Table 5 Schizophrenia PRS modulates the association of cannabis use with self-reported psychotic experiences.

From: Cannabis, schizophrenia genetic risk, and psychotic experiences: a cross-sectional study of 109,308 participants from the UK Biobank

 

Prevalence of self-reported psychotic experiences (ever- vs. never-users), stratified by schizophrenia PRS

PRS-by-ever-use interaction

PRS-by-frequency interaction

0–20th Percentile SCZ PRS

20–40th Percentile SCZ PRS

40–60th Percentile SCZ PRS

60–80th Percentile SCZ PRS

80–100th Percentile SCZ PRS

Any psychotic experience

5.8 vs. 3.6%

AOR = 1.39

[1.16, 1.65]

6.7 vs. 4.0%

AOR = 1.48

[1.26, 1.74]

7.0 vs. 4.2%

AOR = 1.52

[1.30, 1.78]

7.5 vs. 4.3%

AOR = 1.64

[1.40, 1.91]

8.2 vs. 4.7%

AOR = 1.58

[1.36, 1.83]

FDR = 9%

(p = 0.02)

FDR = 2%

(p = 0.01)

Auditory hallucinations

2.2 vs. 1.2%

AOR = 1.36

[1.02, 1.82]

2.4 vs. 1.4%

AOR = 1.32

[1.01, 1.73]

2.7 vs. 1.5%

AOR = 1.54

[1.20, 1.98]

2.8 vs. 1.3%

AOR = 1.91

[1.47, 2.48]

3.2 vs. 1.5%

AOR = 1.73

[1.36, 2.22]

FDR = 9%

(p = 0.04)

FDR = 2%

(p = 0.01)

Visual hallucinations

3.7 vs. 2.4%

AOR = 1.46

[1.18, 1.81]

4.6 vs. 2.7%

AOR = 1.63

[1.34, 1.98]

4.5 vs. 2.7%

AOR = 1.52

[1.25, 1.85]

4.9 vs. 3.0%

AOR = 1.59

[1.31, 1.91]

5.2 vs. 3.0%

AOR = 1.66

[1.38, 1.99]

FDR = 17%

(p = 0.1)

FDR = 7%

(p = 0.06)

Persecutory delusions

0.8 vs. 0.4%

AOR = 1.60

[0.98, 2.60]

1.3 vs. 0.4%

AOR = 2.30

[1.53, 3.45]

1.1 vs. 0.6%

AOR = 1.65

[1.11, 2.44]

1.4 vs. 0.6%

AOR = 1.39

[0.96, 2.00]

1.7 vs. 0.9%

AOR = 1.36

[0.97, 1.89]

FDR = 99%

(p = 1)

FDR = 29%

(p = 0.3)

Delusions of reference

0.5 vs. 0.4%

AOR = 1.07

[0.63, 1.82]

0.7 vs. 0.5%

AOR = 1.29

[0.82, 2.05]

0.8 vs. 0.6%

AOR = 1.07

[0.69, 1.67]

1.1 vs. 0.6%

AOR = 1.54

[1.04, 2.28]

1.5 vs. 0.7%

AOR = 1.68

[1.18, 2.38]

FDR = 9%

(p = 0.05)

FDR = 0.4%

(p = 0.0007)

  1. FDRs for each type of interaction are derived from Benjamini–Hochberg correction for five tests. Square brackets denote 95% confidence intervals. It is noteworthy that our interaction tests (two right-most columns) treat polygenic risk as a continuous variable and do not rely on discretization into quintiles.
  2. FDR false discovery rate.
  3. The bold values are purely for visual emphasis.