Maternal and fetal safety outcomes after in utero stem cell injection: A systematic review

To investigate the maternal and fetal safety of In utero stem cell transplantation (IUSCT).


| INTRODUCTION
In utero stem cell transplantation with haematopoietic (HSC) or mesenchymal stem cells (MSC) may allow prenatal treatment of congenital diseases.The widespread availability of high-resolution fetal ultrasound imaging and advances in prenatal molecular diagnostic techniques means that such disorders are increasingly diagnosed early in gestation.In utero stem cell transplantation may enable treatment of life-threatening disorders before birth, ameliorate in utero damage and potentially provide curative treatment. 1e administration of stem cells to the fetus capitalises on the extensive fetal stem cell migration and expansion that occurs in utero. 2 Given the small size of the mid-trimester fetus, it also permits the administration of a higher cell dose per unit of recipient weight, which may improve engraftment.Compared with postnatal treatment, prenatal treatment offers the physiological advantage of the fetal circulation, which mainly bypasses the pulmonary vasculature, hence avoiding the sequestration of the injected stem cells in the lungs. 3SCT also offers the potential to reconstitute an absent or damaged cell type without the need for myeloablation, or to induce prenatal tolerance to facilitate postnatal transplantation, utilizing fetal immunological naivety.Finally, prenatal therapy may offer a positive psychological benefit to parents, which should not be undervalued. 4e earliest report of IUSCT was in 1967 when a fetus with haemolytic disease due to Rh blood group alloimmunisation received fetal bone marrow HSC. 5 In 1988, the first successful report of IUSCT with HSC emerged, detailing a fetus with Bare Lymphocyte Syndrome who showed full reconstitution of the T cell compartment after umbilical vein injection of fetal liver derived HSC. 6IUSCT with MSC was first successfully reported in 2004, for a fetus with OI. 7,8 Clinical translation of IUSCT must not only present advantages over postnatal treatment but also safe for both the pregnant woman and the fetus.In utero stem cell transplantation is performed via an identical procedure to in utero blood transfusion, with the injection ideally administered into the intrahepatic umbilical vein under ultrasound image-guidance. 9Whilst the immediate post procedural risks of IUSCT are likely to be similar to those of in utero blood transfusion, 9 little is known about complications during the remainder of the pregnancy, particularly maternal adverse events.Two clinical trials of IUSCT are in progress.The first investigates a combined in utero injection of maternal bone marrow-derived HSC and blood transfusion for fetuses with α thalassaemia major (Clin-icalTrials.govIdentifier: NCT02986698).The BOOSTB4 (Boost Brittle Bones Before Birth) trial studies the safety and efficacy of in utero and postnatal transplantation of first trimester human fetal liver-derived MSC for severe OI (ClinicalTrials.govIdentifier: NCT03706482).We undertook a systematic review of fetal and maternal safety of IUSCT providing comprehensive data to inform regulatory authorities, patients and healthcare professionals about IUSCT and to understand the timing of reported adverse events in order to develop optimal monitoring following IUSCT.

| OBJECTIVES
This systematic review investigated the fetal and maternal safety of IUSCT to support regulatory and ethical approval for clinical trials of IUSCT.

| Protocol and registration
This systematic review was conducted in accordance with Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) guidance. 10The protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO-CRD42018110523), with the study title 'Maternal and fetal complications of in utero stem cell therapy'.

| Eligibility criteria
The condition being studied was (1) 'In utero injection of stem cells,' (2) participants were "Pregnant women and their fetuses" and (3) the intervention to be studied was 'In utero injection of haematopoietic or MSC into the fetal circulation'.All publications (randomised, cohort and case-controlled studies, case series, case reports, systematic reviews and narrative review articles) reporting the results of IUSCT in humans were eligible.No comparator groups were considered.No language or date restrictions were applied.where published authors referred to their own unpublished information) were included (Table 1).

