Adolescents who smoked cigarettes heavily, and particularly those who began early, were more prone to develop psychotic symptoms during 15-year follow-up than their nonsmoking counterparts. The association persisted with multiple confounder adjustments, including for baseline psychosis, a first in longitudinal studies that helped establish directionality, according to researchers who published their report in Acta Psychiatrica Scandinavica.

Despite the recognition that early daily tobacco use likely increases the risk for later psychosis, investigation of the effect of covariates such as substance use, existing psychotic features, or parental/family factors, particularly in an adolescent population, was lacking in the literature. Adolescent use could affect the trajectory of a maturing brain, which could affect future mental health, making this research a priority.

The Northern Finland Birth Cohort 1986, a population-based study that relied on national health registers, examined 6081 individuals (47.7% boys) born between July 1, 1985, and June 30, 1986, when they were aged 15 to 16 to ascertain smoking status as well as other substance use, along with psychosis predictors, via 2 surveys (PROD-screen). These participants were categorized as nonsmokers, moderate smokers (1-9 cigarettes/d), or heavy smokers (≥10 cigarettes/d), and also as having early (≤13 years) or later (>13 years) initiation of daily smoking. Diagnoses of psychosis were then assessed through health records for all patients through age 30, with odds ratio (OR) and hazard ratio (HR) calculations performed.

Continue Reading

The sample contained 745 daily smokers (12.3%), and 1883 (31%) individuals met the initial PROD-screen cutoff for potential psychosis prodromal symptoms at baseline. There were 110 cases (1.8%) of psychosis diagnosed over the course of 15 years, according to medical records.

In all analyses, heavy smokers had the highest risk for later psychosis (crude HR, 3.15, 95% CI, 1.94-5.13; P =.000), which remained significant (although somewhat attenuated) even after adjustment for baseline psychosis (adjusted HR, 2.87; 95% CI, 1.76-4.68; P =.000) and final adjustment for all covariates mentioned here (adjusted HR, 2.00, 95% CI, 1.13-3.54; P =.017).

The authors also found a dose-response relationship between number of cigarettes smoked daily and risk for psychosis (adjusted OR, 1.05, 95% CI, 1.01-1.08), such that the heaviest smokers had proportionately greater risk. In addition, those who began smoking earlier demonstrated a greater chance of psychosis compared with later initiation (5.1% vs 2.2%; adjusted HR, 2.84, 95% CI, 1.12-7.18). There was no significant association between lighter smoking and future psychosis risk.

Related Articles

Strengths of the study included a large sample size, allowing robust exploration of associations; use of registers, with minimal loss to follow-up; coverage of a variety of different psychotic disorders within the population; and extensive data that permitted many confounder considerations.

Study limitations included the level of initial attrition, the use of self-report surveys that may lead to underreporting, a possible selection bias against urban areas, the possibility of some reverse causation, assessment at baseline only without time-dependent measurements, potential cannabis use affecting associations detected, and a lack of adjustment for family or childhood adversity.

Even after multivariate adjustments for possible confounders, there appears to be a clear independent relationship between adolescent smoking (especially earlier and heavier consumption) and subsequent development of psychotic disorders. The risk from cigarettes alone, combined with the potential for an adverse effect on mental health, demands rigorous attention from clinicians aimed at preventing young people from ever initiating tobacco use.


Mustonen A, Ahokas T, Nordström T, et al. Smokin` hot: adolescent smoking and the risk of psychosis [published online February 18, 2018]. Acta Psychiatr Scand. doi:10.1111/acps.12863