Nine RCTs with a total of 5,143 patients were included in the review.
In terms of study quality: 1 study was double-blinded; 3 studies used independent adjudication of adverse events; all studies used an intention-to-treat analysis; and 2 studies were terminated early due to safety issues.
All-cause mortality (8 studies).
All-cause mortality was significantly higher in the higher haemoglobin target group than in the lower target group (RR 1.17, 95% CI: 1.01, 1.35, p=0.031). No statistically significant heterogeneity was detected (I-squared 27%, p=0.213). A subgroup analysis showed no statistically significant difference between the two haemoglobin target groups in either pre-dialysis patients (RR 1.33, 95% CI: 0.98, 1.81, p=0.067) or those currently undergoing dialysis (RR 1.11, 95% CI: 0.94, 1.31, p=0.22). A sensitivity analysis that included the 5 excluded studies only had a small impact on the results (RR 1.14, 95% CI: 0.99, 1.32, p=0.07).
MI (7 studies).
There was no significant difference in the occurrence of MI between the two groups (RR 0.98, 95% CI: 0.73, 1.31, p=0.88). There was no statistical evidence of heterogeneity (I-squared 0%, p=0.965). A subgroup analysis in the pre-dialysis patients also showed similar results (RR 0.90, 95% CI: 0.58, 1.41, p=0.66). Only 2 trials with dialysis patients reported MI. A sensitivity analysis that included the 2 excluded trials reporting on MI did not significantly change the results of the analysis.
Poorly controlled BP (4 studies).
There was a significantly higher risk of poorly controlled BP in the higher haemoglobin target group than in the lower target group, based on a fixed-effect meta-analysis (RR 1.27, 95% CI: 1.08, 1.50, p=0.004). There was some suggestion of statistical heterogeneity (I-squared 48%, p=0.119). The random-effects analysis showed no significant difference between the two groups (RR 1.31, 95% CI: 0.97, 1.78, p=0.075). A sensitivity analysis that included the 4 excluded trials reporting on BP did not significantly change the results of the fixed-effect analysis, although the random-effects analysis showed evidence of a beneficial effect of the higher haemoglobin target group (RR 1.62, 95% CI: 1.16, 2,26, p=0.005).
Arteriovenous access thrombosis (6 studies).
There was a significantly higher risk of arteriovenous access thrombosis in the higher haemoglobin target group than in the lower target group (RR 1.34, 95% CI: 1.16, 1.54, p=0.0001). No statistically significant heterogeneity between the trials was detected (I-squared 0%, p=0.612). A sensitivity analysis that included the 2 excluded trials reporting on arteriovenous access thrombosis showed similar results.
Left ventricular mass (5 studies).
None of the studies showed a difference between the higher and lower haemoglobin target groups in the change in left ventricular mass over the study period.