Table 3. Mechanisms by Which Prolonged Nightly Fasting and Related Food Consumption Behaviors Have Been Postulated to Effect Breast Cancer Risk and Prognosisa.
| Mechanism | β (95% CI) | P Value |
|---|---|---|
| Model 1: Glucoregulation, HbA1c, mmol/mol | ||
| Nightly fasting durationb | −0.37 (−0.72 to −0.01) | .04 |
| Eating episodes per day | −0.27 (−0.53 to −0.01) | .045 |
| Eating after 8 pm (yes vs no) | 0.16 (−0.43 to 0.76) | .59 |
| Model 2: Chronic Inflammation, log CRP, mg/L | ||
| Nightly fasting durationb | 0.04 (−0.03 to 0.12) | .27 |
| Eating episodes per day | −0.07 (−0.13 to −0.01) | .01 |
| Eating after 8 pm (yes vs no) | 0.16 (0.04 to 0.29) | .01 |
| Model 3: BMI | ||
| Nightly fasting durationb | 0.04 (−0.32 to 0.39) | .80 |
| Eating episodes per day | −0.48 (−0.73 to −0.23) | <.001 |
| Eating after 8 pm (yes vs no) | 0.64 (0.07 to 1.20) | .03 |
| Model 4: Sleep (Usual Duration), hours per night | ||
| Nightly fasting durationb | 0.20 (0.14 to 0.26) | <.001 |
| Eating episodes per day | 0.09 (0.04 to 0.13) | <.001 |
| Eating after 8 pm (yes vs no) | −0.01 (−0.11 to 0.10) | .93 |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CRP, C-reactive protein; HbA1c, hemoglobin A1c.
a
Models controlled for age, race/ethnicity, educational level, number of comorbidities, tumor stage, grade, radiotherapy, tamoxifen use, kilocalorie intake, menopausal status, study site, and intervention group assignment.
b
A 2-h (approximately 1 SD) unit of analysis was used.