Reference Kendler, Karkowski, Neale and Prescott22). Further studies were carried out as candidate gene studies and GWAS (Ref. Reference Berrettini23). The earliest studies used to establish a genetic contribution to specific addictions focused on the genetic epidemiology of alcoholism (Ref. Reference Berrettini23). Recent epidemiological research has revealed both similarities and differences in the heritability of other addictions, especially OUD (Ref. Reference Nudmamud-Thanoi, Veerasakul and Thanoi24). On the basis of twin and family studies, heritability estimates for OUD risk range from 23% to 54% (Ref. Reference Tsuang, Lyons, Meyer, Doyle, Eisen, Goldberg, True, Lin, Toomey and Eaves10, Reference Kendler, Karkowski, Neale and Prescott22).

Addiction, particularly opioid addiction, is influenced by specific genes and single-nucleotide variants. Knowledge of this aspect of the study will improve understanding of the mechanism of drug addiction and the identification of vulnerable individuals.

Opioidergic system

There are three types of opioid receptors: mu-opioid receptor (MOR), kappa-opioid receptor (KOR) and delta-opioid receptor (DOR). The MOR is encoded by the OPRM1 gene, which has been intensively studied in heroin addiction and other opioid addictions, as well as in methadone and buprenorphine-based treatments for drug addiction. The OPRM1 variant rs1799971, which results in the Asn40Asp amino acid substitution of the MOR (Ref. Reference Haerian and Haerian25), is one of the most extensively researched variants in the field of addiction (Ref. Reference Bond, Laforge, Tian, Melia, Zhang, Borg, Gong, Schluger, Strong, Leal, Tischfield, Kreek and Yu26). The rs1799971 variant significantly reduces OPRM1 mRNA expression, which leads to low levels of MOR and high binding affinity of the endogenous ligand, β-endorphin (Refs Reference Kroslak, Laforge, Gianotti, Ho, Nielsen and Kreek27, Reference Zhang, Wang, Johnson, Papp and Sadée28).

The KOR systems (encoded by OPRK1) and their endogenous KOR agonist ligands, the dynorphins (encoded by PDYN), modulate addictive states, such as addiction to short-acting MOR agonists (Ref. Reference Reed, Butelman, Yuferov, Randesi and Kreek15). The DOR is encoded by the OPRD1 gene. Although the common opioid drugs that cause dependence do not primarily target the DOR, the receptor is involved in reward pathways of the neurobiology of drug addiction (Ref. Reference Mitchell, Margolis, Coker, Allen and Fields29).

GABA-ergic system

GABA (gamma-aminobutyric acid) is the principal inhibitory neurotransmitter in the brain. It acts through the ionotropic GABA-A and metabotropic GABA-B receptors (Ref. Reference Olsen and Sieghart30). The GABA-A receptor belongs to the ligand-gated ion channel receptor superfamily, with α1–6, β1–3, γ1–3, δ, ε, θ, π and ρ1–3 subunits. GABA-A receptors have primarily been linked to alcohol abuse (Ref. Reference Kumar, Porcu, Werner, Matthews, Diaz-Granados, Helfand and Morrow31). As there is a correlation between GABA-A receptor genes and the mechanism of OUD development, it is hypothesised that several GABA-A receptor genes may affect the susceptibility of individuals to drug (opioids) dependence. GABRA1, GABRA2, GABRA4 and GABRB1 are genes on chromosome 4p13–12 that encode GABA-A receptor subunits (Ref. Reference Hill, Shen, Zezza, Hoffman, Perlin and Allan32).

Preclinical research suggests that GABRA2 may play a role in drug addiction (Refs Reference Morris, Dawson, Reynolds, Atack, Rosahl and Stephens33, Reference Steffensen, Stobbs, Colago, Lee, Koob, Gallegos and Henriksen34). It is interesting to note that Enoch et al. (Ref. Reference Enoch, Hodgkinson, Yuan, Shen, Goldman and Roy35) assert that GABRA2 variants can predict addiction vulnerability, specifically for heroin addiction, and that childhood trauma and GABRA2 variants can influence addiction risk, specifically for cocaine dependence.

