Risks of Miscarriage or Preterm Delivery in Dichorionic Triamniotic Triplets with Multifetal Embryo Reduction to Singleton Pregnancy Versus Expectant Management: A Systematic Review

Abstract

Background/Objectives: Dichorionic triamniotic (DCTA) triplet pregnancies are associated with increased rates of placenta-specific complications primarily attributed to vascular anastomoses in the monochorionic (MC) pair. Selective fetal reduction to twins (of one of the MC pair) is a complex and not a widely available procedure. Multifetal reduction (MFR) to singleton pregnancy can reduce adverse pregnancy outcomes but is controversial due to medico-legal and socio-ethical issues. The aim of this study is to identify the rate of miscarriage < 24 weeks or preterm birth < 34 weeks following MFR to singleton pregnancy in DCTA triplets and compare the results with expectant management. Methods: This systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and registered in the Prospective Register of Systematic Reviews System (ID: CRD42023422585). Results: Overall, from 21 citations of relevance, 6 studies with a total of 548 DCTA triplet pregnancies fulfilled the inclusion/exclusion criteria. In comparison with expectant management (n = 336), meta-analysis demonstrated that MFR to singleton pregnancy (n = 212) was associated with a lower rate (9.4% vs. 48.5%) of preterm birth (RR = 0.19, 95%CI 0.07–0.51), whereas the rate of miscarriage (14.6% vs. 9.2%) did not significantly increase (RR = 1.53, 95%CI 0.91–2.55). Conclusions: In DCTA triplet pregnancies, MFR to singleton pregnancy was associated with a reduced preterm birth rate and not associated with an increased miscarriage rate. Given the fact that the MC pair is reduced only to lower the rate of preterm birth, appropriate counselling and justification are important. In the absence of randomized controlled trials, data from systematic reviews are the best available evidence for counseling on the different management options.

1. Introduction

The widespread use of assisted reproductive techniques increased the rate of triplet pregnancies [1,2,3,4,5,6,7,8]. Triplet pregnancies of mixed chorionicity (dichorionic triamniotic, DCTA) are increasingly common, mainly because of the splitting of one of the embryos transferred [8,9]. DCTA triplets are associated with a higher rate of miscarriage (<24 weeks of gestation) and preterm birth (<34 weeks) [7,10,11,12,13,14]. To prevent DCTA triplets, single embryo transfer and selective fetal reduction (SFR), of one of the monochorionic (MC) pair, have been used.

In DCTA triplets, there is ample evidence to support SFR to dichorionic diamniotic (DCDA) twins, thus considered to have a more favorable outcome [7,13,15,16,17,18]. However, SFR to DCDA twins is a complex and not widely available procedure. Multifetal reduction (MFR) to singleton pregnancy can reduce adverse pregnancy outcomes but is controversial due to medico-legal and socio-ethical issues.

Moreover, in DCTA triplets, detailed counselling on the different options available and appropriate risk stratification into a high-risk group for perinatal complications should take place. Therefore, identifying potential risk factors for optimal stratification is important. Both biochemical and ultrasonographical markers used in first-trimester combined screening could serve this purpose [19]. While lower concentrations of pregnancy-associated plasma protein-A and free beta-human chorionic gonadotropin demonstrated associations with outcomes of preterm birth, their individual and collective predictive efficacies for foreseeing such events were found to be suboptimal [20]. In view of scarce literature and ethical dilemmas involved, evidence-based counselling could optimize decision-making and prevent placenta-specific complications [7,21].

There is a need for improved care and support for parents facing the decision of continuing a DCTA triplet pregnancy or deciding on MFR to singleton pregnancy. By summarizing all available English-language literature, this study aims to identify the rate of miscarriage < 24 weeks or preterm birth < 34 weeks after MFR to singleton pregnancy in DCTA triplets and compare the results with expectant management.

2. Materials and Methods

This systematic review was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA) and followed an a-priori-designed search protocol, which was registered prospectively on the Prospective Register of Systematic Reviews System (PROSPERO) (ID: CRD42023422585) [22,23]. Ethical approval was not required for retrospective studies without patient identifiers.

