IUTs have improved foetal survival but the procedure is not without its complications. Performing the IUT procedure requires a high level of skill and so reducing the number of transfusions is desirable to prevent the risk of procedure-related (PR) complications including the risk of infections transmitted via blood components.17,18 PR complications include membrane rupture, preterm delivery, chorioamnionitis, and foetal and neonatal death.8,19 The most serious complication during the procedure is foetal distress which can result in emergency preterm delivery or foetal death. Major complications from IUTs are associated with earlier stages of gestation (< 20 weeks) as foetal vessels are smaller, hence access to foetal circulation is difficult, and foetuses cannot tolerate the transfusion and die in utero due to acute haemodynamic changes.3,6 The risk of foetal death due to PR complications in gestations of less than 20 weeks is 8.5 % and of greater than 20 weeks it is 0.9 % per procedure.20 Foetuses with severe anaemia and hydrops have higher risks of foetal and neonatal death.19,21

Postponing IUT would be beneficial to reduce PR complications. As pregnancy progresses the foetal umbilical vein increases in diameter, making the procedure easier and safer.9 To delay the procedure, a high-dose of intravenous immunoglobulin (IVIG) can be administered. This was explored in a study where the onset of foetal anaemia was delayed by 8–17 weeks after the administration of high-dose IVIG, thus postponing the need for IUT.22 The mechanism of action of IVIG is unknown, but there are several proposed theories.23 IVIG theoretically may block the Fc receptors in the neonatal reticuloendothelial system and induce competition at the placenta, reducing Fc-mediated antibody transport. This prevents foetal haemolysis but does not treat existing foetal anaemia. Foetal mortality improved as IVIG appeared to reduce the mortality rate from 51 % to 20 %.17,20 Hence, theoretically, IVIG is beneficial but further studies on a larger scale are required to provide sufficient evidence.

Foetal death is a devastating event that can occur when a mother is Rh(D) alloimmunised causing foetal anaemia. Ultrasound-guided IUT is a successful technique used to treat foetal anaemia thereby reducing the mortality rate. However, complications can arise and lead to preterm birth, membrane rupture and death of the foetus or neonate. These PR complications may be associated with gestational age. Hence, early diagnosis and identifying the optimal time to perform the procedure would be beneficial in order to start treatment before the development of hydrops thus enhancing the chances of survival.

Postponing the need and decreasing the frequency of IUTs could reduce the chance of PR complications; finding the best possible way to accomplish this remains unknown. There have been multiple studies that examined the use of high-dose IVIG as an alternative therapy however, its mechanism of action is not fully understood and this treatment does not eliminate foetal anaemia, it only helps to reduce foetal haemolysis. IVIG may be successful in delaying the onset of foetal anaemia and thus delay the need for IUT and reduce mortality.

The aim of this study was to determine whether IUT is a safe and effective treatment for foetal anaemia in Rh(D) alloimmunised pregnant women?

This study evaluated the mortality rate of foetuses affected by foetal anaemia due to severe Rh(D) alloimmunisation. The primary aim of this study was to investigate the survival and foetal mortality rates, to determine the effectiveness of IUTs and to examine the risk of PR complications associated to foetal loss so as to determine IUT safety and identify the influence of hydrops.

The Preferred Reporting Item for Systematic Review and Meta-analysis (PRISMA) procedure was implemented in this systematic review to assist in gathering relevant articles to investigate the safety and effective management of IUTs in Rh(D) alloimmunised pregnant women.24

Searching for appropriate literature was conducted using PubMed, Scopus, Google Scholar and Semantic Scholar databases, with the timeline defined as from January 2000 to April 2022. Key search terms included “intrauterine transfusion” AND “alloimmunisation”, “intrauterine transfusion treatment”, “alloimmunisation pregnancy treatment”, “severe Rh(D) alloimmunisation”, “Rh(D) alloimmunised AND foetal anaemia” and “foetal anaemia AND treatment”. A manual search was performed within reference lists of selected articles with one additional study being identified. See Appendix 1 for full search strategy.

To determine the eligibility of the studies obtained in the search, articles were screened based on title and abstract with the application of inclusion/exclusion criteria. Articles that examined the management of IUT and evaluated the mortality rate of affected foetuses were considered eligible. Articles that reported PR complications with foetal outcomes and a comparison between the presence and absence of hydrops were also included. Exclusion criteria comprised of: 1) review articles and book chapters, 2) non-English publications, 3) inexistence of comparison group when investigating hydrops, 4) non-inclusion of foetal anaemia, 5) non-evaluation of mortality rates, 6) case reports, 7) abstract only articles.

