Carolien Zwiers, Mirjam EA Scheffer‐Rath, Enrico Lopriore, Masja de Haas, Helen G Liley

Exchange transfusion and phototherapy have traditionally been used to treat jaundice and avoid the associated neurological complications. Because of the risks and burdens of exchange transfusion, intravenous immunoglobulin (IVIg) has been suggested as an alternative therapy for alloimmune hemolytic disease of the newborn (HDN) to reduce the need for exchange transfusion. To assess the effect and complications of IVIg in newborn infants with alloimmune HDN on the need for and number of exchange transfusions. We performed electronic searches of CENTRAL, PubMed, Embase (Ovid), Web of Science, CINAHL (EBSCOhost), Academic Search Premier, and the trial registers ClinicalTrials.gov and controlled-trials.com in May 2017. We also searched reference lists of included and excluded trials and relevant reviews for further relevant studies. We considered all randomized and quasi-randomized controlled trials of IVIg in the treatment of alloimmune HDN. Trials must have used predefined criteria for the use of IVIg and exchange transfusion therapy to be included. We used the standard methods of Cochrane and its Neonatal Review Group. We assessed studies for inclusion and two review authors independently assessed quality and extracted data. We discussed any differences of opinion to reach consensus. We contacted investigators for additional or missing information. We calculated risk ratio (RR), risk difference (RD) and number needed to treat for an additional beneficial outcome (NNTB) for categorical outcomes. We calculated mean difference (MD) for continuous variables. We used GRADE criteria to assess the risk of bias for major outcomes and to summarize the level of evidence. Nine studies with 658 infants fulfilled the inclusion criteria. Term and preterm infants with Rh or ABO (or both) incompatibility were included. The use of exchange transfusion decreased significantly in the immunoglobulin treated group (typical RR 0.35, 95% CI 0.25 to 0.49; typical RD -0.22, 95% CI -0.27 to -0.16; NNTB 5). The mean number of exchange transfusions per infant was also significantly lower in the immunoglobulin treated group (MD -0.34, 95% CI -0.50 to -0.17). However, sensitivity analysis by risk of bias showed that in the only two studies in which the treatment was masked by use of a placebo and outcome assessment was blinded, the results differed; there was no difference in the need for exchange transfusions (RR 0.98, 95% CI 0.48 to 1.98) or number of exchange transfusions (MD -0.04, 95% CI -0.18 to 0.10). Two studies assessed long-term outcomes and found no cases of kernicterus, deafness or cerebral palsy. Although overall results show a significant reduction in the need for exchange transfusion in infants treated with IVIg, the applicability of the results is limited because of low to very low quality of evidence. Furthermore, the two studies at lowest risk of bias show no benefit of IVIg in reducing the need for and number of exchange transfusions. Based on these results, we have insufficient confidence in the effect estimate for benefit of IVIg to make even a weak recommendation for the use of IVIg for the treatment of alloimmune HDN. Further studies are needed before the use of IVIg for the treatment of alloimmune HDN can be recommended, and should include blinding of the intervention by use of a placebo as well as sufficient sample size to assess the potential for serious adverse effects.