Table 3. Association between inv-8p23 and low BMI in children.
| BMI > 5th-perc. | BMI < 5th-perc. | OR | 95% CI | P-value | ||
|---|---|---|---|---|---|---|
| SHARP | ALL | |||||
| Recessive | ||||||
| N/N-N/I | 68.7% | 57.1% | 1 | |||
| I/I | 31.3% | 42.9% | 1.67 | (1.03, 2.70) | 0.03 | |
| European haplotype ancestry | ||||||
| (inferred ancestry group 1) | ||||||
| Recessive | ||||||
| N/N-N/I | 75.1% | 60% | 1 | |||
| I/I | 24.9% | 40% | 1.92 | (1.11, 3.35) | 0.02* | |
| European-American | ||||||
| (self-reported) | ||||||
| Recessive | ||||||
| N/N-N/I | 68% | 55.4% | 1 | |||
| I/I | 32% | 44.6% | 1.71 | (0.99, 2.95) | 0.05 | |
| INMA | ALL | |||||
| Recessive | ||||||
| N/N-N/I | 66% | 50% | 1 | |||
| I/I | 34% | 50% | 2.10 | (1.09, 4.08) | 0.02 |
Top: association analyses between inv-8p23 genotypes and low BMI in the SHARP sample, adjusted for the first two genome-wide PCA components. While individuals with European haplotype ancestry have a significant risk for low BMI, self-reported European-Americans do not. The P-value with an asterisk is significant after correction for multiple genetic models. Bottom: association analysis between inv-8p23 and low BMI in the INMA sample. Combined effects between the two cohorts resulted in OR = 2.00, P-value = 0.001 and population attributable risk of 2.8%.