Thirty RCTs (n=5,733) were included.
The mean quality score was 9.1 (95% CI: 7, 12.7).
Haemodynamic optimisation interventions were associated with a significant reduction in mortality compared with the control; the OR was 0.61 (95% CI: 0.46, 0.81), the RR 0.75 (95% CI: 0.62, 0.90) and the ARR 0.4% (95% CI: -1.7, 2.6). Statistically significant heterogeneity was detected (p=0.0003). There was no significant correlation between individual study ORs and quality scores (p=0.07).
Interventions to optimise a haemodynamic condition peri-operatively were associated with a significant reduction in mortality compared with the control; the OR was 0.43 (95% CI: 0.28, 0.66) and the RR 0.66 (95% CI: 0.54, 0.81). Statistically significant heterogeneity was detected (p=0.08). The NNT was 31 (95% CI: 20, 63). Analyses were based on 21 RCTs (n=4,174). There was no significant correlation between individual study ORs and quality scores (p=0.3).
For patients with sepsis and/or organ failure, there was no significant difference in morality between optimisation interventions and controls (10 RCTs, n=1,558); the OR was 0.85 (95% CI: 0.58, 1.25) and the RR 0.92 (95% CI: 0.75, 1.11). Statistically significant heterogeneity was detected (p=0.02). There was no significant correlation between individual study ORs and quality scores (p=0.4).
High-quality studies had higher RR (RR 0.0.84, 95% CI: 0.66, 1.07) than lower-quality studies (RR 0.60, 95% CI: 0.48, 0.75).
The results of all other subgroup analyses were also reported.