Table 2.

Evolution of fluid balance between groups

	Control group N = 89	Intervention group N = 77	Mean difference 95% CI	p value
Primary analysis
Multiple imputation (N = 166)			− 4.9 [− 7.4;− 2.5]	< 0.001
Sensitivity analyzes
Complete cases (N = 144, 86.7%)	6.4 [5–11.2] kg	1.4 [1–4.5] kg	− 5.1 CI95 [− 7.4;− 2.8]	< 0.001
Worst case imputation (N = 166)	4.5 [− 1.5;10.5] kg	1.5 [− 2;7.7] kg	− 0.5 [− 3.2;2.3]	0.7
Last observation carried forward (N = 166)	7 [2;15.5] kg	1.5 [− 2;7.7] kg	− 8.4 [− 16.6;− 0.2]	0.047
Per protocol (N = 135, 81.3%)	5 [0.5;11.3] kg	1 [− 2.0;4.6] kg	− 4.7 [− 7.1;− 2.3]	< 0.001

Evolution of fluid balance between successful extubation and randomization. A multiple imputation was performed (SAPS II, acute renal failure, chronic diuretic therapy, chronic renal failure, heart failure, admission to ICU for septic shock and admission to ICU for acute respiratory failure). This table displays the median [quartile1–quartile3] value of measured body weight in kilograms, between the two groups (this data is not available for multiple imputation, which uses multiple imputed datasets) and generalized linear mixed models results (estimated mean difference and its confidence interval)