Loop diuretics in adult intensive care patients with fluid overload: a systematic review of randomised clinical trials with meta-analysis and trial sequential analysis

Wichmann, Sine; Barbateskovic, Marija; Liang, Ning; Itenov, Theis Skovsgaard; Berthelsen, Rasmus Ehrenfried; Lindschou, Jane; Perner, Anders; Gluud, Christian; Bestle, Morten Heiberg

doi:10.1186/s13613-022-01024-6

Annals of Intensive Care

Table 2 Summary of Findings with GRADE evaluation for loop diuretics vs. placebo/no intervention

From: Loop diuretics in adult intensive care patients with fluid overload: a systematic review of randomised clinical trials with meta-analysis and trial sequential analysis

Certainty assessment							No. of patients		Effect		Certainty	Importance
No. of studies	Study design	Risk of Bias	Inconsistency	Indirectness	Imprecission	Other considerations	Loop diuretics	Placebo/no intervention	Relative (95% CI)	Absolute (95% CI)
All-cause mortality
4	RCT	Serious^a	Not serious	Serious^b	Very serious^c	None	33/171 (19.3%)	51/188 (27.1%)	RR 0.72 (0.49–1.06)	76 fewer per 1.000 (from 138 fewer to 16 more)	⨁◯◯◯ VERY LOW	CRITICAL
Quality of life—not reported
–	–	–	–	–	–	–	–	–	–	–	–	CRITICAL
Serious adverse events (SAE)
6	RCT	Serious^a	Not serious	Serious^b	Serious^d	None	87/230 (37.8%)	116/246 (47.2%)	RR 0.81 (0.66–0.99)	90 fewer per 1.000 (from 160 to 5 fewer)	⨁◯◯◯ VERY LOW	CRITICAL
Plasma concentration of creatinine
3^e	RCT	Serious^a	Not serious	Not serious	Serious^f	None	0/0	0/0	The trials showed no significant difference in plasma creatinine at longest follow-up		⨁⨁◯◯ LOW	CRITICAL
Proportion of participants without resolution of fluid overload
2	RCT	Serious^a	Not serious	Very serious^g	Very serious^h	None	4/41 (9.8%)	24/51 (47.1%)	RR 0.22 (0.08–0.58)	367 fewer per 1.000 (from 433 to 198 fewer)	⨁◯◯◯ VERY LOW	CRITICAL
Days on mechanical ventilation
2ⁱ	RCT	Serious^a	Not serious	Serious^b	Serious^j	None	0/0	0/0	The trials showed no significant difference in days in mechanical ventilation		⨁◯◯◯ VERY LOW	IMPORTANT
Length of stay in ICU
2ⁱ	RCT	Serious^a	Not serious	Serious^k	Serious^j	None	0/0	0/0	The trials showed no significant difference in length of stay in the ICU		⨁◯◯◯ VERY LOW	IMPORTANT
Adverse event not considered serious (AE)
2	RCT	Serious^a	Not serious	Serious^l	Serious^m	None	71/120 (59.7%)	61/125 (48.8%)	RR 1.23 (0.98–1.55)	112 more per 1.000 (from 10 fewer to 268 more)	⨁◯◯◯ VERY LOW	NOT IMPORTANT

RCT randomised clinical trials, CI Confidence interval, RR Risk ratio
^aAll trials were at overall high risk of bias for this outcome
^bVariations in experimental intervention and control groups
^cTSA showed lack of data, because only 11.5% of optimal information size had been reached
^dTSA showed lack of data, because only 34.7% of optimal information size had been reached
^eThree trials reported on plasma creatinine. A meta-analysis could not be performed because of unsuitable data (medians and interquartile range or only graphical presentation of data)
^fThe total number of participants were only 193 in the included trials, which is concerning for imprecision
^gDifferences in ICU subpopulation (AKI vs. decompensated heart failure). Resolution of fluid overload is a surrogate outcome
^hTSA showed lack of data. Only 6.2% of optimal information size had been reached
ⁱData was not suitable for meta-analysis. Both trials found no difference between groups
^jThe total number of participants were only 186 in the included trials, which is concerning for imprecision
^kThe two trials were dissimilar regarding ICU population, control group and length of stay in the ICU
^lDifferences in ICU subpopulation (AKI patients vs. mixed population)
^mThe total number of participants were only 244 in the included trials, which is concerning for imprecision

Back to article page