Health conditions linked to lasting changes in perception after hallucinogen use

TriNetX is a global federated health research network providing access to electronic medical records (diagnoses, procedures, medications, laboratory values, genomic information) across large Healthcare Organisations (HCOs). At the time of analysis (10^th April 2025) there were 135 HCOs available, of which 74 are academic or tertiary centres, and 61 are non-academic, community-based institutions, with a total population of over 150 million individuals. The TriNetX platform is compliant with the Health Insurance Portability and Accountability Act (HIPAA) as it only contains de-identified data, which is presented in aggregate form. In this study we used TriNetX to undertake a retrospective cohort analysis, with cumulative incidence as the main outcomes of interest.

The de-identified data was analysed in accordance with the ethical approvals obtained by the TriNetX database. No further ethical approval was required for this study.

We used the TriNetX database to create four cohorts for analysis. Only those aged ≥16 at the index event (date of diagnosis) were included. The four cohorts were:i.A HPPD cohort consisting of individuals coded with F16.983 (Hallucinogen use, unspecified with hallucinogen persisting perception disorder).ii.A population control group consisting of individuals coded with Z00 (General examination and investigation of persons without complaint and reported diagnosis) but never F16.983.iii.A control group of psychedelic users consisting of individuals coded with F16.10 (Hallucinogen abuse, uncomplicated) or F16.90 (Hallucinogen use, unspecified, uncomplicated) but never F16.983.iv.A control visual symptoms cohort (VSC) consisting of individuals coded with H53.8 (Other visual disturbances) but never F16.983. This code likely includes, but is not limited to, people with a diagnosis of visual snow syndrome [20]; a specific ICD code for VSS has been inaugurated, but is not yet in use.

Outcomes were compared prior to or following the index event, which is characterised as the date of first receiving cohort diagnosis (F16.983, Z00, F16.10/F16.90, or H53.8). Diagnoses recorded prior to index event (i.e. cumulative incidences) are presented as percentages.

The analyses utilised multiple comparisons. The largest number of comparisons run per analysis was 41. Via a manual Bonferroni calculation, a significant p-value was calculated as 0.05/41 = 0.00122; this was rounded down to 0.0010, which was the threshold for significance across all analyses.

People with HPPD were more likely to develop Group A, Group B, Group G disorders in comparison to population controls and psychedelic users. The HPPD group was more likely to develop degenerative visual disorders than psychedelic users, but no more than controls (see Fig. 1 and Supplementary Table 4). The HPPD group was more likely to develop neurodegenerative disorders than population controls and psychedelic users, but not visual symptom controls (see Supplementary Table 1).

We further analysed the conditions driving the increased odds of developing neurodegenerative disorders in the HPPD group (2.3%) versus population controls (1.2%). This was driven by increased rates of unspecified neurodegenerative disorders (0.9% vs. 0.3%) and Parkinson’s disease (0.2% vs. 0.1%), see Supplementary Table. The relative increase in odds of developing these conditions may be confounded by the higher rates of antipsychotic prescribing in the HPPD group (see Supplementary Table). 1 1

The increased odds of developing degenerative visual conditions in HPPD (2.7%) vs psychedelic users (1.4%) was driven by greater rates of glaucoma (1.3% vs 0.7%), other retinal disorders (1.0% vs 0.6%), and other disorders of optic pathways (0.5% vs 0.2%). There were no increased odds of developing degenerative visual disorders versus population controls.

There was a high level of cumulative psychiatric morbidity in the HPPD cohort. There were high rates of substance misuse (recorded pre-index), including alcohol use, which seemed to be shared with the psychedelic-using control group, albeit often at higher proportions. ADHD was also notably elevated, which may suggest a role for neurodevelopmental traits in predisposition to using psychedelics, as well as a risk for developing HPPD following psychedelic use.

Almost a third of the HPPD cohort had diagnoses of depressive episodes and unspecified anxiety disorders. There were also significantly higher rates of anxiety in the HPPD group versus psychedelic-using controls (25.1% vs 19.6% for unspecified anxiety). Psychedelic users with a prior diagnosis of anxiety were 1.5x more likely to develop HPPD.

Our findings on high rates of anxiety disorders are in accordance with previous literature on the topic. In a cross-sectional online survey of people with symptoms of Type I HPPD (n = 165), 35% reported a diagnosis of anxiety [9]. A review of HPPD studies also found that anxiety is commonly seen at higher rates in the HPPD population, and the presence of anxiety may increase the frequency and severity of symptoms [2]. Other research has shown anxiety, negative trip experiences, and lack of prior knowledge are associated with negative experiences of persistent visual symptoms following psychedelic use [21].

HPPD has been reported following both positive and negative psychedelics experiences, but individuals who have highly anxious psychedelic trips are more likely to experience long-term distressing symptoms [7, 22]. Similarly, although HPPD like phenomena have been seen in up to 30% of people following a psychedelic experience, in most cases they are not perceived as distressing [23]. Some authors have suggested that HPPD-specific distress correlates more strongly with anxiety and maladaptive cognition than with symptom severity [7]. Our findings are also in line with literature which has shown increased odds of anxiety disorders in people with functional somatic syndromes [24].

Our HPPD cohort also had higher pre-index rates of other systemic and physical disorders, including chronic pain (15.9%), headache syndromes (14.7%), post-viral fatigue (12.3%), and fibromyalgia (6.7%). Psychedelic users with a prior diagnosis of post-viral fatigue syndrome had almost twice the odds of developing HPPD. The association with these conditions suggests avenues for further research to more directly probe any potential mechanistic associations between HPPD and functional somatic syndromes and somatic symptom disorders.

