Risk of psychosis in illicit amphetamine users: a 10 year retrospective cohort study

Background

Amphetamine and its analogues, such as methamphetamine, are addictive psychostimulants inducing CNS excitation.1 The estimated global prevalence of amphetamine use is 0.7%, with dependence affecting 11% of people who use amphetamine.2 Through interacting with nerve terminals that use catecholamines (dopamine and norepinephrine) or indoleamine (serotonin) as neurotransmitters, amphetamine elicits initial euphoric yet chronic toxic effects on the brain.3 Amphetamine toxicity is frequently seen in patients with acute illicit substance complications at the emergency department,4–6 and mortality in people with regular or problematic amphetamine use is elevated.7 Currently, cognitive–behavioural treatment is the first choice therapy,8 whereas no pharmacotherapies for amphetamine use disorder have been approved by the US Food and Drug Administration or the European Medicines Agency.9 10

Abuse of amphetamine frequently leads to psychosis, with positive and cognitive symptoms similar to those of schizophrenia.11–16 By causing an overflow of dopamine in the striatum that generates excessive glutamate release into the cortex, amphetamine may cause damage to cortical interneurons, thereby dysregulating thalamocortical signals and resulting in psychotic symptoms.11 Amphetamine psychosis usually subsides within days, but symptoms such as delusion of persecution, delusion of reference, auditory hallucination and visual hallucination may persist for years in a minority (5–15%) of people who use amphetamines.17

Objective

To understand the association between illicit amphetamine use and subsequent psychosis as well as the effect of rehabilitation for amphetamine misuse in a population based setting, we analysed the incidence of psychosis in illicit amphetamine users by connecting the Taiwan Illicit Drug Issue Database (TIDID) and the National Health Insurance Research Database (NHIRD) during 2007–2016.

Findings

Eligible study participants included 74 601 patients in the amphetamine cohort and 298 404 patients in the non-amphetamine cohort (table 1). The distributions of age and sex on entry were the same in both cohorts. Mean age was 33.3±9.45 years and 84.2% were men. Mean follow-up duration was 4.93±2.93 years in the amphetamine cohort and 4.99±2.93 years in the non-amphetamine cohort. Compared with participants without illicit amphetamine use, a higher proportion of illicit amphetamine users had depression (1.98% vs 0.42%), anxiety (0.91% vs 0.27%), ischaemic heart disease (1.32% vs 0.78%), cardiovascular disease (0.79% vs 0.45%) and stroke (1.31% vs 0.66%).

Table 1

Baseline demographic factors and comorbidities according to illicit amphetamine use

Kaplan–Meier analysis showed that by the end of the 10 years of follow-up, the cumulative incidence of psychosis was six-fold higher for the amphetamine cohort compared with the non-amphetamine cohort (3.75% vs 0.63%; log rank test p<0.0001) (figure 2). Table 2 shows the incidence and HRs for psychosis according to amphetamine status, stratified by demographic factors and comorbidities. Illicit amphetamine users were 6.08 times more likely to experience psychosis than those in the non-amphetamine cohort (4.68 vs 0.77 per 1000 person years), with an HR of 5.28 (95% CI 4.90 to 5.70) after adjusting for age, sex and comorbidities (depression, anxiety, ischaemic heart disease, cardiovascular disease and stroke).

Figure 2Figure 2Figure 2

Comparison of cumulative incidences of psychosis in participants with and without illicit amphetamine use (created by the authors).

Table 2

Incidence rates and hazard ratios for psychosis in subjects with or without illicit amphetamine use by demographic characteristics

The incidence of psychosis was similar across all age groups in both cohorts. The greatest HR magnitude was observed in patients aged ≥45 years (adjusted HR 5.41, 95% CI 4.29 to 6.82). The incidence of psychosis increased with comorbidity in both cohorts. We analysed the association between amphetamine use and the risk of psychosis, stratified by comorbidity, and found higher risks of psychosis in illicit amphetamine users without comorbidities (adjusted HR 6.30, 95% CI 5.79 to 6.86) and with comorbidities (adjusted HR 2.28, 95% CI 1.93 to 2.69).

