Table 4.
Associations between Change in Serum Creatinine from Baseline and Delirium or Coma
Delirium |
Coma |
|||
---|---|---|---|---|
OR | 95% CI | OR | 95% CI | |
Clinical condition of interest | ||||
Associations with peak serum creatinine today according to RRT status (per mg/dl)* | ||||
Not on RRT | 1.35† | 1.18–1.55 | 1.44† | 1.20–1.74 |
On RRT | 1.07 | 0.87–1.31 | 1.16 | 0.94–1.44 |
Regression model predictors (full regression model with interaction term) | ||||
Peak serum creatinine today (per mg/dl)‡ | 1.35† | 1.18–1.55 | 1.44† | 1.20–1.74 |
RRT today§ | 1.61 | 0.64–4.04 | 1.64 | 0.65–4.17 |
Peak creatinine × RRT (interaction term) | 0.79† | 0.63–0.99 | 0.81 | 0.62–1.05 |
Baseline serum creatinine (per mg/dl) | 0.78 | 0.60–1.01 | 0.53† | 0.35–0.80 |
Previous day’s mental status|| | ||||
Delirium | 28.09† | 22.13–35.65 | 16.23† | 10.9–24.16 |
Coma | 48.55† | 32.40–72.76 | 221.35† | 119.57–409.75 |
Age (per yr) | 1.03† | 1.02–1.03 | 1.03† | 1.02–1.04 |
Female | 1.11 | 0.87–1.40 | 0.98 | 0.69–1.38 |
African American | 0.61† | 0.40–0.93 | 0.83 | 0.49–1.40 |
Sepsis | 1.08† | 1.06–1.10 | 1.08† | 1.05–1.11 |
Modified SOFA (per unit) | 1.06† | 1.00–1.12 | 1.23† | 1.15–1.32 |
Previous day’s sedative exposure | ||||
Benzodiazepines (per 1-mg midazolam eq.) | 1.02† | 1.01–1.04 | 1.03† | 1.02–1.05 |
Opiates (per 100-μg fentanyl eq.) | 1.03† | 1.02–1.04 | 1.04† | 1.02–1.05 |
Propofol (per 100 mg) | 1.01 | 0.99–1.02 | 1.01 | 1.0–1.02 |
Definition of abbreviations: CI = confidence interval; OR = odds ratio; RRT = renal replacement therapy; SOFA = Sequential Organ Failure Assessment.
Because of a significant interaction between RRT and peak serum creatinine today, the association between peak creatinine and mental status (normal vs. delirium vs. coma) differed according to the use of RRT. The associations between peak creatinine and mental status (normal vs. delirium vs. coma) are therefore reported here stratified according to whether RRT was present or not and modeled using the regression model predictors below.
Denotes a statistically significant association.
In the full regression model, the regression coefficients (and corresponding ORs) for the associations between peak serum creatinine today and delirium and coma indicate the associations among patients not on RRT.
Because of the significant interaction between RRT and peak serum creatinine today, the association between RRT and mental status (normal vs. delirium vs. coma) differed according to the change in creatinine. Results shown here assume a peak creatinine at the reference value (1 mg/dl).
Compared with a reference (OR of 1.0) of “normal” mental status the previous day.