Periodic Reporting for period 3 - OPIOIDREWARD (How distress alters opioid drug effects and abuse liability)
Période du rapport: 2022-07-01 au 2023-12-31
Why are opioids so addictive? Non-human animal research links addiction with the powerful relief opioids can offer to animals in distress. In humans, epidemiological and clinical studies converge upon social stressors and a poor social support network as key risk factors for addiction. Despite this, it is currently unknown how pre-drug distress might alter opioid drug effects. Tremendous resources are dedicated to charting how people feel after taking a drug, sidestepping the potentially profound influence of how people feel before they take the drug. OPIOIDREWARD turns the current approach on its head. Using acute social distress induction before morphine administration in healthy humans, we create a human model to determine the psychological, physiological and brain underpinnings of how social stressors increase opioids’ abuse liability.
First, we test the hypothesis that pre-drug distress enhances drug wanting (self-administration) but not drug liking (self-report) compared to drug effects in a control condition. Second, we use opioid blockade to confirm or falsify the hypothesis that opioid drugs ‘hijack’ brain mechanisms underpinning social support. Third, we test the link between pre-drug distress and drug self-administration in a real-world clinical setting. In sum, the project aims to achieve a breakthrough in our understanding of how a pre-drug social distress state can alter opioid drug mechanisms. The mechanistic understanding arising from this project could have profound implications for science, as well as for clinical care and new policies designed to contain the opioid epidemic.
First, we test the hypothesis that pre-drug distress enhances drug wanting (self-administration) but not drug liking (self-report) compared to drug effects in a control condition. Second, we use opioid blockade to confirm or falsify the hypothesis that opioid drugs ‘hijack’ brain mechanisms underpinning social support. Third, we test the link between pre-drug distress and drug self-administration in a real-world clinical setting. In sum, the project aims to achieve a breakthrough in our understanding of how a pre-drug social distress state can alter opioid drug mechanisms. The mechanistic understanding arising from this project could have profound implications for science, as well as for clinical care and new policies designed to contain the opioid epidemic.
During the first part of the project, we have established two novel stress induction paradigms optimised for Study 1 and Study 2 of OPIOIDREWARD, respectively. Both setups build on the commonly used Trier Social Stress Task (TSST) but contain several key improvements that enable us to test the hypotheses of Studies 1 and 2.
Study 1 setup and results:
For Study 1, we developed an hybrid stress setup with the participant present in the lab, while the stress induction panel interacts with the participants on zoom. The hybrid solution has two major benefits: i) it is compatible with infection control restrictions due to COVID-19, and ii) it allows for blinding (masking) of test personnel not only to the participant’s drug condition, but also to their stress condition (stress or no-stress control interaction). This blinding is not typically implemented in traditional TSST. Our setup ensures that all personnel interacting with participants outside of the stress/control interaction is fully blinded to stress condition, preventing any bias in the data related to experimenter expectation effects.
Moreover, we designed the setup to be suitable for repeated within-subjects stress induction. Specifically, we adapted a ‘singing TSST’ task for the second stress induction, since previous studies using singing challenges had reported little or no habituation of stress responses to this type of task. Due to the typically high within-subjects correlations in the main outcomes of interest such as subjective affective state (r=0.5-0.9) conducting Study 1 fully within-subjects affords a large boost in statistical power.
We also designed improved control tasks for both stress inductions. The commonly used ‘placebo TSST’ involves giving a speech about a recent event to an empty room, which is not always perceived as stress-free and notably fails to control for the occurrence of a social interaction. The control tasks in our setup include a (hybrid) social interaction in which participants complete colour and sound preference tasks (to match the job talk and singing task) as well as easy maths tasks to control for the mental arithmentic part of the stress inductions.
Pilot testing was initially conducted in N=13 using the traditional TSST setup outside of a drug context, then in 42 sessions with morphine/placebo, before it was redesigned and pilot tested in 30 sessions. Data collection for the study (final N=63, 252 sessions) was completed in April 2022. The stress and drug manipulations were successful and the main paper has been preregistered (https://osf.io/v8dxy/(s’ouvre dans une nouvelle fenêtre)) and submitted for publication. In brief, the results confirm the hypothesis that pre-drug stress enhances opioids’ abuse liability. However, we find evidence of a strong interaction with subject sex, such that stress enhances opioid drug self-administration in male but not in female participants. This finding points to a putative stress-related mechanism that could help explain the overrepresentation of men in addiction.
