Mediation analyses of famine exposure, metabolites, and diabetes
| Effect | a (famine-mediator) | b (mediator-diabetes) | c' (direct effect) | ab (mediated effect) | Proportion of mediation | c (total effect) |
| 2014 | ||||||
| Indole-3-carbinol | ||||||
| Model 1 | −1.20 (−2.10 to −0.31) | −0.15 (−0.37 to 0.07) | 0.44 (−0.72 to 1.61) | 0.18 (−0.09 to 0.68) | 0.18 (−1.88 to 1.28) | 0.05 (−0.03 to 0.12) |
| Model 2 | −0.28 (−0.59 to 0.25) | −0.28 (−0.93 to 0.37) | 0.31 (−0.91 to 1.53) | 0.08 (−0.10 to 0.46) | −0.001 (−0.33 to 1.28) | 0.08 (−0.05 to 0.25) |
| PC(22:6(4Z,7Z,10Z,13Z,16Z,19Z)/16:1(9Z)) | ||||||
| Model 1 | −5.68 (−9.72 to −1.63) | −0.01 (−0.06 to 0.03) | 0.53 (−0.63 to 1.70) | 0.07 (−0.23 to 0.73) | 0.08 (−1.69 to 1.42) | 0.04 (−0.04 to 0.11) |
| Model 2 | −0.35 (−0.66 to −0.44) | −0.003 (−0.68 to 0.67) | 0.39 (−0.93 to 1.62) | 0.001 (−0.35 to 0.56) | 0.0002 (−4.65 to 1.94) | 0.08 (−0.07 to 0.25) |
| Pyrimidine | ||||||
| Model 1 | −0.28 (−0.47 to −0.08) | −1.19 (−2.49 to 0.11) | 0.32 (−0.86 to 1.50) | 0.33 (0.05 to 1.02) | 0.27 (−2.25 to 2.47) | 0.05 (−0.05 to 0.13) |
| Model 2 | −0.29 (−0.59 to 0.01) | −0.62 (−1.50 to 0.25) | 0.23 (−0.99 to 1.45) | 0.18 (−0.01 to 0.67) | −0.003 (−1.18 to 0.79) | 0.07 (−0.03 to 0.20) |
| PC(16:1(9Z)/22:5(4Z,7Z,10Z,13Z,16Z)) | ||||||
| Model 1 | −0.23 (−0.38 to −0.08) | −0.23 (−1.47 to 1.01) | 0.55 (−0.61 to 1.72) | 0.05 (−0.26 to 0.59) | 0.066 (−2.19 to 2.06) | 0.05 (−0.02 to 0.10) |
| Model 2 | −0.39 (−0.69 to −0.08) | 0.06 (−0.62 to 0.73) | 0.42 (−0.82 to 1.66) | −0.02 (−0.36 to 0.51) | −8.68e−03(−1.73e+00 to 2.53) | 0.06 (−0.05 to 0.15) |
| 2019 | ||||||
| Indole-3-carbinol | ||||||
| Model 1 | −0.45 (−0.87 to −0.02) | −0.39 (−0.72 to −0.07) | 1.05 (0.12 to 1.98) | 0.18 (−0.01 to 0.50) | 0.12 (0.01 to 1.11) | 0.13 (0.04 to 0.22) |
| Model 2 | −0.33 (−0.62 to −0.05) | −0.41 (−0.91 to 0.09) | 1.02 (0.06 to 1.98) | 0.14 (−0.05 to 0.47) | 0.09 (−0.03 to 0.38) | 0.16 (0.02 to 0.34) |
| PC(22:6(4Z,7Z,10Z,13Z,16Z,19Z)/16:1(9Z)) | ||||||
| Model 1 | −0.87 (−1.72 to −0.01) | −0.11 (−0.27 to 0.05) | 1.10 (0.18 to 2.02) | 0.09 (−0.07 to 0.33) | 0.06 (−0.13 to 0.18) | 0.14 (0.05 to 0.24) |
| Model 2 | −0.31 (−0.59 to −0.03) | −0.15 (−0.64 to 0.35) | 1.07 (0.12 to 2.02) | 0.05 (−0.15 to 0.28) | 0.01 (−0.46 to 0.20) | 0.15 (0.01 to 0.27) |
| Pyrimidine | ||||||
| Model 1 | −0.10 (−0.19 to −0.01) | −1.62 (−3.13 to −0.11) | 1.05 (0.13 to 1.98) | 0.17 (0.01 to 0.43) | 0.09 (−0.005 to 0.40) | 0.13 (0.04 to 0.21) |
| Model 2 | −0.33 (−0.61 to −0.05) | −0.38 (−0.88 to 0.12) | 1.01 (0.51 to 1.97) | 0.12 (−0.02 to 0.40) | 0.03 (−0.06 to 2.04) | 0.14 (−0.04 to 0.34) |
| PC(16:1(9Z)/22:5(4Z,7Z,10Z,13Z,16Z)) | ||||||
| Model 1 | −0.05 (−0.08 to −0.01) | −2.45 (−6.21 to 1.31) | 1.08 (0.16 to 2.01) | 0.11 (−0.09 to 0.38) | 0.05 (−0.13 to 0.36) | 0.12 (0.06 to 0.21) |
| Model 2 | −0.35 (−0.63 to −0.77) | −0.11 (−0.61 to 0.40) | 1.07 (0.12 to 2.03) | 0.04 (−0.20 to 0.32) | 0.03 (−0.88 to 0.27) | 0.14 (−0.01 to 0.31) |
Mediation analysis was applied to identify the four changed metabolites as mediators in the association between famine and diabetes.
Bold indicates p<0.05.
The results are expressed as a log-odds metric except for the total effect (c), which is expressed as OR (95% CI).
Model 1 was unadjusted. Model 2 was adjusted for sex and body mass index.
PC, phosphatidylcholine.