Dopaminergic system

D1, D2, D3, D4 and D5 are the five subtypes of dopamine receptors. The D1 and D5 dopamine receptors belong to the D1-like family of dopamine receptors, while the D2, D3 and D4 receptors belong to the D2-like family. D1-like family receptor activation is coupled to the G protein, which then activates adenylyl cyclase, whereas D2-like family receptor activation inhibits cAMP formation by inhibiting adenylyl cyclase. The rewarding effects of opioids are mediated by dopamine release and post-synaptic receptor activation. DRD2 is a gene that encodes the dopamine D2 receptor located near ANKK1, a gene that encodes a serine/threonine protein kinase. The DRD2/ANKK1 locus contains two widely studied polymorphisms: rs1800497, a missense variant in ANKK1 exon eight and rs1079597, located in DRD2 intron one. Rs1800497 and rs1079597 are often inherited together rather than by chance in non-African populations (Ref. 36). Both variants are reported among the Han Chinese population with opioid dependence (Ref. Reference Hou and Bin37).

Table 1 provides a brief overview of the genes and the variants studied among the varying ethnicities and their association with OUD. As highlighted in Table 1, various variants within the opioidergic system exhibit associations with OUD that are specific to certain populations. For example, the extensively researched OPRM1 variant rs1799971 (A118G) is more prevalent in non-African populations (Refs Reference Bond, Laforge, Tian, Melia, Zhang, Borg, Gong, Schluger, Strong, Leal, Tischfield, Kreek and Yu26, Reference Kumar, Chakraborty and Das39, Reference Nagaya, Ramanathan, Ravichandran and Navaratnam40, Reference Clarke, Crist, Kampman, Dackis, Pettinati, O’Brien, Oslin, Ferraro, Lohoff and Berrettini41) and has been associated with the risk of OUD across different cohorts, whereas rs1799972 shows greater effects in individuals of African descent (Refs Reference Bond, Laforge, Tian, Melia, Zhang, Borg, Gong, Schluger, Strong, Leal, Tischfield, Kreek and Yu26, Reference Crystal, Hamon, Randesi, Cook, Anastos, Lazar, Liu, Pearce, Golub, Valcour, Weber, Holman, Ho and Kreek38). Additionally, OPRD1 variants such as rs2236861 and rs2236857 have been consistently linked to heroin dependence among European and Israeli populations (Refs Reference Reed, Butelman, Yuferov, Randesi and Kreek15, Reference Beer, Erb, Pavlic, Ulmer, Giacomuzzi, Riemer and Oberacher43, Reference Levran, Londono, O’Hara, Nielsen, Peles, Rotrosen, Casadonte, Linzy, Randesi, Ott, Adelson and Kreek44).

Table 1. Genes and SNVs of the reward system of OUD which have demonstrated an association

Variants in the GABAergic pathway also display risk signals that are specific to ethnicity: GABRA2 rs11503014 has been linked to heroin addiction in African American individuals, while polymorphisms in GAD1 show significant associations within Han Chinese groups. Furthermore, GABRG2 rs211014 has been independently associated with OUD in both American and Chinese male populations, indicating shared biological mechanisms across different ancestries.

In the dopaminergic system, significant variants include DRD2/ANKK1 rs1800497 (Taq1A) and DRD2 rs1079597, both of which are related to OUD risk in European and Han Chinese populations. Other polymorphisms, like COMT rs4680 (Val158Met), demonstrate associations that vary by sex, particularly being linked to opiate addiction in Hispanic women.

Collectively, these findings highlight that while numerous risk loci are involved in different neurotransmitter systems, the robustness and reproducibility of these associations tend to differ based on ancestry, sex or environmental factors. This reinforces the necessity for replication efforts that are ancestrally diverse and adequately powered to authenticate candidate variants identified in smaller studies.