Clinical setting:
In addition, our team gained the opportunity to test the same hypothesis in a real-world, clinical setting. Drawing on data from a quality control study of 259 day surgery patients treated at Kongsberg Hospital in Norway between 2018 and 2021, we quantified subjective responses to opioid treatment on the operating table in the minutes before anaesthesia. The minutes before surgery are associated with substantial stress, which varies naturally between individuals. After calculating the main group effect of each drug (high doses of iv. remifentanil and oxycodone), we were therefore able to analyse the relationship between pre-surgical negative affect and opioid responses. As reported in Eikemo et al, 2023, Anaesthesia, pre-surgical negative affect was (weakly) associated with a more positive response to the medications. We did not find that pre-surgical negative affect or a positive medication response significantly predicted postsurgical opioid use however, based on phone interview data from 1-2 days after surgery and long-term follow-up data (>4 months after surgery), as reported in Meier et al, 2023, Regional Anaesthesia and Pain Medicine). Finally, we probed the relationship between early life stress (measured using the Childhood Trauma Questionnaire) and medication responses, as reported in a registered report by Carlyle et al, 2023 (Peer Community In and Addiction Research and Theory). Following up on the inconclusive results from this analysis of surgery patients, we were also able to conduct a secondary analysis of data collected in a series of psychopharmacology studies in the lab of collaborator Harriet de Wit. The results have been preprinted (Carlyle et al, 2023, https://osf.io/zvnse/(s’ouvre dans une nouvelle fenêtre)) and submitted for publication.
Study 2 setup and results:
For Study 2, we developed a novel, dyadic implementation of the TSST, the Dyadic Stress and Support Task (DSST). The DSST is optimised to measure the effects of social support on stress recovery, which will allow us in turn to determine the importance of endogenous opioids for support-induced stress relief. The DSST has two parts, stress induction and recovery. During stress induction, pairs of friends complete the TSST in parallel, separated by a physical barrier such that they can hear, but not see each other. During recovery, the two friends are left to talk freely with each other (social support condition). In the non-support control condition, each participant instead interacts with a panel member instructed to appear professional, but not supportive.
The dyadic setup has two major benefits. First, we test two participants at the same time, which reduces the number of test personnel needed for TSST sessions. Second, by recruiting pairs of long-standing, close friends, we ensure the ecological validity of the social support.
Pilot testing to validate the paradigm was conducted in N=42 (21 pairs of friends) randomised to social support or non-support after stress induction. Ratings of affective state, heart rate and salivary cortisol data support the validity and utility of the DSST stress induction. Social support caused a large boost in positive affect that was not observed during recovery in the control condition. The pilot data and a detailed description of the DSST has been written up and preprinted (Løseth et al, 2022, PsyArXiv link).
Data collection for the main study (final N=258) was completed autumn 2023. Data analysis is ongoing.
Scientifically, the project has so far contributed with several publications in scientific journals, as well as a number of talks and presentations at conferences, seminars and workshops. Of the results so far, the below publications can be highlighted (presented in roughly chronological order):
1. The manuscript presenting the DSST, as described above (Løseth et al, 2022, Psychoneuroendocrinology).
2. Our team has also published several peer-reviewed comments and review articles on the topics studied in OPIOIDREWARD. In Leknes & Atlas, 2020, British Journal of Anaesthesia, we present a checklist for psychopharmacology studies, highlighting and explaining the importance of placebo control and double-blinding; counterbalancing of conditions; choice of drug, dose and administration method; construct validity; study population and generalisability; and reproducibility and reporting. Eikemo, Løseth and Leknes (2021, Pain) review effects of opioid antagonists and present a guide for how to interpret results of full and partial blockade of mu-opioid receptors. Meier, Eikemo and Leknes (2021, Current Addiction Reports) review the psychopharmacological evidence on opioid modulation of reward and threat using a translational perspective. In brief, we find that many results in the human literature point to a preserved function of the endogenous opioid system across rodents and humans, however effect sizes are typically smaller. We conclude that the human endogenous opioid system modulates human subjective experience and related responses in concert with other neurotransmitter systems.