Table 3 shows the incidence rates and HRs for psychosis in patients with different numbers of arrests for illicit amphetamine use. After adjusting for age, sex and comorbidities (depression, anxiety, ischaemic heart disease, cardiovascular disease and stroke), illicit amphetamine users, even if only captured once, had a significantly higher risk of psychosis compared with the control group. We found a trend for the adverse effects of psychosis in illicit amphetamine users arrested a higher number of times (adjusted HR 4.49, 95% CI 4.09 to 4.93 for one arrest; adjusted HR 6.25, 95% CI 5.42 to 7.20 for five or more arrests, p for trend 0.038; table 3). Furthermore, we divided the amphetamine cohort into two subgroups based on rehabilitation treatment during deferred prosecution to examine the combined effects of psychosis and treatment for amphetamine addiction. We observed that illicit amphetamine users who received treatment had a significantly lower risk of psychosis than amphetamine users who received no treatment (adjusted HR 0.74, 95% CI 0.61 to 0.89), as shown in table 4.

Table 3

Incidence rates and hazard ratios for psychosis associated with number of arrests for illicit amphetamine use

Table 4

Incidence rates and hazard ratios for psychosis in illicit amphetamine users with or without deferred prosecution for rehabilitation treatment

Conclusions and clinical implications

Our analysis revealed that the incidence of psychosis was associated with a history of illicit amphetamine use in a population based setting. The risk existed for all age groups, particularly for women. In addition, the risk of psychosis increased as the number of arrests increased, and decreased when patients received psychotherapy for amphetamine addiction.

The annual cumulative incidence rates for psychosis in controls and amphetamine users were 77 and 468 per 100 000 persons, respectively (table 2). The international incidence of all psychotic disorders, 2002–17, was 26.6 per 100 000 person years.20 Estimated true incidence rates of psychotic symptoms were 86 per 100 000 per year among those aged 15–29 and 46 per 100 000 in those aged 30–59 in the USA.21 Pooled incidence of all psychoses in England, 1950–2009, was 31.7 per 100 000 person years.22 Compared with other published data, the incidence of psychosis in our controls was slightly higher. Psychotic symptoms are frequently experienced in methamphetamine users, with estimates of up to nearly 40% of users affected.14 To the best of our knowledge, no cumulative incidence of psychosis in amphetamine users has been reported before, so we were unable to compare our findings with other published data.

Consistent with the systematic review supporting amphetamine use (rather than comorbidities or psychosis risk) as the primary risk factor for amphetamine associated psychosis,23 we found that the association of amphetamine and subsequent psychosis was more prominent in patients without comorbidities, inferring a direct effect of amphetamine on psychotic symptoms. In line with a former study by McKetin et al 24 demonstrating the dose–response relationship of methamphetamine use and psychotic symptoms, our results showed that the risk of psychosis increased as the number of arrests for amphetamine use increased. But we could not assert the dose dependence since there were no validated measures of dose, frequency, amount of use or severity of dependence of amphetamine in TIDID.

In accordance with previous research, we discovered that illicit amphetamine use was associated with anxiety and depressive symptoms25 as well as cardiovascular complications.26 The risk of amphetamine associated psychosis was identified across all age groups. Because persistent psychotic symptoms could represent a risk for cognitive decline in amphetamine users,27 identifying the illicit amphetamine users with psychosis and providing treatment early might prevent subsequent damage of cognitive functions. We observed that women were at a greater risk of amphetamine induced psychosis than men, corresponding to previous animal studies. One explanation could be that amphetamine induced behaviour is enhanced by the presence of oestrogen, whereas testicular hormones may have the opposite effect.28 Another possibility is that women arrested for illicit amphetamine use were particularly disadvantaged in comparison with men, with higher levels of trauma, lack of psychosocial support and stigma.29