The selected studies were assessed according to the Strengthening the Reporting of Observational studies in Epidemiology (STROBE) checklist for their methodological quality.25

Data was extracted from each eligible study including the primary author, year of publication, study design, country, study period, sample size (both the total number of foetuses and the total number of transfusions performed) and the parameters measured for analysis. These parameters included survival rates, mortality rates, and PR complications and their associated foetal loss. Only data for survival and foetal mortality were used in the systematic review for articles that contained a comparison between the presence and absence of hydrops after the first IUT.

To conduct the meta-analysis, OpenMeta-Analyst software was downloaded via the Brown University website.26 A single-arm proportion analysis was used for overall survival, foetal mortality and PR complications with associated foetal loss parameters and a two-arm proportion analysis was used for the presence and absence of hydrops comparing the survival and foetal mortality rates. Each used the arcsine transformed proportion with the binary random-effects model and maximum likelihood method. The results for each parameter are presented as forest plots. The software calculated the overall p-value to ensure statistical significance with 95 % confidence intervals and I2 to assess heterogeneity and its corresponding p-value. An overall p-value < 0.05 was considered statistically significant.

The study selection based on inclusion/exclusion criteria is shown in Figure 1. The database search obtained 2789 potentially relevant studies. Duplicates were identified and removed leaving 2227 studies. These were screened based on title and abstract with 1547 studies being excluded. The remaining 680 studies were assessed for eligibility. Irrelevance to the research question, the major exclusion criteria, excluded another 601 studies. Review articles & book chapters, case reports, lack of a comparison group for hydrops, non-English publications and abstract only papers were also excluded. Thus, 14 studies matched the eligibility criteria, and one extra study was added after a manual search. All 15 studies were included in the meta-analysis.

Figure 1.

PRISMA flowchart. Study selection process from searching four databases, removing duplicates, screening based on title and abstract, eligibility testing and a manual search leading to the included studies used in the meta-analysis.

(0.47MB).

The characteristics of the 15 studies that evaluated the survival and mortality rates of anaemic foetuses are shown in Table 1. Of the included studies, 11 highlighted the risk of PR complications and foetal loss rates and six studies emphasised the influence of hydrops contributing to mortality. Most of these studies had a retrospective study design across a 1988 to 2019 time period spanning 11 different countries. The sample size included the total number of foetuses which ranged between 42 and 334, and the total number of transfusions which ranged between 87 and 937. The gestational age at first IUT for most studies was 25–28 weeks. However, one study conducted by Yinon et al. performed the first IUT at a mean gestational age of 20 weeks,3 and another study conducted by Tiblad et al. performed the first IUT at a mean gestational age of 31 weeks.31

Study data used for the meta-analysis are shown in Table 2. The overall survival and foetal mortality rates were calculated per foetus and the PR complications and associated foetal loss were calculated per procedure. Data referring to mortality only relate to foetal mortality as neonatal morality was excluded. When investigating foetal mortality related to hydrops, data on survival and mortality of foetuses with or without hydrops were used to determine if it impacted survival. Data from one study only included results from the cohort from 2001 onwards.32

Studies included in the systematic review were analysed using the STROBE checklist to assess the methodological quality as shown in Table 3. All studies had a high quality providing sufficient detail for the title/abstract and introduction sections. Most studies had a high quality in the methods section, however two studies failed to outline eligibility criteria for participant selection and three studies only mentioned the inclusion criteria. A summary of participant characteristics was included in all studies however three had limited results. The majority of studies discussed limitations usually based on the retrospective study design or the sample size, however three studies failed to mention any limitations. Also, it should be noted that three studies failed to describe any efforts to address potential sources of bias within the methodology.3,19,30Table 3 shows that the studies were mostly of high methodological quality which reduces the risk of inappropriate study selection for the analysis.

A meta-analysis and forest plots, constructed for overall survival, mortality rate, the risk of PR complications and the risk of foetal loss due to PR complications, are displayed in Figure 2, Parts A, B, C and D, respectively from the data presented in Table 2.

Figure 2.

Forest plots of meta-analysis on survival and mortality rates, risk of complications and risk of foetal mortality. (A) Overall survival, (B) foetal mortality, (C) risk of procedure-related complications, and (D) risk of foetal loss due to PR complications were expressed after single-arm proportion data analysis over the total number of events with the arcsine transformed proportion, the binary random-effects model and maximum likelihood for each parameter evaluated with a 95 % confidence interval. Statistical significance was assessed by an overall p-value and heterogeneity was evaluated by the I2 value and associated p-value. Relative sample sizes were determined by weighting for each study.

