Table 3.
Descriptive statistics and univariate analyses for 14 candidate covariates for 4945 fecal samples
| Covariates | All samples (n = 4945) | Univariate analysis p valuea | |||
|---|---|---|---|---|---|
| Continuous variable | Mean ± SD | Frequency (%) of missing | Mean age of O. formigenes-positive/O. formigenes-negative | Logistic | Negative binomial |
| Age | 46.08 ± 17.28 | 246 (4.97) | 48.39:44.78 | <.001 | <.001 |
| Categorical covariate | Frequency (%) | Frequency (%) of missing | Frequency (%) of O. formigenes-positive | Logistic | Negative binomial |
| Sex | 4722 (95.49) | 223 (4.51) | |||
| Female | 2576 (52.09) | 897 (34.84) | Reference | ||
| Male | 2146 (43.40) | 775 (36.11) | 0.791 | 0.224 | |
| Race | 4867 (98.42) | 78 (1.58) | |||
| Caucasian | 4369 (88.35) | 1602 (36.67) | Reference | ||
| Asian/Pacific Islander | 232 (4.69) | 40 (17.24) | <.001 | <.001 | |
| Hispanic | 86 (1.74) | 28 (32.56) | 0.315 | 0.082 | |
| African American | 50 (1.01) | 9 (18.00) | 0.012 | 0.075 | |
| Other | 130 (2.63) | 41 (31.54) | 0.178 | 0.674 | |
| BMI | 4620 (93.43) | 325 (6.57) | |||
| Underweight | 453 (9.16) | 113 (24.94) | <.001 | <.001 | |
| Normal | 2720 (55.01) | 1093 (40.18) | Reference | ||
| Overweight | 985 (19.92) | 347 (35.23) | 0.003 | <.001 | |
| Obese | 462 (9.34) | 114 (24.68) | <.001 | <.001 | |
| Frequency of alcohol consumption | 4855 (98.18) | 90 (1.82) | |||
| Never | 1146 (23.17) | 317 (27.66) | Reference | ||
| Rarely (a few times a month) | 1228 (24.83) | 447 (36.40) | <.001 | 0.038 | |
| Occasionally (1–2 times/week) | 1076 (21.76) | 400 (37.17) | <.001 | 0.055 | |
| Regularly (3–5 times/week) | 888 (17.96) | 343 (38.63) | <.001 | 0.098 | |
| Daily | 517 (10.46) | 213 (41.20) | <.001 | 0.192 | |
| Last exposure to antibiotics | 4822 (97.51) | 123 (2.49) | |||
| >365 days | 3281 (66.35) | 1286 (39.20) | Reference | ||
| <365 days | 717 (14.50) | 223 (31.10) | <.001 | 0.018 | |
| <180 days | 584 (11.81) | 136 (23.19) | <.001 | <.001 | |
| <30 days | 149 (3.01) | 44 (29.53) | 0.070 | 0.317 | |
| <7 days | 91 (1.84) | 23 (25.27) | 0.002 | 0.369 | |
| Presence of appendix | 4784 (96.74) | 161 (3.26) | |||
| No | 4297 (86.90) | 1537 (35.77) | Reference | ||
| Yes | 487 (9.85) | 160 (32.85) | 0.170 | 0.347 | |
| Country | 4945 (100.00) | 0 (0) | |||
| USA | 3779 (76.42) | 1176 (31.12) | Reference | ||
| UK–Ireland | 819 (16.56) | 411 (50.18) | <.001 | <.001 | |
| Europe Continental | 111 (2.24) | 52 (46.85) | 0.001 | 0.088 | |
| Australia and NZL | 137 (2.77) | 71 (51.82) | <.001 | 0.030 | |
| Canada | 73 (1.48) | 31 (42.47) | 0.099 | 0.442 | |
| Others | 26 (0.53) | 13 (50.00) | 0.099 | 0.656 | |
| Drinking water source | 4821 (97.49) | 124 (2.51) | |||
| City | 2283 (46.17) | 855 (37.45) | Reference | ||
| Filtered | 1709 (34.56) | 587 (34.35) | 0.169 | 0.389 | |
| Bottled | 426 (8.61) | 126 (29.58) | 0.002 | 0.524 | |
| Well | 403 (8.15) | 136 (33.75) | 0.366 | 0.928 | |
| Level of education | 2791 (56.44) | 2154 (43.56) | |||
| ≤High school | 245 (4.95) | 72 (29.39) | 0.007 | 0.146 | |
| College/bachelor’s | 1024 (20.71) | 405 (39.55) | Reference | ||
| Graduate school | 1522 (30.78) | 613 (40.28) | 0.418 | 0.514 | |
| Dog | 4834 (97.76) | 111 (2.24) | |||
| Absent | 3362 (67.99) | 1219 (36.26) | Reference | ||
| Present | 1472 (29.77) | 489 (33.22) | 0.058 | 0.118 | |
| Born by C-section | 4626 (93.55) | 319 (6.45) | |||
| False | 4147 (83.86) | 1485 (35.81) | Reference | ||
| True | 479 (9.69) | 152 (31.73) | 0.047 | 0.584 | |
| Vegetable consumption frequency | 2876 (58.16) | 2069 (41.84) | |||
| <1 time/week | 127 (2.57) | 33 (25.98) | 0.017 | 0.034 | |
| 1–2 times/week | 288 (5.82) | 84 (29.17) | 0.024 | 0.046 | |
| 3–5 times/week | 1025 (20.73) | 381 (37.17) | Reference | ||
| Daily | 1436 (29.04) | 616 (42.90) | 0.001 | 0.007 | |
| Thyroid disease | 2853 (57.69) | 2092 (42.31) | |||
| No condition | 2490 (50.35) | 986(39.60) | Reference | ||
| Diagnosed | 363 (7.34) | 120(33.06) | 0.001 | 0.070 | |
a Univariate analyses: The p values reported in the table were estimated based on logistic regression models or negative binomial regression models for the effect of each candidate covariate on abundance of O. formigenes. Both logistic or negative binomial regression models include the log of total read count per sample as the offset variable. For the purpose of univariate analysis, we consider p<0.1 as statistically significant as stated in the 'Methods', which is shown in italics