| Data extraction
Two reviewers (RS and LWJ) independently extracted data and entered it into a standardised Excel form.Disagreements were resolved by discussion.Characteristics noted included publication type, underlying fetal condition, type of stem cells administered, gestational age of fetus, route of administration, cell dose given and number of injections.To classify adverse events, we used definitions and grading 1−5 for maternal and fetal adverse events as per the Maternal and Fetal Adverse Event Terminology (MFAET) version 1.1 terminology, which is mapped to the Medical Dictionary of Regulatory Activities (MedDRA). 77,78 he primary outcome measures were maternal or fetal mortality or morbidity possibly or probably related to the stem cell injection procedure.Where complications were reported, both reviewers assessed whether the adverse event was possibly or probably related to the IUSCT procedure or if it was unrelated.Unrelated adverse events were defined if they were related to a separate procedure from the IUSCT injection or due to an underlying fetal/maternal condition.To capture this information, outcomes documented included complications at the time of procedure, pregnancy and delivery details, long term fetal outcome and maternal complications.The acute fetal procedure complications and mortality rates are provided in relation to the number of needling procedures performed rather than the number of pregnancies in which IUSCT took place, as per best practise guidance. 9

| Analysis strategy
A narrative synthesis was planned as it was anticipated that the publications identified would mainly consist of individual case reports detailing IUSCTs for a wide variety of conditions.Quality assessment of studies, assessment of heterogeneity and meta-analysis were not considered possible in these circumstances.

| Study selection
The study selection process was carried out in two stages in order to identify all relevant original reports (Figure 1).During the primary search, both original reports and review articles were included.A secondary review of all references generated from the review articles then took place to identify any original reports not found by the primary search.The electronic literature search identified 618 studies published from 1967 to 24/06/2023.Following the study selection, 69 papers detailing 66 IUSCT procedures in 52 women were included in the final data set (Figure 1).

| Study characteristics
The 69 papers were independently read by both reviewers and all data were extracted.When cases were reported in a number of publications, all data were combined from the relevant publications to give the most robust data set.52 cases were found, which contained sufficient information to analyze.Fourteen additional cases of IUSCT were identified in the search but were not included in the statistical data analysis as there was insufficient information to analyze (Supporting Information S1).In these cases, gestational age was given in weeks post fertilization rather than in weeks post last menstrual period.In each of these cases 2 weeks has been added to the gestational age to make the values comparable to the other cases that reported gestational age post last menstrual period.

| DETAILS OF IUSCT CASES
b In these cases, a cell dose/kg was stated, but it was not specified whether this was per injection or in total for both injections.
c All injections in this case series were reported as taking place between 21 and 25 weeks of gestation.
SAGAR ET AL.
injections.Details pertaining to the congenital diseases treated, cell source, cell dose, gestational age at injection and injection type are in Supporting Information S1.

| ANALYSIS OF FETAL AND MATERNAL SAFETY OUTCOMES FOLLOWING IUSCT
We considered maternal and fetal outcomes separately at three timepoints (Table 2).Immediate post-procedural complications were defined as those which occurred during or within 48 h of the IUSCT injection procedure.The second timepoint considered was the remainder of the pregnancy and delivery.Finally, long-term outcomes after delivery were also considered.

| Immediate adverse events within 48 h of IUSCT
Of the 52 cases of IUSCT identified (66 injections), comments were found relating to the health of the fetus or pregnancy within 48 h of the IUSCT procedure in 36 cases; in 33 cases (39 injections) there were no reported concerns, whereas there were complications reported in three cases (4 injections).The presence or absence of immediate procedural complications within 48 h was not documented in 16 cases (23 injections).Only 16 cases (30.8%) contained a direct or implied reference to maternal health in the first 48 h of the procedure.