(0.37MB).

The one-arm proportion analysis on overall survival was found to be statistically significant with an overall p-value < 0.001. The total survival was 89.79 % of foetuses. The results of the 15 studies were found to be highly heterogeneous (p-value < 0.001; I2 = 88.64 %) therefore, there is variability in the data. The results indicate that there is greater foetal survival when IUT is used to treat anaemic foetuses (95 % confidence interval [CI]: 0.841–0.933).

Similarly, the one-arm proportion analysis of foetal mortality was found to be statistically significant with an overall p-value < 0.001. The mortality rate was 7.61 % of foetuses. The results showed an I2 value of 78 % and a het p-value < 0.001 indicating high heterogeneity. The forest plot suggests a low foetal mortality rate with IUT treatment (95 % CI: 0.051–0.110).

Data on PR complications were used to determine the safety of IUT with a total PR complication rate of 3.95 %. Single-arm proportion analysis was conducted giving an overall p-value < 0.001 indicated the results were statistically significant, and an I2 value of 85 % and a het p-value of < 0.001 suggesting high heterogeneity. The results indicate a low risk, albeit not fully eliminated, of complications (95 % CI: 0.026–0.062).

Data on foetal loss due to PR complications (1.29 %) were also used to investigate the safety of IUTs. Single-arm proportion analysis was used to identify the risk of foetal mortality. An overall p-value < 0.001 indicates the results were statistically significant (95 % CI: 0.009–0.018). The results of the included studies were found to have low to moderate heterogeneity (p-value = 0.04; I2 = 28.8 %). The results suggest a low risk of foetal mortality due to PR complications.

The influence of hydrops on foetal survival and mortality was examined by meta-analysis and forest plot as shown in Figure 3 (Part A: presence of hydrops; Part B: absence of hydrops). Two-arm proportion analysis was used for each parameter to compare the influence of hydrops using data on the survival rate and number of foetal mortalities compared to the total number of foetuses treated with IUT. Of the foetuses with hydrops, 71.86 % survived and 28.14 % died. Of those without hydrops, 89.54 % of the foetuses survived and 10.46 % died. The overall p-value of 0.009 calculated for foetuses with hydrops proved statistically significance with high heterogeneity (I2 = 87.18 %; het p-value < 0.001; 95 % CI: 0.113–0.773). The forest plot suggests a shift towards foetal mortality when hydrops is present. The results for foetuses without hydrops showed an overall p-value < 0.001 hence statistically significant (95 % CI: 0.070–0.135). However, the heterogeneity I2 value was 63.73 % with a het p-value of 0.010, which indicates moderate heterogeneity.

Figure 3.

Forest plots of meta-analysis comparing survival & foetal mortality rates related to hydrops. (A) Hydrops present and (B) hydrops absent were expressed using two-arm proportion data analysis compared to the total number of events using the arcsine transformed proportion, the binary random-effects model and maximum likelihood for each parameter assessed with a 95 % confidence interval. Statistical significance was considered for an overall p-value with heterogeneity being evaluated by the I2 value and associated p-value. Relative sample sizes were determined by weighting for each study.

(0.35MB).

This systematic review and meta-analysis revealed with good power that IUT is an effective treatment for foetal anaemia. The overall survival rate was statistically significant and showed that 1486/1656 foetuses survived after IUT (95 % CI: 0.841–0.933: p-value < 0.001). The high heterogeneity is represented by I2 = 88.64 % and p-value < 0.001 categorising the variability of data across studies as heterogeneous rather than by chance. The results obtained can be explained by the number of studies included in the meta-analysis. The results provide confidence for reproducibility as highlighted by the narrow blue diamond on the forest plot. One of the studies showed a 50 % chance of survival which is a considerably different result to the other studies.5 This survival rate was the lowest recorded, but this was due to the large number of severely hydropic cases in the cohort. It was found that this study included severely anaemic foetuses at first IUT and the majority of participants had severe hydrops which probably impacted on the result. In addition, most of the mortality that occurred was neonatal and not foetal death. Overall, the results showed a greater rate of the anaemic foetuses survived after IUT treatment however, this is what the results suggest rather than the conclusion due to the high heterogeneity of the study.

