Transcript IS CANNABIS A RISK FACTOR FOR SCHIZOPHRENIA?

IS CANNABIS A RISK FACTOR
FOR SCHIZOPHRENIA?
Jouko Miettunen
Department of Public Health and Primary Care
Institute of Public Health
University of Cambridge
February 3, 2003
CONTENTS OF THE PRESENTATION
 cannabis and cannabis use
 schizophrenia
 association and causality
 summaries and limitations
of the studies
 conclusions
SOURCE OF CANNABIS
 hemp plant, Cannabis sativa
 contain cannabinoids
 major active component

9-tetrahydrocannabinol
 preparations of cannabis

illicit drugs
• marijuana (leaves, stalks, flowers, seeds)
• hashish (resin)

also legal drugs
 conflicting attitudes among researchers
CANNABIS USE
 measured by questionnaires and urine/hair test
 known effects





10% become dependent and gateway to other drugs
depression and anxiety
somatic disorders (e.g. cancer)
impair cognitive and driving skills
brain effects (releases dopamine)
 use as a therapeutic drug


multiple sclerosis, epilepsy, cancer, AIDS, etc.
BMA (1997): “Therapeutic Uses of Cannabis”
PREVALENCE OF CANNABIS USE
Annual prevalence estimates of cannabis use in the late 1990s
(“official statistics” i.e. various questionnaires, surveys and estimates)
TOTAL
EUROPE
United Kingdom
Netherlands
NORTH AMERICA
United States
SOUTH AMERICA
ASIA
China
India
AFRICA
OCEANIA
0
3.5%
4.9%
9.4%
4.1%
6.6%
8.3%
4.7%
1.6%
0.5%
3.2%
8.1%
18.8%
5
10
15
20
% of population age 15 and above
United Nations Office on Drugs and Crime
CANNABIS USE BY AGE
current monthly use (survey in New York, N=1,160)
Chen et al. 1995
use among UK students (Webb et al. 1996)
• any use 60% and regular use 20%
use is increasing in most countries
• especially among people under age 16
• in some parts of the world more common than alcohol use
SCHIZOPHRENIA
 chronic, severe, and disabling mental disease
 diagnosed using structured interviews (ICD-10: F20)
 life-time prevalence approximately 1%

not increasing in general, though e.g. in south London
 prevalence of some psychotic symptoms in general
population (Eaton et al. 1991):

paranoid symptoms 10%

hallucinations 5-8%

bizarre delusions 2%
AGE AT ONSET OF SCHIZOPHRENIA
patients (%)
30
Female
Male
20
10
0
12-14 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59
years
age group
Häfner et al. 1993
PREVALENCE OF CANNABIS USE
AMONG PSYCHOTIC PATIENTS
 difficult to compare due to the selection of cases
(inpatients/outpatients) and controls
 most case-control studies report that cannabis use is
about 2 times more common among psychotic patients
than among general population based controls
 among schizophrenia patients

prevalence estimates vary between samples from 5 to 50 %

4 times more often any drugs (UK study, McCreadie 2002)
CANNABIS USERS AMONG
SCHIZOPHRENIA PATIENTS
 younger age at onset and more males
 more unemployment and alcoholism
 worse course of schizophrenia




more positive symptoms
poorer compliance with treatment
more frequent hospitalisation (unclear?)
less negative symptoms in short-term (unclear?)
 more patients with catatonic subtype of
schizophrenia (Hambrecht and Häfner 2000)
EFFECTS OF CANNABIS USE ON
VULNERABLE CASES
 cannabis use is a risk for psychotic diagnosis in
subjects who have already have symptoms
(van Os et al. 2002)
 patients with cannabis associated psychosis have
increased familial risk for schizophrenia
(McGuire et al. 1995)
 some recent high-risk studies
(Phillips et al. 2002, Miller et al. 2001)
CANNABIS USERS IN
GENERAL POPULATION
 have more psychotic symptoms than nonusers at age 18-20 (Fergusson et al. 2003)

adjusted OR 1.8 (95% CI: 1.2-2.6)
 have more often schizotypal personality
traits (Williams et al. 1996, Dumas et al.
2002)
POTENTIAL CONFOUNDERS