| Fetal/pregnancy outcomes
Two fetuses suffered a bradycardia and subsequently died in utero within 1 hour of the procedure.The first fetus was receiving a first intravenous injection of second trimester fetal liver-derived HSC for β thalassaemia major at 19 weeks of gestation. 52The second fetus was undergoing a second intravenous injection of second trimester fetal liver-derived HSC for Chronic Granulomatous Disease at 23 weeks of gestation. 52Fetal death during pregnancy, irrespective of cause, constitutes a Grade 5 fetal complication. 78e fetus undergoing transplantation of paternal bone marrow HSC intraperitoneally at 13 weeks for Globoid Cell Leukodystrophy was reported to have 'a small leakage of amniotic fluid' 1 hour following the procedure, with mild oligohydramnios noted on ultrasound.The following day, the liquor volume was reported to be normal, and no further leakage was reported.The fetus was later born alive at term. 26The early gestational age at preterm prelabour rupture of the membranes (PPROM) could render a Grade 4 adverse event but as the oligohydramnios did not persist, the liquor-volume normalised and there was no change in management we therefore determined this was a Grade 2 adverse event. 78l three complications, two of which were Grade 5 fetal adverse events, were determined to be procedure related and were included in the descriptive statistics for acute procedure-related complica- T A B L E 2 Immediate, medium and long term maternal and fetal outcomes after in utero stem cell transplantation. -163

Reference
The 10 healthy fetuses who received IUSCT as part of a study of tissue distribution and concentrations of transplanted fetal liver cells in the human fetus all underwent planned termination of pregnancy 48−72 h following the procedure, as per the study protocol.All fetuses were documented as being alive 6 hour prior to termination. 76ese terminations were therefore not included within the complication rate and were not graded.

| Maternal outcomes
Maternal wellbeing in the first 48 h after the IUSCT procedure was documented in only 16 cases (30.8%).In seven of these, phrases such as 'the procedure was uncomplicated' have been interpreted as reporting both fetal and maternal wellbeing.Individual, specific mention of maternal wellbeing was found only in 9 cases of IUSCT (17.3%).
In the three cases where acute procedural complications were described, these are reported from the fetal perspective.For example, no report is given on the health of the mother in the two reported cases of fetal bradycardia and in utero fetal death. 52Given the fetal gestation at demise of 19 and 23 weeks, respectively, the mother is likely to have been admitted to hospital for management of mid-trimester miscarriage, which would be a Grade 3 adverse event according to terminology. 78Finally, in the pregnancy complicated by PPROM immediately following IUSCT at 13 weeks of gestation, the baby was delivered at term, meaning that severe maternal complications from the membrane rupture are unlikely. 26This was defined as a Grade 2 maternal adverse event. 78 conclusion, no publications report details of maternal health or the specific maternal complications outside of the fetal complications already discussed.However, it is likely that at least one Grade 2 complication and two Grade 3 (or above) maternal complications related to the IUSCT procedure occurred.

| Pregnancy and delivery safety outcomes following IUSCT
From 48 h following the IUSCT procedure, there were 40 ongoing pregnancies.Seventeen cases were reported to have experienced no maternal or fetal complications, whilst a further two cases had no documented fetal complications.In the final four cases, it was documented only that there was no engraftment of cells in the pregnant woman.No information was provided about the remainder of the pregnancy in eight cases.In seven cases, there were further in utero procedures, all of which were fetal blood transfusion for fetal anemia due to α thalassaemia major (n = 4) or Rh blood group alloimmunisation (n = 3). 5,11,31,79,80 Bot conditions are recognised to cause fetal anemia, requiring in utero blood transfusion in order to allow the pregnancy to progress to a viable gestation. 79Indeed, the requirement for ongoing fetal blood transfusions was a pre-specified part of the trial of IUSCT in which two women were partaking.Note: Highlighted in black are the cases of fetal loss, highlighted in dark gray are the cases with additional pregnancy complications, highlighted in light gray are the cases in which specific mention to maternal health was made.Abbreviations: BMT, Bone marrow transplantation; CGD, Chronic Granulomatous Disease; CMV, Cytomegalovirus; CS, Caesarean section; FHR, fetal heart rate; N/A, Not applicable; N/S, Not specified; NVD, Normal vaginal delivery; OI, Osteogenesis Imperfecta; PPROM, preterm prelabour rupture of the membranes; SCID, Severe Combined Immunodeficiency; SE's, Side effects.