Foetal mortality rates also supported IUT treatment. Results suggested a low risk of foetal mortality after IUT treatment. The overall foetal mortality rate of 126/1656 was found to be statistically significant (95 % CI: 0.051–0.110; p-value < 0.001) and had confidence for reproducibility. The results showed an I2 value of 78 % and a het p-value < 0.001 indicating high heterogeneity, hence variability present in the data with the results obtained being used as a suggestion rather than a firm conclusion. Most studies identified a similar low range of foetal death, but four studies showed higher weighting on mortality.3,5,27,29 These studies had more cases of hydrops and more procedures performed which could have led to an increase in PR complications hence, higher mortality. One of the studies had a mean gestational age of 20.4 weeks at first IUT with treatment being shown to be less effective.3

IUTs have proven to be a safe and beneficial procedure as results indicate a low risk of PR complications. A total PR complication rate of 144/3646 procedures was obtained by 11 studies and was found to be statistically significant with high heterogeneity (95 % CI: 0.026–0.062; p-value < 0.001; I2 = 85 %; het p-value < 0.001). Results suggest there is a low risk for PR complications, however there is an increased risk per procedure performed and some foetuses may require multiple transfusions. Hence, postponing the need for IUTs at early gestational ages would be beneficial to limit the frequency of transfusions. IVIG could be used as a non-invasive alternative to avoid early transfusions and delay the onset of foetal anaemia.20 Two studies showed a higher complication risk.28,29 These studies had a larger proportion of participants with hydrops which could have caused a shift in the results of the meta-analysis.

Data on foetal loss due to PR complications ensured that IUTs are safe procedures. The results indicate a low risk of foetal mortality due to PR complications. An overall PR foetal loss rate of 47/3646 procedures was obtained and was found to be statistically significant with low to moderate heterogeneity (95 % CI: 0.009–0.018; p-value < 0.001; I2 = 28.83 %; het p-value = 0.04). This suggests that the studies reported comparable results, with confidence of reproducibility; there is a low risk of foetal mortality due to PR complications. One of the studies showed a much higher risk of foetal loss.29 It was found that 33 % of foetuses had hydrops, some of the foetuses were treated at an early gestational age of 22 weeks and all foetuses required more than one IUT. There can be difficulty to access small vessels for transfusions at early gestational ages, hence a higher foetal mortality rate. Therefore, treatment is not as effective in early gestation. Delaying IUT to a later gestational age with IVIG would be beneficial as the vessel diameter would have increased, making the procedure easier and safer; this could also reduce the number of IUTs required per foetus.20 The mechanism of action of IVIG is unknown, but theoretically it stimulates competition at the placenta by reduces Fc-mediated antibody transport, preventing foetal haemolysis. However, this treatment cannot treat existing foetal anaemia on its own and so IUTs are still required.23

The presence of hydrops can influence foetal mortality. Hydrops occurs if the baby's developing organs struggle to overcome anaemia, leading to heart failure and death. The condition has a higher risk of intrauterine foetal death and is harder to treat.33 Data from six studies were used to compare the survival and mortality data in respect to the total number of foetuses being treated by IUT; hydrops was present in 120/167 foetuses that survived and in 47/167 foetal mortalities. The results were statistically significant with high heterogeneity (95 % CI: 0.113–0.773; p-value = 0.009; I2 = 87.18 %; het p-value < 0.001) indicating variation in the data. The results suggest a shift towards foetal mortality when hydrops is present. Hydrops was absent in 291/325 foetuses that survived and 34/325 foetal mortalities. Even though there is a higher survival rate for foetuses without hydrops, it does not exclude the risk of mortality. Multiple IUTs had been performed for the 34 foetuses that died in utero, hence an increased risk of PR complications exists. The results were statistically significant with moderate heterogeneity (95 % CI: 0.070–0.135; p-value < 0.001; I2 = 63.73 %; het p-value = 0.010), indicating some variation between studies. The results suggest that there is a better chance of survival when hydrops is absent. Therefore, early detection of anaemia and the prevention of the development of hydrops are important to increase the chance of survival.

The major limitation of this review was the lack of prospective studies. The included studies were mainly retrospective which can introduce bias. Some of the studies failed to address potential sources of bias and discuss any limitations that might have existed. Some studies only had one specialist conducting the study and gathering results from patients, however the majority of selected studies eliminated this element of bias by having multiple specialists oversee the study. Some studies also relied on previous medical records, hence assumed that the documentation had been filled out correctly. Finally, the hydrops data in the meta-analysis only included six studies, therefore there were less variations in the data.