Introduction
Methods
Patients and data
Variables | Value | |
---|---|---|
Cool | Warm | |
Total patients, number (n) | 200 | |
Median age, years | 61 [51, 72] | |
Female gender, number (%) | 40 (20.6%) | |
ICU mortality, number (%) | 85 (45.6%) | |
Diabetes status, number (%) | 26 (13.0%) | |
Total treatment, hours (h) | 4219 | 4303 |
Blood glucose, median (mmol/L) | 7.6 [6.3,9.7] | 6.8 [5.9,8.0] |
Insulin rate, median (U/hr) | 3.4 [1.3,8.0] | 3.5 [1.6,7.0] |
Glucose rate, median (g/hr) | 2.7 [1.0,5.3] | 5.4 [2.7,8.1] |
Model-based insulin sensitivity
Analyses and metrics
Day | Period | Analysis | Block | Hour range |
---|---|---|---|---|
1 | Cool | 6-hour block | 1 | 0 – 6 hours |
2 | 6 – 12 hours | |||
3 | 12 – 18 hours | |||
4 | 18 – 24 hours | |||
Idle 2-hour period in between cool and warm | ||||
2 | Warm | 6-hour block | 5 | 24 – 30 hours |
6 | 30 – 36 hours | |||
7 | 36 – 42 hours | |||
8 | 42 – 48 hours |
Results
SIlevel analyses
SIlevel | Cohort analysis | Per-patient analysis | ||
---|---|---|---|---|
analysis | % SImedian increase | Pvalue | % SImedian increase | Pvalue |
(6-hr blocks) | ||||
Block 1-2 (C) | 35.1 | <0.05 | 26.4 | <0.05 |
(0–6 vs. 6–12 hr)
| ||||
Block 2–3 (C) | 19.2 | <0.05 | 31.1 | <0.05 |
(6–12 vs. 12–18 hr)
| ||||
Block 3-4 (C) | 31.8 | <0.05 | 42.4 | <0.05 |
(12-18 vs. 18-24 hr)
| ||||
Block 4-5 (C-W) | 23.4 | <0.05 | 18.3 | <0.05 |
(18-24 vs. 24-30 hr)
| ||||
Block 5-6 (W) | 23.9 | <0.05 | 23.2 | <0.05 |
(24-30 vs. 30-36 hr)
| ||||
Block 6-7 (W) | 13.1 | 0.06 | 15.8 | 0.2 |
(30-36 vs. 36-42 hr)
| ||||
Block 7-8 (W) | 4.4 | 0.4 | 3.2 | 0.5 |
(36-42 vs. 42-48 hr)
|
6 -12 hr | 12-18 hr | 18-24 hr | 24-30 hr | 30-36 hr | 36-42 hr | 42-48 hr | |
---|---|---|---|---|---|---|---|
0 - 6 hr | 0.72 | 0.74 | 0.79 | 0.83 | 0.84 | 0.85 | 0.86 |
6 - 12 hr | 0.66 | 0.72 | 0.74 | 0.76 | 0.82 | 0.82 | |
12 - 18 hr | 0.69 | 0.70 | 0.75 | 0.79 | 0.79 | ||
18 - 24 hr | 0.66 | 0.65 | 0.70 | 0.72 | |||
24 - 30 hr | 0.64 | 0.68 | 0.66 | ||||
30 - 36 hr | 0.58 | 0.61 | |||||
36 - 42 hr | 0.52 |
SIvariability analyses
SIvariability | Cohort analysis | Per-patient analysis | ||
---|---|---|---|---|
analysis | % reduction of IQR | Pvalue | % median decrease | Pvalue |
[6-hr blocks] | ||||
Block 1-2 (C) | 11.1 | <0.05 | 33.6 | <0.05 |
(0-6 vs. 6-12 hr)
| ||||
Block 2-3 (C) | 20.7 | <0.05 | 15.8 | <0.05 |
(6-12 vs. 12-18 hr)
| ||||
Block 3-4 (C) | 14.4 | <0.05 | 22.6 | <0.05 |
(12-18 vs. 18-24 hr)
| ||||
Block 4-5 (C-W) | −19.7 | <0.05 | −14.9 | <0.05 |
(18-24 vs. 24-30 hr)
| ||||
Block 5-6 (W) | 23.1 | <0.05 | 26.4 | 0.05 |
(24-30 vs. 30-36 hr)
| ||||
Block 6-7 (W) | 4.6 | <0.05 | 0.8 | 0.05 |
(30-36 vs. 36-42 hr)
| ||||
Block 7-8 (W) | 13.0 | 0.08 | 17.1 | 0.06 |
(36-42 vs. 42-48 hr)
|
Discussion
Insulin sensitivity level
Insulin sensitivity variability
Implications for glycaemic control
Limitations
Conclusions
Key messages
-
OCHA patients treated with TH have significantly lower and highly variable SI during the first 24 hours of the cool period, compared to the later warm period in their ICU stay.
-
There is an overall trend of increasing SI over the first 36 hours, both per-cohort and per-patient results.
-
SI variability decreases consistently over time, except for a large, statistically significant increase during the cool to warm transition at 24 hours.
-
This increase requires special consideration for glycaemic control as it increases risk of hypoglycaemia, BG variability and thus mortality.