Some of the visual phenomena reported by people with HPPD, such as transient floaters and flashes (known as entoptic phenomena) are also reported in healthy populations [22]. Benign visual phenomena are common in healthy people and in those who have not used psychedelic drugs [2, 25]. Similar phenomena to those associated with HPPD, including derealisation and visual symptoms, have been reported as adverse events following meditation and mindfulness practices [26] and conventional antidepressant medications [27]. Non-distressing visual phenomena are also common following psychedelic use based on self-reported samples [28].

There have been studies suggesting the clinical, phenotypic, and phenomenological overlap between HPPD and visual snow syndrome (VSS). In VSS, reported to be present in around 3.7% of the UK population [29], people report the presence of dots in the visual field which can appear akin to static on a television screen. Of note, VSS does not require a precipitating event, such as precipitant psychedelic use is required for a diagnosis of HPPD, and the inconsistent migraine association with HPPD — stronger than psychedelic controls but weaker than VSS — may point to distinct pathophysiological mechanisms.

Visual snow syndrome has recently been suggested to be a multi-network disorder of predictive processing, whereby inert visual phenomena are processed with aberrant salience and attention, implicating the role of increased self-monitoring in the perpetuation of symptoms [16, 17, 30]. Severity of VSS symptoms have also been correlated with scores on the Modified Tellegen Absorption Scale [31].

Previous authors have suggested that HPPD may be ‘explained in terms of a heightened awareness of and concern about ordinary visual phenomena’ [2, 13]. This current study was not designed to make any inferences about underlying pathophysiology, however the associations of HPPD with anxiety and functional somatic symptoms may represent avenues for future research. It is possible that HPPD may involve hyperawareness of benign visual stimuli, which are given undue prominence in the conscious processing of visual experience.

The role of narrative and diagnostic practices in HPPD may also interplay with symptoms. Despite sometimes presenting very similarly to non-drug-induced cases, HPPD is defined around drug induction and specifically separated from VSS by its association with ‘illicit’ drugs. This may contribute to stigma, negative self-conceptions, and catastrophising cognitions. Similarly, a narrowed focus on toxic or neurological causes may contribute to hyperfocus on symptoms, which are interpreted solely through a neuropharmacological lens, limiting psychosocial understanding and with potential attendant effects on recovery.

In this study, the presence of HPPD was associated with greater risk of developing psychiatric and physical disorders in future, including anxiety, depression, pain disorders, and migraine. This may be a consequence of the burden of living with the distressing symptoms of HPPD, or alternatively may represent underlying shared risk factors that pertain to mechanisms underlying symptom production and maintenance. Having a diagnosis of HPPD was also associated with a doubled risk of developing functional somatic syndromes versus psychedelic-using controls.

People with HPPD were more likely to develop neurodegenerative disorders than controls (but not visual symptom controls). The absolute risk remained small (2.0% of sample) and may be in part explained by the significantly higher rates of prescription of antipsychotic medications in this group. Antipsychotics are associated with the development of, for example, secondary Parkinsonism [32].

The limited association with neurodegenerative disorders is in keeping with neuropsychological profiles of patients with HPPD from a recent, small study, which showed no differences in cognitive performances versus psychedelic using and healthy controls [12]. This is also consistent with literature on psychedelic users which indicate no negative effects on cognition [33]. Nevertheless, further longitudinal studies should aim to confirm suggestions of baseline population rates of neurogenerative disorders in HPPD populations.

Using codes from the TriNetX database is dependent on contemporaneous and accurate recording via electronic health records. It cannot always be distinguished when a diagnosis was first recorded and were unable to necessarily exclude conditions which had been diagnosed pre-index, but which had been recoded post-index. The database features predominantly North American users and may not be applicable globally (although south American and European HCOs have been added).

Patients captured in the TriNetX database are those who receive care within participating health systems, and as a result, the database may overrepresent individuals with more complex presentations or higher healthcare utilisation. This means that incidence estimates may appear higher than in the general population, comorbidity profiles may be skewed toward patients with greater healthcare needs, and generalisability must be interpreted with caution. This is not a limitation unique to TriNetX, but rather one that applies to all EHR-based studies. We believe highlighting this shared limitation provides useful context for situating our study within the broader EHR-based research literature.

The fact that ICD-10 codes were used to establish psychedelic use likely biases the sample to heavier users and may miss out on occasional psychedelic users who do not present to medical services. Granular data on substance use was not available, and therefore our data lacks sufficient specificity to account for differences in substance type, dose, route of administration, usage context (recreational versus clinical), and polysubstance co-use. Future research should aim to use large datasets with more granular information on psychedelic use, which would improve upon the generalisability of the current results.

The ICD-10 codes used for HPPD in this study are broad and non-specific. Diagnostic criteria for HPPD may vary across institutions [28], and we could not distinguish between type I and type II HPPD.

The control group, defined by ICD-10 code Z00 (our population controls), may not represent the general healthy population. This group may have been seen for routine check-ups that are not necessarily indicative of good health, and may include individuals undergoing, for example, pre-surgical evaluations, cancer screenings, or follow-ups for prior health issues. Many of these individuals carry significant underlying comorbidities. As such, the medical or psychiatric background of population controls may be comparably or even more clinically complex than the HPPD cohort.

Due to the lack of a specific ICD-10 code for visual snow syndrome, we used a broader category likely encompassing patients with other conditions. EHRs may not capture all dementia cases due to limited follow-up periods, and we were unable to control for antipsychotic use in assessing dementia risk.