Epigenetics and OUD

Previous research has demonstrated a genetic link to OUD. The genetic role refers specifically to the association between OUD and particular genes, including their single-nucleotide variants. The other component of OUD causation is the influence of the individual’s environment. In this context, the interaction of genetics and environment can be described through epigenetic processes. The risk of drug abuse is associated with epigenetic mechanisms that explain some of the interactions between genes and the environment (Ref. 65). Therefore, epigenetics is the study of DNA-independent alterations in the regulation of gene activity and expression. These modifications can occasionally be inherited (Ref. Reference Vanyukov66).

Chronic substance use (repeated exposure) is partially linked to neuro-adaptive epigenetic changes in people’s brains that result in SUDs (Refs Reference Cadet67, Reference Robison and Nestler68, Reference Cadet and Jayanthi69). DNA methylation, chromatin remodeling, non-coding RNA and histone modifications all contribute to epigenetic events (Ref. Reference Robison and Nestler68). Even though all cells in the human body contain the same genetic information, epigenetic regulatory systems permit the development of several types of cells that, in response to the environment, form skin, heart, kidney, liver and nerve cells. These epigenetic marks can affect the health of offspring and even the expression of inherited traits. When a person uses heroin, for example, the drug can alter the DNA, increasing the production of addiction-related proteins. There is a correlation between elevated levels of these altered proteins and drug-seeking behaviour. There are studies widely carried out in vivo animal studies investigating these. Histones are similar to protein threads in that they provide an organisational framework for genes. The loosening or tightening of DNA around histones, exposing or concealing genes, in turn regulates gene expression. Drug exposure can modify gene expression in localised brain regions by affecting specific histones (Ref. Reference Maze, Covingtoni, Dietz, Laplant, Renthal, Russo, Mechanic, Mouzon, Neve, Haggarty, Ren, Sampath, Hurd, Greengard, Tarakhovsky, Schaefer and Nestler70). According to scientific evidence, manipulating enzymes that alter histones and proteins that bind to them may be an effective method for treating SUDs (Refs Reference Covington, Maze, Sun, Bomze, DeMaio, Wu, Dietz, Lobo, Ghose, Mouzon, Neve, Tamminga and Nestler71, Reference Demontis, Rajagopal, Thorgeirsson, Als, Grove, Leppälä, Gudbjartsson, Pallesen, Hjorthøj, Reginsson, Tyrfingsson, Runarsdottir, Qvist, Christensen, Bybjerg-Grauholm, Bækvad-Hansen, Huckins, Stahl, Timmermann, Agerbo, Hougaard, Werge, Mors, Mortensen, Nordentoft, Daly, Stefansson, Stefansson, Nyegaard and Børglum72).

Changes in DNA methylation have been observed after chronic exposure to multiple rewarding drugs, including opioids (Refs Reference Manzardo, Henkhaus and Butler73, Reference Jayanthi, Torres, Ladenheim and Cadet74, Reference Host, Dietrich, Carouge, Aunis and Zwiller75, Reference Hwang, Song, Kim, Choi, Guo, Law, Wei and Loh76). In particular, Hwang et al. (Ref. Reference Hwang, Song, Kim, Choi, Guo, Law, Wei and Loh76) found that increased DNA methylation of the promoter regions of the opioid receptor gene resulted in downregulation of the gene. It is consistent with previous research that MeCP2 can suppress the expression of the MOR gene (Refs Reference Manzardo, Henkhaus and Butler73, Reference Jayanthi, Torres, Ladenheim and Cadet74). Although evidence is limited, it has been suggested that the levels of DNA methylation vary between individuals of different racial groups who are exposed to opioids (Ref. Reference Jayanthi, Torres, Ladenheim and Cadet74). Opioid exposure is known to be linked to germline changes that can be carried down to future generations via transgenerational epigenetic inheritance (Refs Reference Byrnes, Johnson, Schenk and Byrnes77, Reference Sarkaki, Assaei, Motamedi, Badavi and Pajouhi78, Reference Cicero, Nock and O’Connor79). Transgenerational epigenetic inheritance refers to phenotypic variation that does not result from variations in DNA base sequences but is passed down through the germline to the progeny, even when no direct opioid exposure occurs. More extensive genetic and epigenetic studies will be required to compare individual and group racial differences in OUD susceptibility or resistance (Ref. Reference Vassoler, Byrnes and Pierce80). The previous studies have shown that genetic and epigenetic markers vary between ethnic groups (Ref. Reference Nielsen, Hamon, Yuferov, Jackson, Ho, Ott and Kreek81). Thus, analytical studies should focus more on determining the contribution of different epigenetic mechanisms to OUD development, which can be used for personalised treatments.