1632
- Therefore, we did not consider these subsequent in utero blood transfusions to be a complication of IUSCT.

| Fetal outcomes
Of the 40 continuing pregnancies, three fetuses died before birth.
Two pregnancies, one in which the fetus was affected by α thalassaemia major and one with SCID, were terminated when fetal blood samples showed no evidence of engraftment of maternal bone marrow HSC. 80At postmortem examination, one of these fetuses showed evidence of engraftment. 11We did not include these deaths in the mortality rate attributed to IUSCT.One fetus who received paternal bone marrow HSC intraperitoneally at 13 weeks for globoid cell leukodystrophy died in utero at 20 weeks of gestation due to "excessive infiltration", death which should be considered directly related to IUSCT and included as a Grade 5 fetal complication. 26The cell dose given was 5.0 � 10 9 cells/kg.Two further fetuses with the same condition underwent similar IUSCT treatment but received a ten-fold lower cell dose, with no reports of similar complications during the pregnancies. 26 37 fetuses surviving to birth, gestational age at delivery was reported in 24 cases; 18 fetuses were born at term, whilst 6 were born preterm at gestations between 32 and 36 weeks.All three fetuses who received IUSCT for Rh blood group alloimmunisation were delivered preterm, as was one of the fetuses who received IUSCT for α thalassaemia major. 5,24,51,81 Itis likely that the preterm deliveries in these cases were due to the underlying congenital disorder; it is standard practise to deliver a fetus with a need for ongoing transfusions once a late preterm gestation is reached, usually between 34 and 36 weeks of gestation, in preference to performing further in utero transfusions. 9,82,83 Atus who received one intravenous infusion of fetal liver HSC at 28 weeks for SCID was delivered at 36-37 weeks of gestation. 52Finally, a fetus who received an uncomplicated intravenous IUSCT of fetal liver MSC for OI at 32 weeks of gestation was delivered at 35 weeks after spontaneous PPROM occurred. 18,19 Te study authors concluded that the preterm delivery 3 week after the procedure was due to the fetus' underlying disease as this is associated with a high rate of preterm birth. 84Thus, these occurrences of preterm delivery were not included in the complication rate.
There was one neonatal death. 5This was a complicated case where bone marrow HSC had been given to the fetus via hysterotomy at 11 weeks of gestation for Rh blood group alloimmunisation.
Five further needling procedures occurred during the pregnancy: twice to perform amniocentesis and three times for in utero blood transfusion.The pregnancy was complicated by PPROM following a fetal transfusion at 26 weeks, vaginal bleeding, and spontaneous preterm labor at 32 weeks.During the last blood transfusion at 31 weeks of gestation, the needle was seen to enter the fetal bowel.
The baby died on day one of life after developing respiratory distress and signs of infection.After delivery, there was evidence that the fetal anemia may have been caused by congenital rubella infection and that the fetus may have been the Rh blood group negative.There was additional post-mortem evidence of fetal bowel perforation and meconium peritonitis, which corresponded to the complications of the fetal blood transfusion at 31 weeks.Whilst this constituted a grade 4 intraoperative injury adverse event as the unintended damage to the fetal organ was life-threatening, these complications are undoubtedly a result of the last intrauterine blood transfusion needling procedure at 31 weeks.We therefore considered it unlikely that the complications were related to IUSCT, which took place at and intrauterine death due to 'overwhelming engraftment' (n = 1)), the maternal outcome was documented in only three cases. 