2.1. Study Inclusion Criteria

2.1.1. Types of Studies

To gather any available information, the study design included prospective randomized controlled trials (RCTs), prospective cohort studies, and retrospective studies [24,25].

2.1.2. Types of Participants

DCTA triplets with 3 live fetuses at 8–14 weeks, data on both expectant management and MFR to singleton pregnancy, outcomes of miscarriage < 24 weeks or preterm delivery < 34 weeks.

2.1.3. Types of Therapeutic Interventions

• Expectant management;
• MFR to singleton pregnancy by ultrasound-guided intrathoracic potassium chloride injection or ultrasound-guided laser ablation/radiofrequency ablation of the pelvic vessels of the MC pair.

2.1.4. Types of Outcome Measures (With Their Respective Measurement Units)

• Miscarriage < 24 weeks (rate);
• Preterm birth < 34 weeks (rate). To compare our data current literature, the cut-off for preterm birth is represented by a range (from 24⁺⁰ to <33⁺⁶ weeks). Summary outcomes are provided as relative risks (RR) and 95% confidence intervals (95% CI).

2.1.5. General Admission Requirements

Conclusively, studies were included when the following criteria were met:

• DCTA triplets with 3 live fetuses at 8–14 weeks;
• Available data on both expectant management and MFR to singleton pregnancy;
• Outcomes of miscarriage before 24 weeks or preterm delivery before 34 weeks;
• Presented data allowed calculation of miscarriage or preterm delivery rate;
• Data or subsets of data were not published more than once.

2.2. Study Exclusion Criteria

Unmatched case–control studies with different or inconsistent gestational age at recruitment between the 2 arms were excluded, because the rate of miscarriage is higher early in pregnancy. Studies with gestational age at reduction ≥ 14 weeks were similarly excluded. We excluded the following studies: review articles, case reports, case series (reporting on ≤5 participants), letters to the editor, comments, conference abstracts, retracted articles and study protocols.

2.3. Identification of Studies

A systematic literature review was conducted to identify studies reporting on MFR to singleton pregnancy and expectant management in DCTA triplets. Primary online medical databases including MEDLINE via PubMed, EMBASE, Scopus, Cochrane Database of Systematic Reviews (CDSR), Cochrane Central Register of Controlled Trials (CENTRAL), and ClinicalTrials.gov registry accessed on 31 December 2025 were searched from 1 January 2000 until 31 December 2025, due to potentially markedly altered practice and techniques over an extensively elongated period. Reference lists from identified articles and relevant systematic reviews were screened manually to locate further eligible studies. The search algorithm included: combinations of medical subject heading (MeSH) terms, keywords, and word variants for: ‘triplet pregnancy’, ‘mixed chorionicity’, ‘fetal reduction’, ‘selective termination’, ‘selective reduction’, ‘dichorionic triplets’, ‘monochorionic multiple pregnancy’, ‘multifetal reduction’, and ‘monochorionic component’. After reading the abstracts, two reviewers (C.A. and E.I.) independently assessed all potentially eligible articles in which expectant management and MFR to singleton pregnancy were evaluated. The filters that were applied were English language and human female population. Grey literature, including conference proceedings, dissertations, and unpublished studies, was excluded from this review.

2.4. Study Selection

Duplicates were removed and two independent reviewers (C.A. and E.I.) initially screened the title and abstract of the retrieved studies. If the inclusion/exclusion criteria were satisfied, journal articles were obtained and full-text evaluation was independently conducted by both reviewers to determine final eligibility. Discrepancies during the selection process between the reviewers were resolved through discussion and consensus involving a third investigator (T.T.).

2.5. Data Extraction

For each eligible study, two independent reviewers (C.A. and E.I.) extracted appropriate data using a predefined form designed specifically for this systematic review. Standardized preformulated tables were used to ensure completeness and allow subsequent data analysis and outcome synthesis. The reviewers were not blind to journal names, article authors or affiliations as this was deemed unnecessary. If required, corresponding authors were contacted via institutional email.

