Table 2.
Overall day-and-night glucose control during closed-loop and control periods based on sensor glucose measurements
| Closed-loop period (n=29) | Control period (n=28) | Paired difference*or paired ratio†(95% CI) | p value | ||
|---|---|---|---|---|---|
| Proportion of time with glucose concentration in range (%) | |||||
| 3·9–10·0 mmol/L‡ | 76·2% (6·4) | 65·6% (8·1) | 10·5 (7·6 to 13·4) | <0·0001 | |
| >10·0 mmol/L | 20·4% (6·3) | 27·4% (9·6) | −6·9 (−10·2 to −3·5) | 0·0003 | |
| >13·9 mmol/L | 3·8% (2·6) | 6·9% (3·9) | −3·0 (−4·4 to −1·6) | 0·0002 | |
| >16·7 mmol/L | 0·9% (0·8) | 2·1% (1·8) | −1·2 (−1·9 to −0·6) | 0·0009 | |
| <3·9 mmol/L | 2·9% (2·3 to 4·0) | 5·3% (3·5 to 10·0) | 0·50 (0·41 to 0·63)† | <0·0001 | |
| <3·5 mmol/L | 1·3% (0·8 to 2·3) | 3·4% (1·9 to 7·2) | 0·35 (0·26 to 0·47)† | <0·0001 | |
| <3·3 mmol/L | 0·9% (0·5 to 1·7) | 2·6% (1·3 to 5·5) | 0·30 (0·22 to 0·43)† | <0·0001 | |
| <2·8 mmol/L | 0·3% (0·1 to 0·5) | 1·0% (0·5 to 2·6) | 0·24 (0·14 to 0·41)† | <0·0001 | |
| AUCday <3·5 mmol/L (min × mmol/L) | 9·1 (3·7 to 18·2) | 26·7 (13·1 to 65·5) | 0·27 (0·18 to 0·41)† | <0·0001 | |
| Mean glucose concentration (mmol/L) | 7·9 (0·5) | 8·3 (0·9) | −0·4 (−0·7 to −0·1) | 0·0226 | |
| SD of glucose concentration (mmol/L) | 2·8 (0·4) | 3·3 (0·5) | −0·5 (−0·7 to −0·3) | <0·0001 | |
| Coefficient of variation of glucose concentration | |||||
| Within days (%) | 35·3 (3·0) | 40·3 (5·1) | −5·0 (−7·1 to −3·0) | <0·0001 | |
| Between days (%) | 12·8 (3·3) | 20·2 (4·6) | −7·5 (−9·7 to −5·3) | <0·0001 | |
| Sensor glucose concentration <3·5 mmol/L for at least 20 min | |||||
| Number of nights | 2·1 (1·0) | 5·6 (3·5) | −3·6 (−4·9 to −2·2) | <0·0001 | |
| Mean duration of each episode (min) | 45 (13) | 75 (25) | −29 (−38 to −20) | <0·0001 | |
Data are mean (SD) or median (IQR), unless otherwise stated. No significant period effect was observed. AUCday=sensor glucose area under the curve per day.
Unless specified otherwise, data are normally distributed and presented as mean difference of closed-loop period minus control period, with 95% CI for mean; a positive value indicates that the measurement was higher during the closed-loop period than in the control period.
Non-normally distributed data are presented as ratio of closed-loop data over control data, with 95% CI for ratio; a value greater than unity indicates that the measurement was higher in the closed-loop period than in the control period.
Primary outcome.