Skip to main content
. 2016 Sep 7;6:32730. doi: 10.1038/srep32730

Table 2. Odds ratio (OR) estimates of bipolar disorder, major depressive disorder (MDD) and schizophrenia per 1-standard deviation increment in BMI based on IVW, MR-Egger and weighted median approaches.

Included SNPsa Parameter Statistic IVW Weighted median MR-Egger MR-Egger (SIMEX)b
Outcome: Bipolar disorder
All Intercept Odds (95% CI) 0.99 (0.97; 1.01) 0.99 (0.97; 1.01)
(97 SNPs)   P 0.281 0.294
  OR OR (95% CI) 0.90 (0.69; 1.16) 0.88 (0.62; 1.25) 1.23 (0.65; 2.31) 1.26 (0.63; 2.52)
    P 0.416 0.482 0.517 0.516
Non-influential Intercept Odds (95% CI) 0.99 (0.98; 1.01) 0.99 (0.98; 1.01)
(92 SNPs)   P 0.319 0.335
  OR OR (95% CI) 0.93 (0.74; 1.17) 0.88 (0.62; 1.26) 1.19 (0.69; 2.05) 1.21 (0.67; 2.20)
    P 0.532 0.487 0.518 0.526
Outcome: Schizophrenia
All Intercept Odds (95% CI) 0.99 (0.98; 1.00) 0.99 (0.97; 1.00)
(96 SNPs)   P 0.101 0.103
  OR OR (95% CI) 0.98 (0.80; 1.19) 0.93 (0.78; 1.11) 1.41 (0.87; 2.27) 1.46 (0.86; 2.47)
    P 0.806 0.420 0.162 0.161
Non-influential Intercept Odds (95% CI) 0.99 (0.98; 1.00) 0.99 (0.98; 1.01)
(92 SNPs)   P 0.280 0.283
  OR OR (95% CI) 1.01 (0.86; 1.18) 0.93 (0.78; 1.10) 1.22 (0.83; 1.81) 1.25 (0.81; 1.92)
    P 0.944 0.406 0.310 0.315
Outcome: Major depressive disorder
All Intercept Odds (95% CI) 1.00 (0.98; 1.01) 1.00 (0.98; 1.01)
(90 SNPs)   P 0.669 0.622
  OR OR (95% CI) 1.15 (0.92; 1.44) 1.40 (1.03; 1.90) 1.28 (0.74; 2.24) 1.33 (0.72; 2.47)
    P 0.221 0.035 0.374 0.364
Non-influential Intercept Odds (95% CI) 0.99 (0.98; 1.01) 0.99 (0.98; 1.01)
(86 SNPs)   P 0.385 0.328
  OR OR (95% CI) 1.25 (1.02; 1.52) 1.45 (1.05; 1.99) 1.52 (0.93; 2.49) 1.60 (0.93; 2.75)
    P 0.030 0.026 0.094 0.087

95% CI: 95% Confidence interval. P: P-value. SIMEX: Simulation Extrapolation.

aAll corresponds to Mendelian randomization analysis using all BMI-associated SNPs available in the correspondent psychiatric dataset. “Non-influential” is similar, but excludes SNPs classified as influential using statistical tests based on studentized residuals and Cook’s distance (see Materials and Methods for details).

bThis differs from regular MR-Egger regression because it uses the SIMEX method to correct for regression dilution bias.