5,31 n two of these cases, the mother was well during a follow-up, up to a year after delivery.In the other case, however, the mother is described as having evidence of salpingitis at delivery.This maternal complication is likely unrelated to the IUSCT procedure performed over 20 weeks prior, but if thought related to the IUSCT would constitute a Grade 2 maternal adverse event. 5 considered the requirement for in utero blood transfusions in the additional four cases to be related to the underlying disorder and thus not related to the IUSCT.However, in the case of intrauterine demise secondary to overwhelming engraftment, a complication caused by IUSCT, the resulting mid-trimester management of miscarriage is likely to have required hospital admission, which would constitute a Grade 3 maternal adverse event.There was one IUSCT-related fetal complication reported in the intermediate term, which the authors attribute to a direct fetal reaction to the high cell dose given, as subsequent use of a ten-fold lower dose of cells in two fetuses was not related to similar complications. 26For long term fetal outcomes, the majority of reports focused on efficacy rather than on reporting safety parameters.
There was convincing evidence of benefit in 10 cases of immunodeficiency (HSC) and two cases of OI (MSC), with partial benefit in two cases of thalassaemia.There was no demonstrable benefit in storage diseases or clotting disorders.These cases have been discussed in detail elsewhere and further comment is outside the scope of this review.Whilst non-tolerance to donor cells was reported in some cases, there were no reports of immune response, GvHD, or tumorigenicity in any of the fetal recipients, and no reported direct longterm complications.
Maternal outcomes were less well reported than fetal.In only nine cases was it specifically documented that there were no acute procedural complications in the mother.Likewise, only 13 cases specifically mention maternal wellbeing during pregnancy and delivery.An additional 10 publications made general comments about there being no complications during pregnancy, which has been interpreted as stating that there were no maternal or fetal complications (44.2% overall).There are only two reports of maternal health after delivery of the baby; in both cases the mothers were well a year following delivery.This lack of explicit maternal safety reporting is not unique to IUSCT and has, for example, been commented upon in a recent systematic review of fetal surgery for spina bifida. 86ilst there were no specific maternal-only complications related to IUSCT reported in the 52 pregnant women studied in this systematic review, in three cases which reported acute or intermediate fetal/pregnancy related complications, it is likely that the complication would have additionally constituted at least a grade 3 maternal adverse event according to MFAET taxonomy.
Despite the poor reporting of safety outcomes overall, certain conclusions can reasonably be drawn from our results.Three of the four fetal complications related directly to IUSCT occurred within 24 h of the IUSCT procedure.This time course is consistent with the complications which were reported to occur after in utero blood transfusion by Zwiers et al, where seven of the 11 documented procedure related complications (in 334 fetuses) occurred within 24 h of the IUT procedure. 9We therefore recommend that monitoring for adverse events should focus most on the 24−48 h immediately following IUSCT.However, one intrauterine death directly related to IUSCT occurred 7 weeks following the procedure; hence, we recommend that monitoring takes place throughout the pregnancy and delivery.Very little data is available regarding long-term safety of IUSCT, and for this reason, we recommend long-term follow-up in order to establish reliable safety data.Certainly,