3. Our primer on opioid antagonism to understand endogenous opioid function and two open-access web apps to facilitate study planning (planoxone and plantrexone): Opioid antagonism in humans: a primer on optimal dose and timing for central mu-opioid receptor blockade (Trøstheim, M Eikemo, J Haaker, JJ Frost, S Leknes
Neuropsychopharmacology 48 (2), 299-307)
4. Three papers summarizing the link between pre-drug negative affect, positive medication responses and subsequent opioid self-administration in surgery patients (by Eikemo, Meier and Carlyle et al, details described above)
5. The main paper of the project has been preprinted and describes the test of the core hypothesis, namely that receiving opioids when in a stress state increases subsequent opioid self-administration (the gold-standard measure of drug abuse liability). Our results confirm the hypothesis, but the analysis clearly shows that the effect is driven by male participants (N=31) with no stress-enhanced opioid drug wanting in female participants (N=32). Eikemo et al, 2023: https://osf.io/preprints/v8dxy/(s’ouvre dans une nouvelle fenêtre))
Study 1 setup and results:
For Study 1, we developed an hybrid stress setup with the participant present in the lab, while the stress induction panel interacts with the participants on zoom. The hybrid solution has two major benefits: i) it is compatible with infection control restrictions due to COVID-19, and ii) it allows for blinding (masking) of test personnel not only to the participant’s drug condition, but also to their stress condition (stress or no-stress control interaction). This blinding is not typically implemented in traditional TSST. Our setup ensures that all personnel interacting with participants outside of the stress/control interaction is fully blinded to stress condition, preventing any bias in the data related to experimenter expectation effects.
Moreover, we designed the setup to be suitable for repeated within-subjects stress induction. Specifically, we adapted a ‘singing TSST’ task for the second stress induction, since previous studies using singing challenges had reported little or no habituation of stress responses to this type of task. Due to the typically high within-subjects correlations in the main outcomes of interest such as subjective affective state (r=0.5-0.9) conducting Study 1 fully within-subjects affords a large boost in statistical power.
We also designed improved control tasks for both stress inductions. The commonly used ‘placebo TSST’ involves giving a speech about a recent event to an empty room, which is not always perceived as stress-free and notably fails to control for the occurrence of a social interaction. The control tasks in our setup include a (hybrid) social interaction in which participants complete colour and sound preference tasks (to match the job talk and singing task) as well as easy maths tasks to control for the mental arithmentic part of the stress inductions.
Pilot testing was initially conducted in N=13 using the traditional TSST setup outside of a drug context, then in 42 sessions with morphine/placebo, before it was redesigned and pilot tested in 30 sessions. Data collection for the study (final N=63, 252 sessions) was completed in April 2022. The stress and drug manipulations were successful and the main paper has been preregistered (https://osf.io/v8dxy/(s’ouvre dans une nouvelle fenêtre)) and submitted for publication. In brief, the results confirm the hypothesis that pre-drug stress enhances opioids’ abuse liability. However, we find evidence of a strong interaction with subject sex, such that stress enhances opioid drug self-administration in male but not in female participants. This finding points to a putative stress-related mechanism that could help explain the overrepresentation of men in addiction.
Clinical setting:
In addition, our team gained the opportunity to test the same hypothesis in a real-world, clinical setting. Drawing on data from a quality control study of 259 day surgery patients treated at Kongsberg Hospital in Norway between 2018 and 2021, we quantified subjective responses to opioid treatment on the operating table in the minutes before anaesthesia. The minutes before surgery are associated with substantial stress, which varies naturally between individuals. After calculating the main group effect of each drug (high doses of iv. remifentanil and oxycodone), we were therefore able to analyse the relationship between pre-surgical negative affect and opioid responses. As reported in Eikemo et al, 2023, Anaesthesia, pre-surgical negative affect was (weakly) associated with a more positive response to the medications. We did not find that pre-surgical negative affect or a positive medication response significantly predicted postsurgical opioid use however, based on phone interview data from 1-2 days after surgery and long-term follow-up data (>4 months after surgery), as reported in Meier et al, 2023, Regional Anaesthesia and Pain Medicine). Finally, we probed the relationship between early life stress (measured using the Childhood Trauma Questionnaire) and medication responses, as reported in a registered report by Carlyle et al, 2023 (Peer Community In and Addiction Research and Theory). Following up on the inconclusive results from this analysis of surgery patients, we were also able to conduct a secondary analysis of data collected in a series of psychopharmacology studies in the lab of collaborator Harriet de Wit. The results have been preprinted (Carlyle et al, 2023, https://osf.io/zvnse/(s’ouvre dans une nouvelle fenêtre)) and submitted for publication.