We also learnt that illicit amphetamine users who received rehabilitation treatment via deferred prosecution30 were associated with a lower risk of psychosis. This finding implies a preventive effect of psychotherapy for amphetamine use disorder on subsequent psychosis development. This is in line with previous findings on the efficacy of psychosocial interventions for cocaine and amphetamine use disorders.31 There is also evidence that comprehensive and sustained psychological treatment is more effective than less intensive comparison conditions in reducing methamphetamine use and improving psychiatric symptoms.32 33

Our study had several limitations. Illicit amphetamine use could precipitate and aggravate schizophrenic symptoms,34 35 so we could not rule out the possibility that amphetamine is the inducer, instead of the cause, of the psychotic symptoms seen in amphetamine users. Also, the arrest can often be a reflection of police activity/priorities (which can change over time) and while it may be a proxy for amphetamine use (especially given the biological confirmation), it does not capture all amphetamine use (ie, many people who use amphetamines will never be detected/arrested, and it is possible that the ‘non-amphetamine’ group includes a number of people who use or have used amphetamine but have never been arrested for doing so). Further, it may be that it is the more frequent amphetamine users who are more likely to come to the attention of the police and be arrested, while higher functioning or more 'recreational' amphetamine users may be less likely to come to the attention of police.

The arrest records for illicit amphetamine use in Data of Drug Crime and Penalty from the National Police Agency are primarily used for judicial punishment, and the severity of amphetamine addiction of each offender was not documented for scientific purposes. But the urine test was required for the criminal charge, suggesting that the use of amphetamine was validated. In addition, amphetamine users often have lower socioeconomic status, which could be the issue leading to amphetamine use as well as the development of psychosis. However, there is controversy regarding the association between low socioeconomic status and psychosis,36 so the omission of socioeconomic status in our matching of users and non-users might not have affected our results.

We could not ignore the potential selection bias from prosecutors because deferred prosecution was offered to only ~10% of the illicit amphetamine users with promising treatment outcomes, as judged by the prosecutors, not by patients’ willingness to receive treatment for amphetamine misuse. Since individuals who received rehabilitation were having their treatment at the psychiatric department in a hospital setting during deferred prosecution, the prosecution process/procedure might increase the likelihood of being diagnosed. Also, gene–environment interactions have been linked to psychosis spectrum disorders, involving multiple common genetic variants and several environmental factors that constitute a dense network of exposures, named the exposome.37 However, TIDID and NHIRD contain no genetic data and lack detailed information about environmental exposures, such as ethnic group, obstetric complications and childhood adversity.38 Therefore, we were unable to control for these confounders.

Because analysing the entire population is time consuming and the gain of statistical power is little when increasing the control to case ratio beyond five, we chose four controls to match each case.39 However, this might make the study sample less representative of the whole population. In particular, controls who were eligible to be matched will not be included once the corresponding case had reached the matching ratio of four. Also, our analysis reported a single average HR throughout the whole of the following period. But the average HR may be misleading because of potentially time varying period specific HRs, and because the period specific HRs may be time varying due to built-in selection bias.40

Similar to the effect of smoking on the risk of cancer, our study was attempting to investigate the effect of amphetamine use on the risk of psychosis. In common with smoking status that is ever changing during the follow-up period, amphetamine use status is also a time varying covariate. A time varying covariate could be handled with a time dependent Cox model, where the time varying coefficient is described using a step function and a continuous function.41 Unfortunately, due to the illicit nature of recreational amphetamine use, we could not obtain the exact periods and doses of amphetamine administration, and hence were not able to perform the time dependent analysis.

Our study indicated that psychosis is associated with illicit amphetamine use across all age groups, especially in women and those arrested multiple times. The influence of illicit amphetamine use on subsequent psychosis development appears to be mitigated through psychotherapy for amphetamine addiction. The relation of an induced paranoid psychosis with amphetamine abuse has been known for many decades. Nonetheless, our findings are from a detailed and comparative analysis using a comprehensive and large population dataset. Furthermore, it would be worthwhile to investigate the health benefits and cost effectiveness of deferred prosecution for drug crime offenders by providing appropriate therapy for drug addiction.

留言 (0)

沒有登入
gif