Table 2.
Associations of genetically predicted risk factors with gastroesophageal reflux disease in Mendelian randomization sensitivity analyses
| Exposure | Used SNPs | Cochrane’s Q | Weighted median | MR-Egger | ||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | |||
| Body mass index | 312 | 781 | 1.37 | 1.27, 1.48 | 6.50 × 10–15 | 1.05 | 0.90, 1.23 | 0.529 |
| Waist circumference | 38 | 81 | 1.06 | 0.95, 1.20 | 0.307 | 0.98 | 0.75, 1.29 | 0.906 |
| Waist circumference adjusted for BMI | 62 | 144 | 1.09 | 0.99, 1.20 | 0.082 | 1.47 | 1.03, 2.10 | 0.041 |
| Type 2 diabetes | 278 | 770 | 1.03 | 1.00, 1.07 | 0.055 | 0.98 | 0.93, 1.03 | 0.425 |
| Smoking initiation | 202 | 546 | 1.33 | 1.23, 1.45 | 6.27 × 10–12 | 1.12 | 0.83, 1.51 | 0.454 |
| Alcohol drinking | 71 | 210 | 0.98 | 0.78, 1.22 | 0.839 | 0.92 | 0.61, 1.37 | 0.676 |
| Coffee consumption | 11 | 21 | 1.07 | 0.95, 1.20 | 0.256 | 1.23 | 0.99, 1.53 | 0.100 |
| Exposure | Ppleiotropya | Pdistortion testb | MR-PRESSOc | Contamination mixture | ||||
|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | |||
| Body mass index | 2.57 × 10–6 | 0.847 | 1.48 | 1.39, 1.58 | 2.64 × 10–29 | 1.62 | 1.51, 1.72 | 1.65 × 10–35 |
| Waist circumference | 0.265 | 0.795 | 1.15 | 1.05, 1.26 | 0.006 | 1.14 | 1.05, 1.26 | 0.003 |
| Waist circumference adjusted for BMI | 0.086 | 0.744 | 1.07 | 0.99, 1.15 | 0.115 | 1.05 | 0.98, 1.12 | 0.134 |
| Type 2 diabetes | 6.52 × 10–5 | 0.216 | 1.06 | 1.03, 1.09 | 6.84 × 10–6 | 1.05 | 1.03, 1.07 | 1.80 × 10–7 |
| Smoking initiation | 0.121 | 0.293 | 1.47 | 1.37, 1.58 | 5.33 × 10–22 | 1.57 | 1.43, 1.67 | 2.62 × 10–25 |
| Alcohol drinking | 0.618 | 0.652 | 1.09 | 0.92, 1.30 | 0.310 | 1.13 | 0.90, 1.84 | 0.137 |
| Coffee consumption | 0.131 | 0.077 | 0.98 | 0.85, 1.14 | 0.814 | 1.09 | 1.00, 1.20 | 0.052 |
BMI Body mass index; CI Confidence interval; NA Not available; OR Odds ratio; SNPs Single-nucleotide polymorphisms. Sensitivity analyses could not be performed for caffeine consumption due to few SNPs (< 3 SNPs)
aP values for pleiotropy were p values for MR-Egger intercept test and a p value < 0.05 indicates a statistically significant pleiotropic effect
bP values for distortion were obtained from MR-PRESSO test and a p value < 0.05 indicates a statistically significant difference between estimates before and after outlier removal. P of distortion test was not available for the analysis of coffee consumption due to no outlier detected
cThere were 6 outliers detected in MR-PRESSO analysis of BMI, 2 in waist circumference, 4 in waist circumference adjusted for BMI, 10 in type 2 diabetes, 7 in smoking initiation, 4 in alcohol drinking, and 1 in coffee consumption