Study 2 setup and results:
For Study 2, we developed a novel, dyadic implementation of the TSST, the Dyadic Stress and Support Task (DSST). The DSST is optimised to measure the effects of social support on stress recovery, which will allow us in turn to determine the importance of endogenous opioids for support-induced stress relief. The DSST has two parts, stress induction and recovery. During stress induction, pairs of friends complete the TSST in parallel, separated by a physical barrier such that they can hear, but not see each other. During recovery, the two friends are left to talk freely with each other (social support condition). In the non-support control condition, each participant instead interacts with a panel member instructed to appear professional, but not supportive.
The dyadic setup has two major benefits. First, we test two participants at the same time, which reduces the number of test personnel needed for TSST sessions. Second, by recruiting pairs of long-standing, close friends, we ensure the ecological validity of the social support.
Pilot testing to validate the paradigm was conducted in N=42 (21 pairs of friends) randomised to social support or non-support after stress induction. Ratings of affective state, heart rate and salivary cortisol data support the validity and utility of the DSST stress induction. Social support caused a large boost in positive affect that was not observed during recovery in the control condition. The pilot data and a detailed description of the DSST has been written up and preprinted (Løseth et al, 2022, PsyArXiv link).
Data collection for the main study (final N=258) was completed autumn 2023. Data analysis is ongoing.
Scientifically, the project has so far contributed with several publications in scientific journals, as well as a number of talks and presentations at conferences, seminars and workshops. Of the results so far, the below publications can be highlighted (presented in roughly chronological order):
1. The manuscript presenting the DSST, as described above (Løseth et al, 2022, Psychoneuroendocrinology).
2. Our team has also published several peer-reviewed comments and review articles on the topics studied in OPIOIDREWARD. In Leknes & Atlas, 2020, British Journal of Anaesthesia, we present a checklist for psychopharmacology studies, highlighting and explaining the importance of placebo control and double-blinding; counterbalancing of conditions; choice of drug, dose and administration method; construct validity; study population and generalisability; and reproducibility and reporting. Eikemo, Løseth and Leknes (2021, Pain) review effects of opioid antagonists and present a guide for how to interpret results of full and partial blockade of mu-opioid receptors. Meier, Eikemo and Leknes (2021, Current Addiction Reports) review the psychopharmacological evidence on opioid modulation of reward and threat using a translational perspective. In brief, we find that many results in the human literature point to a preserved function of the endogenous opioid system across rodents and humans, however effect sizes are typically smaller. We conclude that the human endogenous opioid system modulates human subjective experience and related responses in concert with other neurotransmitter systems.
3. Our primer on opioid antagonism to understand endogenous opioid function and two open-access web apps to facilitate study planning (planoxone and plantrexone): Opioid antagonism in humans: a primer on optimal dose and timing for central mu-opioid receptor blockade (Trøstheim, M Eikemo, J Haaker, JJ Frost, S Leknes
Neuropsychopharmacology 48 (2), 299-307)
4. Three papers summarizing the link between pre-drug negative affect, positive medication responses and subsequent opioid self-administration in surgery patients (by Eikemo, Meier and Carlyle et al, details described above)
5. The main paper of the project has been preprinted and describes the test of the core hypothesis, namely that receiving opioids when in a stress state increases subsequent opioid self-administration (the gold-standard measure of drug abuse liability). Our results confirm the hypothesis, but the analysis clearly shows that the effect is driven by male participants (N=31) with no stress-enhanced opioid drug wanting in female participants (N=32). Eikemo et al, 2023: https://osf.io/preprints/v8dxy/(s’ouvre dans une nouvelle fenêtre))
The methodological advances described above constitute progress beyond the state of the art with respect to psychological interventions. We have also made substantial progress with respect to pharmacological interventions:
First, we have conducted pilot testing to establish the drug, dose and administration method optimised to induce drug reward in healthy non-opioid users in the non-stress condition. Initial dose testing was conducted using intravenous morphine administered in a five-minute infusion. Participants were asked to report qualitatively on their responses and where relevant, to compare the effects to those of alcohol. Two participants received 5 mg/70 iv morphine, which was insufficient (‘I’m out drinking but I’m one beer short’). The next participants received 10 mg/70 kg, which was too much (‘I’m out drinking, it’s 3 am and I want to go to bed soon’). 7 mg/70 kg yielded more consistently positive reponses during dose piloting.
