n | Hypoglycemia | Normoglycemia | Estimated rate ratio | 95% CI | ||||
---|---|---|---|---|---|---|---|---|

End point | Geometric mean (CV) | Range | Geometric mean (CV) | Range | P value | |||

Number of arousals | ||||||||

0−4 h | 20 | 31 (88) | 3–145 | 34 (67) | 4–132 | 0.92 | 0.64; 1.33 | 0.6547 |

4−8 h | 20 | 20 (113) | 1–183 | 30 (61) | 2–99 | 0.81 | 0.53; 1.24 | 0.3070 |

0−8 h | 20 | 54 (96) | 4–328 | 66 (62) | 6–231 | 0.86 | 0.59; 1.24 | 0.3955 |

Number of awakenings | ||||||||

0−4 h | 20 | 14 (44) | 3–30 | 15 (37) | 5–26 | 0.87 | 0.72; 1.06 | 0.1694 |

4−8 h | 20 | 10 (54) | 2–26 | 14 (51) | 6–31 | 0.73 | 0.56; 0.95 | 0.0237^{*} |

0−8 h | 20 | 25 (41) | 11–54 | 30 (39) | 12–55 | 0.80 | 0.68; 0.95 | 0.0112^{*} |

*P*values and 95% CIs are based on the negative binomial regression model using logarithm of total sleep time (h) during the corresponding exposure hours as an offset. The model included type of night and period as fixed effects and subject as a random effect. The analysis was based on the completers analysis set.↵*Statistically significant.