Table 2. Candidate predictors for development of prediction model for early prediction of Intensive Care Unit – acquired weakness.
| candidate predictors* | distribution | p-value | selection percentage inbootstrap samples | |
| ICU-AW(N:103) | no ICU-AW(N:109) | |||
| patient characteristics | ||||
| females, n (%) | 52 (50) | 40 (37) | 0.06 | 37.1 |
| age, mean ± SD | 63±15 | 59±16 | 0.08 | 57.6 |
| risk factor for a polyneuropathyin medical history, n (%) | 35 (34) | 40 (37) | 0.79 | 13.4 |
| pre-existing polyneuropathy priorto ICU admission, n (%) | 3 (3) | 1 (1) | 0.57 | n.a. |
| systemic corticosteroid use priorto ICU admission, n (%) | 7 (7) | 9 (8) | 0.89 | 10.7 |
| clinical parameters | ||||
| suspected sepsis, n (%) | 78 (76) | 70 (64) | 0.09 | 14.7 |
| unplanned admission, n (%) | 85 (83) | 83 (76) | 0.33 | 10.9 |
| presence of shock, n (%) | 75 (73) | 67 (61) | 0.11 | 24.6 |
| RASS score, median (IQR) | –3 (–5 to −1) | –2 (–3 to 0) | <0.01 | 48.2 |
| laboratory parameters | ||||
| average urine production,median ml/h (IQR) | 70 (20 to 122) | 102(64 to 134) | <0.01 | 14.4 |
| highest glucose, mean mg/dl ± SD | 243.8±78.5 | 220.5±67.3 | 0.02 | 38.5 |
| lowest glucose, mean mg/dl ± SD | 85.8±22.3 | 89.6±25.8 | 0.25 | 22.2 |
| lowest pH, mean ± SD | 7.21±0.1 | 7.25±0.1 | 0.02 | 17.3 |
| lowest P/F ratio, median (IQR) | 186 (127 to 245) | 178 (134 to 246) | 0.98 | 27.9 |
| lowest platelet count,median×109/L (IQR) | 103 (45 to 151) | 127 (85 to 197) | 0.01 | 21.0 |
| highest lactate, median mmol/L(IQR; 17 missing) | 4.5 (3.0 to 7.0) | 2.8 (1.7 to 4.8) | <0.01 | 89.5‡ |
| lowest ionized Ca2+, meanmmol/L ± SD | 0.97±0.11 | 0.98±0.13 | 0.53 | 51.6 |
| medication | ||||
| treatment with any corticosteroid,n (%) | 81 (79) | 63 (58) | <0.01 | 33.9 |
| repeated treatment with anyneuromuscular blocker§, n (%) | 17 (17) | 18 (17) | 1.00 | 20.3 |
| treatment with anyaminoglycoside, n (%) | 51 (50) | 30 (28) | <0.01 | 80.4 |
*all clinical, laboratory and medication parameters were scored using information from the first two ICU days, except for the RASS score which was scored around two days after ICU admission;
‡
logarithmically transformed;
§
more than one administration of any neuromuscular blocker.
ICU-AW: Intensive Care Unit – acquired weakness; RASS: Richmond Agitation and Sedation Scale; n.a.: not applicable.
Table displaying distributions and differences between patients with and without Intensive Care Unit – acquired weakness for the candidate predictors. In the final column selection percentages of the candidate predictors in bootstrap samples based on backward selection are presented.