Skip to main content
. 2012 Jun;21(3):223–242. doi: 10.1177/0962280210394459

Table 6.

IV estimates of the effect of fat mass on bone mineral density (BMD) using all available dataa

SNPs used as instrumental variable N First stage regression coefficient (95% CI) First stage R2 First stage F-statistic Ratio of geometric mean BMDb (95% CI) SE of estimate (log scale) Hausman test p-value Sargan test p-value
OLS 5509 NA NA NA 1.22 (1.18, 1.25), p < 0.001 0.014 NA NA
IV: SNP(s) used as IV
FTO 5091 0.12 (0.08, 0.15) 0.0088 45.35 1.41 (1.05, 1.89), p = 0.023 0.15 0.320 NA
MC4R 5412 0.09 (0.05, 0.13) 0.0037 19.95 2.42 (1.42, 4.12), p = 0.001 0.27 0.002 NA
TMEM18 5323 −0.06 (−0.11, −0.02) 0.0013 6.99 2.17 (0.92, 5.12), p = 0.077 0.44 0.130 NA
GNPDA2 5303 0.05 (0.01, 0.08) 0.0013 6.90 0.92 (0.42, 2.01), p = 0.84 0.40 0.463 NA
FTO, MC4R 5007 NA 0.0125 31.61 1.60 (1.24, 2.07), p < 0.001 0.13 0.029 0.221
FTO, MC4R, TMEM18 4881 NA 0.0138 22.75 1.69 (1.32, 2.17), p < 0.001 0.13 0.006 0.227
a

Analyses adjusted for height and height squared.

b

For a 1 unit increase in z-score of age and gender standardised fat mass.