Nevertheless, pilot testing using this dose in conjunction with the full setup taught us that morphine reward was inconsistent in healthy volunteers who were not opioid users. For instance during debrief, one participant stated that they would ‘pay money to never receive this drug again’. We also saw no clear evidence of opioid-induced stress relief in the initial pilot data (42 sessions). We therefore changed from morphine to a comparable dose of oxycodone, a drug associated with a more favourable proportion of drug liking/drug disliking responses. Pilot testing with oxycodone showed that many participants chose to work for another dose of the drug. This strong evidence of incentive salience of oxycodone was even observed in participants who had vomited after the initial dose.
Second, we have conducted an in-depth investigation into the blockade profile of commonly used opioid antagonists. Synthesizing the available positron emission tomography and dual detector data on blockade of mu-opioid receptors from these drugs at varying doses and time points, we have generated several drug-to-blockade models, a primer manuscript (Trøstheim et al, 2022, BioArXiv link) and two user-friendly, open access web applications that allow researchers to convert naloxone and naltrexone doses to opioid receptor blockade. The manuscript includes a detailed set of recommendations for researchers designing new studies and evaluating existing literature using opioid antagonists in humans.
Expected results until the end of the project include tests of the key hypotheses of OPIOIDREWARD:
1. That social distress enhances drug wanting (self-administration of oxycodone) but not self-reported drug liking.
2. Distress relief induced by social support depends on brain opioids.
3. The reward/ relief effects of opioid drugs depend on dopaminergic signalling.
First, we have conducted pilot testing to establish the drug, dose and administration method optimised to induce drug reward in healthy non-opioid users in the non-stress condition. Initial dose testing was conducted using intravenous morphine administered in a five-minute infusion. Participants were asked to report qualitatively on their responses and where relevant, to compare the effects to those of alcohol. Two participants received 5 mg/70 iv morphine, which was insufficient (‘I’m out drinking but I’m one beer short’). The next participants received 10 mg/70 kg, which was too much (‘I’m out drinking, it’s 3 am and I want to go to bed soon’). 7 mg/70 kg yielded more consistently positive reponses during dose piloting.
Nevertheless, pilot testing using this dose in conjunction with the full setup taught us that morphine reward was inconsistent in healthy volunteers who were not opioid users. For instance during debrief, one participant stated that they would ‘pay money to never receive this drug again’. We also saw no clear evidence of opioid-induced stress relief in the initial pilot data (42 sessions). We therefore changed from morphine to a comparable dose of oxycodone, a drug associated with a more favourable proportion of drug liking/drug disliking responses. Pilot testing with oxycodone showed that many participants chose to work for another dose of the drug. This strong evidence of incentive salience of oxycodone was even observed in participants who had vomited after the initial dose.
Second, we have conducted an in-depth investigation into the blockade profile of commonly used opioid antagonists. Synthesizing the available positron emission tomography and dual detector data on blockade of mu-opioid receptors from these drugs at varying doses and time points, we have generated several drug-to-blockade models, a primer manuscript (Trøstheim et al, 2022, BioArXiv link) and two user-friendly, open access web applications that allow researchers to convert naloxone and naltrexone doses to opioid receptor blockade. The manuscript includes a detailed set of recommendations for researchers designing new studies and evaluating existing literature using opioid antagonists in humans.
Expected results until the end of the project include tests of the key hypotheses of OPIOIDREWARD:
1. That social distress enhances drug wanting (self-administration of oxycodone) but not self-reported drug liking.
2. Distress relief induced by social support depends on brain opioids.
3. The reward/ relief effects of opioid drugs depend on dopaminergic signalling.