Four RCTs (226 infants) were included.
The Jadad scores for study quality ranged from 1 to 3 out of a possible 5 points. There were several methodological flaws: failing to report intention-to-treat analysis; a lack of reporting when phototherapy treatment was started; and unclear criteria for late transfusion.
When compared with phototherapy alone, HDIVIG plus phototherapy significantly reduced the need for exchange transfusion (3 RCTs, 189 infants); the RR was 0.28 (95% CI: 0.17, 0.47). No significant heterogeneity was detected (P=0.58).
For infants with Rh disease (2 RCTs, 73 infants), HDIVIG treatment significantly reduced the need for exchange transfusion to a greater extent (RR 0.21, 95% CI: 0.10, 0.45). No significant heterogeneity was detected (P=0.77).
HDIVIG treatment plus phototherapy significantly reduced multiple exchange transfusions compared with phototherapy alone (3 RCTs, 198 infants); the RR was 0.22 (95% CI: 0.08, 0.61). No significant heterogeneity was detected (P=0.38).
The addition of HDIVIG treatment to phototherapy significantly reduced the length of hospital stay (2 RCTs, 153 infants); the WMD was -1.06 (95% CI: -1.65, -0.46). Significant heterogeneity was detected (P=0.042). One RCT found that both treatment groups had a longer stay than the other RCT.
The addition of HDIVIG treatment to phototherapy significantly reduced the duration of phototherapy treatment (2 RCTs, 153 infants); the WMD was 0.87 (95% CI: -1.37, -0.37). No significant heterogeneity was detected (P=0.36).
The increase in the number of infants receiving red cell transfusions for late anaemia just reached statistical significance, but the CIs were wide (3 RCTs, 189 infants); the RR was 8.0 (95% CI: 1.03, 62.2). No significant heterogeneity was detected (P=0.7).
The risk difference and NNT were also reported in the review.
In terms of adverse effects, in 3 studies infants were monitored for possible febrile, allergic, haemolytic and volume overload effects of HDIVIG. None of the 3 studies reported any such events. Two studies reported problems in infants in control groups, but no causal link was established; one infant had inspissated bile syndrome; one infant had bacterial sepsis; two infants had hypoglycaemia and hypocalcaemia after exchange transfusions.