Table 3.
Multivariate regression models of anthropometric and metabolic parameters with regular consumers as the referent group.
| Overall † | <-------------------- Skipper --------------------> | <-------------------- Intermittent --------------------> | |||||
|---|---|---|---|---|---|---|---|
| Variable | p | β | (95% CI) | p | β | (95% CI) | p |
| Anthropometric Parameters (n = 671) | |||||||
| Waist circumference (cm) a | 0.09 | 0.59 | (−1.98, 3.16) | 0.62 | 2.09 | (0.08, 4.10) | 0.03 |
| Total body fat (%) a | 0.39 | −0.05 | (−1.95, 1.86) | 0.76 | 0.82 | (−0.67, 2.31) | 0.17 |
| BMI z-score a | 0.12 | 0.12 | (−0.12, 0.36) | 0.33 | 0.19 | (0.01, 0.38) | 0.04 |
| Systolic blood pressure (mmHg) a | 0.58 | −0.10 | (−2.51, 2.31) | 0.99 | 0.84 | (−1.05, 2.73) | 0.33 |
| Diastolic blood pressure (mmHg) a | 0.72 | 0.63 | (−1.46, 2.72) | 0.64 | 0.43 | (−1.21, 2.06) | 0.43 |
| Metabolic Parameters (n = 344) | |||||||
| Fasting glucose (mg/dL) b,c,d | 0.89 | 0.47 | (−2.23, 3.17) | 0.73 | −0.21 | (−2.42, 2.01) | 0.85 |
| Insulin (µIU/mL) b,c,d | 0.67 | 0.89 | (−2.29, 4.08) | 0.39 | 1.26 | (−1.35, 3.87) | 0.90 |
| Triglycerides (mg/dL) b,c,d | 0.37 | 4.14 | (−7.46, 15.74) | 0.61 | −5.98 | (−15.49, 3.52) | 0.32 |
| Total cholesterol (mg/dL) b,c,d | 0.95 | 0.42 | (−6.82, 7.66) | 0.91 | 0.93 | (−5.00, 6.87) | 0.76 |
| HDL (mg/dL) c | 0.96 | −0.30 | (−2.95, 2.36) | 0.89 | −0.31 | (−2.49, 1.87) | 0.77 |
| Non-HDL (mg/dL) c | 0.91 | 0.69 | (−5.86, 7.24) | 0.84 | 1.18 | (−4.19, 6.54) | 0.67 |
| LDL (mg/dL) c | 0.54 | −0.15 | (−6.05, 5.74) | 0.96 | 2.45 | (−2.38, 7.28) | 0.32 |
| HbA1c (%) d | 0.49 | −0.09 | (−0.15, 0.09) | 0.50 | 0.04 | (−0.06, 0.15) | 0.53 |
Abbreviations: BMI, body mass index; HbA1c, glycolated hemoglobin A1C; HDL, high-density lipoprotein; LDL, low-density lipoprotein. Multivariate regression assessed relationships between anthropometric parameters (n = 671) and metabolic parameters (n = 344) by breakfast consumption groups. Regular consumers served as the referent group for all analyses. † Overall effect of breakfast consumption groups. a Model included waist circumference, total body fat, BMI z-score, and blood pressure. b Model included fasting glucose, insulin, total cholesterol, and triglycerides. c Model included fasting glucose, insulin, triglycerides, HDL, non-HDL, and LDL. d HbA1c was assessed only in the last two waves of the intervention, indicative of the smaller sample size (n = 237), so it was assessed in an independent model including fasting glucose, insulin, total cholesterol, and triglycerides. A priori covariates included: age, sex, race/ethnicity, free/reduced lunch status, daily energy, and BMI z-score (for metabolic parameters).