Table 2.
Regression Models for Predictors of Change in Serum Creatinine, Global VAS, Cystatin C and NT-proBNP at 72 hours
Covariate | Change in Serum Creatinine at 72 h (N=301) | Change in Global VAS at 72 h (N=303) | % change in Serum Cystatin C at 72 h (N=247) | % change in NT-proBNP at 72 h (N=245) |
---|---|---|---|---|
| ||||
P value | P value | P value | P value | |
|
||||
Baseline creatinine | <.0001 (−) | .43 (+) | NA | NA |
Baseline covariate | NA | <.0001 (+) | <.0001 (−) | .40 (−) |
Outpatient furosemide dose | 22 (+) | .84 (+) | .01 (+) | .44 (+) |
Continuous mode of therapy | .40 (+) | .48 (+) | .16 (+) | .85 (−) |
High initial IV diuretic dose | .21 (+) | .06 (+) | .30 (+) | .01 (−) |
Baseline covariate indicates the results of regression using the baseline value of the covariate (e.g., baseline cystatin C level) as a covariate to predict the dependent variable (e.g., change in cystatin C level). The (+) and (−) after the P values signify the directionality of the regression coefficients. NA indicates that the covariate was not tested in the regression model. The interaction terms between furosemide dose and mode and furosemide dose and intensification were tested in a separate model and found to be not significant. They were subsequently not included in the model results presented. The number (N) of patients with data available for each regression model is specified. Abbreviations: NT-proBNP, N-terminal pro-B-type natriuretic peptide; VAS, visual analog scale.