GWAS studies in the field

Large, global genome-wide association studies (GWAS) have started to identify the highly polygenic architecture of OUD and provide replicable risk loci. When broader meta-analytic datasets were included, cross-ancestry meta-analysis in the Million Veteran Program (MVP) with additional cohorts (such as the Psychiatric Genomics Consortium) revealed genome-wide significant signals at OPRM1 and a novel locus at FURIN, as well as other loci (e.g., RABEPK, FBXW4, NCAM1, KCNN1, TSNARE1) (Ref. Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90). Utilizing genetic correlations between OUD, alcohol use disorder and cannabis use disorder, complementary work employing multi-trait GWAS enhanced the power of discovery and produced approximately nineteen independent loci (Ref. Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91). When taken as a whole, these efforts put the current total of independent OUD risk loci across single-trait cross-ancestry and multi-trait frameworks at about 14–19, which is an order of magnitude higher than previous research (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91).

OPRM1 rs1799971 (A118G/Asn40Asp), a long-standing candidate variant, now has strong support because it was recaptured in subsequent meta-analyses (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91) and achieved genome-wide significance in MVP European-ancestry samples with replication in two independent cohorts (Ref. Reference Zhou, Rentsch, Cheng, Kember, Nunez, Sherva, Tate, Dao, Xu, Polimanti, Farrer, Justice, Kranzler and Gelernter92). The importance of very large samples and strict control for population structure is highlighted by the fact that, significantly, genome-wide signals still concentrate at OPRM1 (and FURIN) across the largest analyses to date, while a number of previously discussed candidates have not achieved genome-wide significance for OUD in single-trait analyses (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91).

These findings support the idea that OUD is extremely polygenic, meaning that meaningful prediction necessitates combining data from the entire genome, and that individual common variants have negligible effects. In line with an early-stage predictor for a complex, multifactorial trait, polygenic risk scores (PRS) derived from current OUD GWAS explain a modest fraction of variance (2.4–3.8% in case status across European-ancestry test sets) (Refs Reference Kember, Davis, Feuer and Kranzler93, Reference Na, Deak, Kranzler, Pietrzak and Gelernter94). PRS performance should increase as discovery GWAS expands and diversifies; however, OUD PRS currently offers little additional predictive value beyond conventional clinical and environmental factors and is not clinically actionable (Ref. Reference Kember, Davis, Feuer and Kranzler93).

Due to allele frequency and Linkage Disequilibrium (LD) differences, PRS accuracy is crucially dependent on ancestry match between the discovery and target samples; scores trained primarily in European cohorts perform worse in other populations, underscoring the urgent need for ancestrally diverse OUD GWAS (Refs Reference Martin, Kanai, Kamatani, Okada, Neale and Daly95, Reference Ding, Hou, Xu, Pimplaskar, Petter, Boulier, Privé, Vilhjálmsson, Olde Loohuis and Pasaniuc96). It is apparent that GWAS consistently demonstrates genetic associations between OUD and other substance use and mental health conditions, indicating a shared polygenic liability that may eventually inform individualised, integrated prevention and treatment strategies (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Holen, Shadrin, Icick, Filiz, Hindley, Rødevand, O’Connell, Hagen, Frei, Bahrami, Cheng, Parker, Tesfaye, Jahołkowski, Karadag, Dale, Djurovic, Smeland and Andreassen97).