age and sex
urban birth, social class and marital status
alcohol use, smoking and use of other drugs
stressful life-events
migrant/minority status (e.g. Afro-Caribbeans in UK)
premorbid symptoms (e.g. social adjustment difficulties)
personality traits and IQ
familial risk of schizophrenia and/or cannabis use
CAUSALITY BETWEEN CANNABIS
USE AND SCHIZOPHRENIA
 generally accepted that cannabis intoxication
can cause brief psychotic episodes
 can cannabis use cause schizophrenia?
 or can the direction of causality be reversed?
PROBLEMS WITH CHRONOLOGY
What is the temporal order?
CANNABIS USE
first use
regular use
heavy use
AGE
premorbid symptoms
SCHIZOPHRENIA
psychotic symptoms
diagnosis
PROBLEMS WITH POOLING THE STUDIES
Various exposure and outcome combinations in the studies:
CANNABIS USE
•any use
•regular use
•heavy use
•times in a life-time
•times in a year/month/…
•current use
•cannabis abuse/dependence
•etc.
SCHIZOPHRENIA
•any psychotic symptoms
•symptoms in a year/month
•pathological level of symptoms
•need for care due to symptoms
•any psychotic diagnosis
•schizophreniform disorder
•schizophrenia
•etc.
SCHIZOPHRENIA AS AN OUTCOME
Swedish conscript study (1)
 cohort of 18-20 year old males (N=50,045)
 questionnaires at conscription 1969/70
 hospital register follow-up until 1995

ICD-8/9 schizophrenia diagnosis
 Andréasson et al. 1987
 Andréasson et al. 1989
 Zammit et al. 2002
Swedish conscript study (2)
risk for schizophrenia:
 ever used cannabis

adjusted OR 1.9 (95% CI: 1.1.-3.1)
 used cannabis more than 50 times

adjusted OR 6.7 (95% CI: 2.1.-21.7)
 significant linear trend for frequency of use
 cannabis use was not associated with other
psychoses than schizophrenia
Swedish conscript study (3)
limitations:
 no information on possible confounding factors in
the follow-up period
 no information on familial risk for schizophrenia
 validity of the exposure (underreporting?)
 validity of the outcome (underreporting?)
 not many cannabis users got schizophrenia

1.4% if ever used

3.8% if used >50 times

0.6% in controls
SYMPTOMS AS AN OUTCOME
Netherlands 1996-99
 population based survey (N=4,045; 18-64 years)
 any cannabis use predicted the presence of
psychotic symptoms at 3-year follow-up



any symptoms: adjusted OR = 2.8 (95% CI: 1.2-6.5)
pathology level of symptoms: adj. OR = 24.2 (5.4-107.5)
statistically significant trend for dose-response
 cannabis use was a risk for psychotic diagnosis in
subjects who already have psychotic symptoms
 limitations: no information on familial risk for
schizophrenia, short follow-up and 43% drop-outs
van Os et al. 2002
New Zealand 1983-99
 general population birth cohort 1972-73 (N=759)
 cannabis use ≥3 times prior to age 15 predicted
 schizophrenia symptoms at 26
 adjusted OR = 6.6 (4.8-8.3)
 and schizophreniform disorder at age 26
 adjusted OR = 3.1 (0.7-13.3) (non-significant)
 use of other drugs was not associated with outcome
 strength: psychiatric symptoms at age 11
 limitations: no information on familial risk for
schizophrenia and did not use schizophrenia as an
outcome
Arseneault et al. 2002
LIMITATIONS OF THE STUDIES
 misclassification bias

lack of confirmation of the biological presence of
cannabis in the organism

reliability of psychiatric diagnoses may be worse in
subjects with comorbid cannabis use
 not always adjusted for all potential confounders
 short follow-up times
 attitude of the researchers

difficult to interpret results and conclusions
PROBLEMS WITH CHRONOLOGY
Schizophrenia patients using cannabis can be defined into groups chronologically
CANNABIS USE
AGE
self-medicating
patients
similar risk factors
for cannabis use and
schizophrenia
or
cannabis is the trigger
vulnerable patients
or
increased dopamine level
increases positive symptoms
of schizophrenia
SYMPTOMS OF SCHIZOPHRENIA
all the groups include also people who have schizophrenia independently on cannabis use, and vice versa!
CONCLUSIONS
 use of cannabis can cause psychotic symptoms and even
schizophrenia especially in some vulnerable cases
 BUT:


would schizophrenia have occurred in these individuals in any
case (cannabis use only precipitates schizophrenia)?
does not count for many schizophrenia cases?
 IN FUTURE:


large prospective studies with long follow-up time, schizophrenia
diagnosis as an outcome and comprehensive information on
confounding variables
case-control study starts in South London 2003