Impact of pre-trauma recreational drug use on mental health outcomes among survivors of the Israeli Nova Festival terrorist attack

On October 7, 2023, about 4,000 civilians attending the Nova open-air music festival in southern Israel were the victims of a sudden terrorist attack. They had to swiftly react to the attack by running and hiding for extended periods of time to protect their lives.

At the time of the attack, a significant proportion of these people were under the influence of various recreational drugs. We hypothesized that the pre-trauma use of psychostimulants or hallucinogens would be significantly associated with the severity of peri-traumatic dissociation, anxiety, depression, and acute stress disorder (ASD) symptoms in survivors of the attack.

Two hundred thirty-two survivors sought assistance at the Chaim Sheba Medical Center and underwent clinical evaluation. They were considered for this study if they had no severe physical injuries; no first-degree family member killed during the attack; and no history of mental disorders, including post-traumatic stress disorder (PTSD).

Of the 232 survivors screened for the study, 126 met the above criteria and provided informed consent to participate. However, two of them who reported using hallucinogenic mushrooms, and one who reported using ketamine prior to the traumatic event, were excluded from the analysis, due to the small sample size for these drugs, leaving a sample of 123 participants. Their mean age (±SE) was 28.4±0.7 years; 75 of them (60.9%) were male; 68.9% were never married, and 68.2% were holding a high-school degree or equivalent.

Seventy-one of them (57.7%) reported using psychoactive drugs at the festival – 12 only alcohol, nine only lysergic acid (LSD), seven only 3,4-methylenedioxymethamphetamine (MDMA), six only cannabis, three only methylmethcathinone (MMC), 15 various drug combinations including alcohol, and 19 various drug combinations excluding alcohol.

All participants completed several questionnaires, assessing peri-traumatic dissociation (Peritraumatic Dissociative Experiences Questionnaire, PDEQ), post-traumatic anxiety (Generalized Anxiety Disorder-7, GAD-7; and Visual Analog Scale for Anxiety, VAS-A), depression (Patient Health Questionnaire-9, PHQ-9), and ASD symptoms (Posttraumatic Diagnostic Scale, PDS-5).

Both the GAD-7 scores and the PDS-5 hyperarousal scores were significantly higher in the drug-user than in the drug-free group (p<0.05 and p<0.008, respectively). The scores of most participants were above the clinical threshold for these instruments (>10 for GAD-7 in 70.4%, and >28 for PDS-5 in 81.3% of the participants), indicating a very high level of anxiety- and hyperarousal-related symptoms in both groups. Both the PDEQ and PHQ-9 scores were higher in the drug-user than in the drug-free group, but the differences were not significant. No significant differences were found between the groups in the VAS-A, total PDS-5, and PDS-5 subscales. The VAS-A scores of 51.9% of the participants were higher than 6, which is the clinical threshold for this instrument.

A multiple regression analysis was performed on the scores of the PDEQ, PHQ-9, GAD-7, PDS-5 (total and subscales), and VAS-A questionnaires, employing nine sets of independent variables related to drug use, gender and age (see supplementary information).

The multiple regression model for the PDEQ scores was statistically significant (p=0.018). The severity of peri-traumatic dissociation was significantly correlated with alcohol consumption prior to the event (β=0.25, p<0.008), but not with the consumption of any other drug. Consuming alcohol prior to the traumatic event, as compared with consuming other drugs, significantly increased the likelihood of experiencing peri-traumatic dissociation (PDEQ score = 24.8±2.0 vs. 19.3±1.0, p<0.015).

The model for the PHQ-9 scores was statistically significant (p=0.02). The severity of depressive symptoms was significantly correlated with alcohol consumption prior to the event (β=0.32, p<0.001), but not with the consumption of any other drug. Consuming alcohol prior to the event, as compared with consuming other drugs, significantly increased the likelihood of depressive symptoms (PHQ-9 score = 18.7±1.8 vs. 13.8±0.6, p<0.0015).