Limitations related to available studies

According to Vereczkei et al. (Ref. Reference Vereczkei, Demetrovics, Szekely, Sarkozy, Antal, Szilagyi, Sasvari-Szekely and Barta61), Hou & Li (Ref. Reference Hou and Bin37), Levran et al. (Ref. Reference Levran, Londono, O’Hara, Nielsen, Peles, Rotrosen, Casadonte, Linzy, Randesi, Ott, Adelson and Kreek44), Zhang et al. (Ref. Reference Zhang, Kranzler, Yang, Luo and Gelernter45) and Beer et al. (Ref. Reference Beer, Erb, Pavlic, Ulmer, Giacomuzzi, Riemer and Oberacher43), several early candidate–gene reports (including studies of OPRD1 and DRD2) used cohorts of approximately 300–600 cases (with comparable controls). These studies were underpowered, which helps explain inconsistent replication across cohorts and ancestries because OUD risk variants exert small effects (OR ≈ 1.05–1.15). This picture has been clarified by recent large-scale GWAS (PGC/MVP and others), which have confirmed signals such as OPRM1 while finding modest support for a number of previously suggested candidate loci when tested at sufficient scale with strict stratification control (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91, Reference Zhou, Rentsch, Cheng, Kember, Nunez, Sherva, Tate, Dao, Xu, Polimanti, Farrer, Justice, Kranzler and Gelernter92, Reference Holen, Shadrin, Icick, Filiz, Hindley, Rødevand, O’Connell, Hagen, Frei, Bahrami, Cheng, Parker, Tesfaye, Jahołkowski, Karadag, Dale, Djurovic, Smeland and Andreassen97).

A persistent problem in OUD genetics has been the lack of replication: in the largest GWAS/meta-analyses to date, genome-wide significant signals concentrate at OPRM1 (and FURIN), while previously suggested candidates like OPRD1 and DRD2/ANKK1 failed to achieve genome-wide significance for OUD in single-trait analyses (Refs Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90, Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91, Reference Zhou, Rentsch, Cheng, Kember, Nunez, Sherva, Tate, Dao, Xu, Polimanti, Farrer, Justice, Kranzler and Gelernter92). For instance, OPRD1 and DRD2/ANKK1 were not included in the genome-wide significant set of a cross-ancestry Million Veteran Program meta-analysis (N ≈ 425,944), which revealed genome-wide loci at OPRM1, FURIN, RABEPK, FBXW4, NCAM1, KCNN1 and TSNARE1 (Ref. Reference Kember, Vickers-Smith, Xu, Toikumo, Niarchou, Zhou, Hartwell, Crist, Rentsch, Davis, Justice, Sanchez-Roige, Kampman, Gelernter and Kranzler90). OPRM1 and FURIN were also found to be the main single-variant signals in a complementary meta-analysis (total N ≈ 639,063; 20,686 OUD cases) (Ref. Reference Deak, Zhou, Galimberti, Levey, Wendt, Sanchez-Roige, Hatoum, Johnson, Nunez, Demontis, Børglum, Rajagopal, Jennings, Kember, Justice, Edenberg, Agrawal, Polimanti, Kranzler and Gelernter91).

This calls into question the accuracy of those early reports and emphasises the necessity of interpreting them with caution (Refs Reference Ioannidis98, Reference Button, Ioannidis, Mokrysz, Nosek, Flint, Robinson and Munafò99, Reference Border, Johnson, Evans, Smolen, Berley, Sullivan and Keller100, Reference Crist, Reiner and Berrettini101). In the past, false-positive associations could have been caused by potential confounding factors such as population stratification (differences in ancestry between cases and controls) (Refs Reference Price, Patterson, Plenge, Weinblatt, Shadick and Reich102, Reference Devlin and Roeder103, Reference Li, Reilly, Rader and Wang104). Future research must address this by emphasizing independent replication, pre-registering analyses when feasible and rigorously controlling population structure and other covariates (Refs Reference Price, Patterson, Plenge, Weinblatt, Shadick and Reich102, 105, Reference Kraft, Zeggini and Ioannidis106). The significance of large, ancestrally diverse GWAS and replication is further supported by the fact that, despite the large number of candidate–gene associations that have been proposed for OUD, many of them suffer from small sample sizes, lack of replication across cohorts and potential stratification biases (Refs Reference Crist, Reiner and Berrettini101, Reference Levran, Randesi, Adelson and Kreek107).

These insights further suggest that while there is significant progress in this field of study, there is a requirement for more robust evidence.