Table 2.

Summary of findings.

Hyperoncotic albumin compared to non-albumin containing fluids in the management of sepsis
Patient or population: the management of sepsis Setting: hospitalised adults with a diagnosis of sepsis Intervention: hyperoncotic albumin Comparison: non-albumin containing fluids
Outcomes	Anticipated absolute effects * (95% CI)		Relative effect (95% CI)	No. of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with non-albumin containing fluids	Risk with hyperoncotic albumin
Mortality in patients with sepsis follow-up: 28 days	314 per 1000	303 per 1000 (268–339)	OR 0.95 (0.80–1.12)	2733 (4 RCTs)	⨁⨁◯◯	The evidence suggests hyperoncotic albumin results in little to no difference in mortality in patients with sepsis.
					Low a
Mortality in patients with septic shock follow-up: 28 days	413 per 1000	365 per 1000 (323–408)	OR 0.82 (0.68–0.98)	2013 (3 RCTs)	⨁⨁⨁◯	Hyperoncotic albumin probably results in a reduction in mortality in patients with septic shock.
					Moderate b
Need for renal replacement therapy follow-up: 90 days	235 per 1000	219 per 1000 (122–363)	OR 0.91 (0.45–1.85)	1910 (2 RCTs)	⨁⨁◯◯	Hyperoncotic albumin may result in little to no difference in need for renal replacement therapy. There is no evidence linking hyperoncotic albumin to acute kidney injury.
					Low c
Cumulative fluid balance assessed with: litres follow-up: range 1 days to 7 days	The mean cumulative fluid balance was 0 L	MD 1.96 L lower (3.81 to 0.11 lower)	-	1965 (3 RCTs)	⨁◯◯◯	Hyperoncotic albumin may reduce cumulative fluid balance in patients with early sepsis.
					Very low d
Duration of mechanical ventilation assessed with: days	Studies measured dependency on mechanical ventilation differently: one reported median [IQR] duration of mechanical ventilation in days, another reported mean (±SD) duration and another simply the number (%) who required mechnical ventilation in each group. Data were not suitable for pooling, but comparisons in individual studies were similar between groups.			1941 (3 RCTs)	⨁⨁◯◯	Hyperoncotic albumin was not associated with any difference in the use or duration of mechanical ventilation.
					Low e
Duration of shock assessed with: various	Included studies measured duration of shock differently: one measured the number of days until vasoprerssors were suspended, another used the number of patients for whom shock was reversed at 6, 12, 24 and 48 h. A third study stated that patients receiving albumin had a higher number of days free from receiving catecholamines, but there were no evaluable data reported.			2687 (3 RCTs)	⨁⨁◯◯	The use of hyperoncotic albumin may result in faster resolution of shock and shorter duration of cardiovascular support
					Low f

CI: confidence interval; MD: mean difference; OR: odds ratio.

GRADE Working Group grades of evidence

High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different.

Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect.

Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Explanations

^*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

^aAll studies at some or high risk of bias around blinding. Comparator fluid in Dolecek et al. ²⁸ was 6% starch, now shown to be harmful in sepsis. Point estimates of smaller studies suggest the largest treatment effect.

^bAll studies at risk of bias due to lack of blinding.

^cSmallest study suggests the largest effect size; heterogeneity in populations and duration of intervention between studies.

^dCFB measured at different time points across included studies (ranging from 1 to 7 days). CFB also vulnerable to measurement error and reliance on medical record keeping. High statistical heterogeneity (I² = 95%). 95% CI either large (Dolecek et al. ²⁸ ) or not overlapping (Caironi et al. ²⁶ and Maiwall et al. ²⁹ ).

^eStudies used different methods to measure the use and duration of mechanical ventilation and were all at risk of bias to lack of blinding.

^fStudies used different methods to measure reversal of shock and dependence on cardiovascular support. Studies with longest duration of intervention reported greatest effects, indicating a possible dose-response relationship.