The model for the GAD-7 scores was statistically significant (p=0.04). The severity of anxiety symptoms was significantly correlated with alcohol consumption prior to the event (β=0.29, p<0.002). Consuming alcohol, as compared with consuming other drugs prior to the event, significantly increased the likelihood of anxiety (GAD-7 score = 16.3±1.0 vs. 12.7±0.6, p<0.004). None of the other drugs consumed prior to the event significantly affected symptoms of anxiety.

The model for the PDS-5 arousal-hyperactivity scores was statistically significant (p=0.03). The severity of arousal and hyperactivity symptoms was significantly correlated with alcohol (β=0.24, p<0.011) and MMC (β=0.24, p<0.011) consumption prior to the traumatic event. Both alcohol and MMC consumption significantly increased the likelihood of experiencing arousal and hyperactivity symptoms, as compared with the consumption of other drugs.

The multiple regression analysis of the PDEQ dissociation score with VAS-A, GAD-7, PHQ-9, and PDS-5 subscale scores, including gender and age, was statistically significant (p=0.0001). Experiencing peri-traumatic dissociation was significantly correlated with the VAS-A score (β=0.31, p<0.02) and with the PDS-5 mood score (β=0.28, p<0.045).

A mediation analysis showed that pre-trauma alcohol consumption positively predicted PDS-5 mood scores (β=0.15, p=0.03), PDS-5 arousal scores (β=0.16, p=0.015), PDS-5 intrusive scores (β=0.23, p=0.001), GAD-7 anxiety scores (β=0.20, p=0.0015), PHQ-9 depression scores (β=0.24, p=0.00025), and PDEQ peri-traumatic dissociation scores (β=0.20, p=0.01). With the introduction of the peri-traumatic dissociation variable into the model as a potential mediator, the association between alcohol consumption and PDS-5 mood, PDS-5 intrusive, GAD-7 anxiety and PHQ-9 depressive scores became less significant (p<0.04, p<0.015, p<0.015 and p<0.02, respectively). This finding suggests that peri-traumatic dissociation partially mediated the association between alcohol consumption and mood, intrusion, anxiety and depressive symptoms.

So, in marked contrast to our expectations, we found that only pre-trauma alcohol consumption, with or without other drugs, significantly increased the risk of peri-traumatic dissociation, anxiety, depression, and ASD symptoms.

Alcohol consumption exerts various effects on brain functions and behavior, ranging from anxiolytic and mild disinhibitory effects to sedation, motor incoordination, altered memory and emotional processing^{1, 2}. Therefore, pre-trauma alcohol consumption may have interfered with the cognitive, emotional and physiological processes necessary to cope with the traumatic event. The previously available evidence on the effect of pre-trauma alcohol use on the development of post-traumatic symptomatology was mixed^3-7.

Importantly, the traumatic event was prolonged (participants had to run and hide for 8-20 hours until they were rescued). Therefore, the survivors may have experienced a hangover, which could have increased their anxiety and traumatic stress⁸.

We also found that peri-traumatic dissociation significantly increased the likelihood of subsequent anxiety and mood symptoms in participants who consumed alcohol. Peri-traumatic dissociation can potentially disrupt the processing and integration of traumatic memories, which may impede recovery and increase the probability of developing post-traumatic disorders, because trauma-related memories persist in a fragmented and unprocessed state⁹.

The Nova massacre provides a unique opportunity to study how pre-trauma drug consumption affects post-trauma mental health outcomes. Yet, our study is not devoid of limitations. First, as in all naturalistic studies, the sample size is constrained. Second, we relied only on the reports of the participants on their consumption of drugs, rather than on measurements of blood concentrations and data regarding the actual quantities of the consumed substances, which hampers a more nuanced understanding of the correlation between drug consumption and the observed outcomes.

Nonetheless, this study presents novel insights into the relationship between pre-trauma alcohol consumption and increased vulnerability to peri-traumatic dissociation, anxiety, depression, and ASD symptoms. Given the widespread prevalence of alcohol consumption in social gatherings – and the increasing occurrence of sexual assaults, physical assaults and vehicular accidents – these findings can have a social and clinical interest. Moreover, since alcohol is a compound with a known pharmacology and mode of action, they can be relevant to the elucidation of the biology of response to traumatic experiences.