1634
- competent authorities require that trials investigating the administration of advanced therapy medicinal products such as stem cells follow-up participants for minimum 5 years. 87Tumorigenicity and immune response are pertinent negatives to report in the long-term follow-up.
Reporting the presence or absence of not only fetal but also neonatal and maternal complications is recommended in future cases of IUSCT.Indeed, adverse event reporting for all studies involving pregnant women should use maternal and fetal definitions and grades according to MFAET, which has been mapped to MedDRA and for neonates should use the National Cancer Institute INC Neonatal Adverse Events Terminology (NAESS). 88,89 en the limited evidence available currently to support the safety and efficacy of IUSCT, we recommend that it should take place only in the setting of clinical trials, where safety is recorded as the primary outcome, and with fetal, maternal and neonatal outcomes documented in the immediate, medium and long term according to accepted clinical trial definitions.Subsequently, should both safety and efficacy be demonstrated for IUSCT for a given condition, IUSCT could then be performed as a standard of care according to best practise protocols developed during these trials.

3 . 4 |
Search strategy A systematic search was conducted in Medline®, Embase and the Cochrane library databases using free text and Medical Subject Headings.Reference lists of relevant review articles were manually checked.Covidence (Veritas Health Innovation Ltd, Melbourne, Australia) was used to eliminate duplicate articles and manage study screening.The initial searches were performed electronically on 25/9/2018, and repeated on 24/6/2023.Search terms were: � Title (prenatal OR fetus OR fetal OR "in utero" OR intrauterine) � AND ("Stem cell") � AND (Transplant*) Study selection Two reviewers (RS and LWJ) reviewed titles and abstracts independently and excluded irrelevant studies.The same two reviewers independently performed full-text screening; disagreements were resolved by discussion.Studies were excluded at the full text screening stage if the full text was unavailable and the abstract contained insufficient information or if duplication had occurred.Studies involving only animals, postnatal stem cell administration, or stem-cell administration into the uterus outside of pregnancy were excluded.Studies were divided into original case publications and review articles.The same two reviewers analyzed the references of all review articles and identified relevant original publications referenced in these reviews.For inclusion in the final data set, publications had to contain details of the IUSCT along with at least one outcome measure.Only publications detailing original cases (or Gestational age is presented in full weeks.Abbreviations: CGD, Chronic Granulomatous Disease; HSC, haematopoietic stem cells; IP, intraperitoneal; IV, intravenous; SCID, Severe Combined Immunodeficiency.a unanticipated safety events in either mother/child pair There were no unanticipated safety events in either mother/child pair, 4 in utero blood transfusions unanticipated safety events in either mother/child pair There were no unanticipated safety events in either mother/child pair, 4 in utero blood transfusions

B L E 1 Summary table of the final data set after the selection procedure. Reference Also referenced in Disease Cell type Cell source Donor Cell dose Injection route Gestational age No of injections
T A

Reference Also referenced in Disease Cell type Cell source Donor Cell dose Injection route Gestational age No of injections
Flow diagram of study selection (adapted from PRISMA 2009).The electronic literature search identified 618 studies published from 1967 to 24/06/2023.249 studies were removed as duplicates.The remaining studies (369) were screened by title and abstract, and a further 268 were excluded as irrelevant.Full texts of the remaining 101 articles were reviewed, and 23 were excluded for the following reasons: 14 wrong population (12 animal studies, 2 postnatal population) and 9 wrong study type (6 opinion pieces without references, 1 conference abstract of this systematic review study and 1 study on fetal immunology with no In utero stem cell transplantation (IUSCT)).78 studies were identified based upon this initial search strategy.38 publications were case reports/series of original data and were directly included.Forty review articles generated from the primary search were then reviewed independently by both reviewers and a secondary review of references performed.33 additional studies detailing original cases were identified from this secondary review of references.Two publications detailing cases of IUSCT were excluded due to insufficient information.Eventually, 69 papers detailing 66 IUSCT procedures in 52 women were included in the final data set.[Colour figure can be viewed at wileyonlinelibrary.com] tions.Of the 66 IUSCT procedures, the acute per-procedure fetal complication rate was therefore 4.5% (3/66) and the acute procedural fetal mortality rate was 3% (2/66).F I G U R E 1 1628 -SAGAR ET AL.
that IUSCT procedures, performed today and using the knowledge from these best practise points from fetal blood transfusion studies, will have a lower procedure related complication rate than was seen in cases carried out 3 decades ago.
Of the 36 surviving fetal recipients of IUSCT, there was evidence of benefit in 12 cases, and possible benefit in two additional cases.These long term outcomes have been extensively discussed elsewhere, and are not analyzed in this systematic review of fetal and 1.6%.A retrospective safety analysis was performed and after implementation of practise improvement points, such as routinely using fetal paralysis, and avoiding injection into free cord loops, 937 procedures were subsequently performed.In this second cohort of patients treated from 2001 to 2015, the authors report a procedure complication rate of 1.2% and a procedure loss rate of 0.6%.It is therefore likely