The following data were recorded for each article:

• Author(s);
• Publication year;
• Number of participants in each arm;
• Conception method;
• Maternal age;
• Gestational age at recruitment;
• Chorionicity and amnionicity based on the presence or absence of the lambda sign [26];
• MFR technique;
• Miscarriage rate;
• Preterm delivery rate;
• Gestational age at delivery;
• Number of live births and survivors.

If overlapping samples were detected, the study with the largest sample size was selected for data extraction.

2.6. Study Design Characteristics

The following study design characteristics were extracted from each included study: study type (descriptive or analytical), investigator role (experimental or observational), allocation and control (randomization/blinding), direction of inquiry (prospective or retrospective), timeframe (longitudinal or cross-sectional), and comparison method.

2.7. Quality Assessment

Two independent reviewers (C.A. and E.I.) assessed the quality of each included study using GoodReseArch for Comparative Effectiveness checklist (GRACE) [27,28], and the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (Ref. https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort (accessed on 31 December 2025)). Included studies should meet at least 50% of the points for each scale. Discrepancies were discussed and, if consensus could not be reached between the two reviewers, a third senior reviewer (T.T.) was consulted.

2.8. Statistical Analysis

For data analysis, x² and Fisher’s exact test were used appropriately. Dor data synthesis, 2 × 2 contingency tables were constructed and RRs, 95% CIs, and number needed to treat (NNT) or harm (NNH) were calculated (when RRs did not cross 1.0). If 2 × 2 contingency table contained a zero cell, 0.5 was added.

The following statistical software packages were used for data analysis and outcome synthesis:

• Comprehensive Meta-Analysis (CMA), Version 3.3.070, 2014 (Biostat Inc., 14 North Dean Street, Englewood, NJ 07631, USA);
• Medcalc for Windows, version 12.7 (Medcalc Software, Mariakerke, Belgium);
• Review Manager (RevMan), Version 5.3, 2014 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration).

3. Results

3.1. Search Results and Study Selection

The electronic database search yielded 162 results. Following duplicate removal and exclusion due to predefined criteria (n = 108), 54 studies underwent title and abstract screening, of which 33 were excluded. A total of 21 studies underwent full-text review, of which 6 studies were ultimately included in this systematic review (Figure 1). Exclusion reasons were: (1) unavailable data on expectant management (n = 8); (2) unavailable data on preterm birth (n = 5); (3) case series (n = 1); and (4) different or inconsistent gestational age at recruitment between the 2 arms (n = 1) (

Table 1. Results of computerized database search and study selection for embryo reduction to singletons vs. expectant management in dichorionic triplet pregnancies.

Search Outcome Reason for Exclusion	Retrieved References to Studies
	Management of DCTA Triplets (n = 21)
Excluded	15 (71.4)
1. No data on expectant management of triplets	8 (38)
2. No data on preterm birth < 34 weeks	5 (23.8)
3. Case series reporting five cases or fewer	1 (4.8)
4. Difference in gestational age at recruitment between expectant and MFR groups	1 (4.8)
Included	6 (28.6)

Note: Numbers in parentheses are percentages (%). DCTA, dichorionic triamniotic. MFR, multifetal reduction.

, Figure 2 and

Table A1. Excluded studies list and reason(s) for exclusion [7,18,40,41,42,43,44,45,46,47,48,49,50,51].

Author	Year	Reason for Exclusion
Liu et al. [18]	2022	No data on preterm birth < 34 weeks Different outcome measures
2. Kim et al. [40]	2021	No data on expectant management of triplets Different MFR technique
3. Shaw et al. [41]	2021	No data on preterm birth < 34 weeks
4. Zemet et al. [42]	2020	No data on expectant management of triplets
5. Zhou et al. [43]	2019	No data on expectant management of triplets Different MFR technique
6. Liu et al. [44]	2019	No data on preterm birth < 34 weeks Different outcome measures
7. Sun et al. [51]	2018	No data on preterm birth < 34 weeks
8. Lin et al. [45]	2016	No data on expectant management of triplets
9. Yanaihara et al. [46]	2017	Case series reporting on five cases or fewer (1 MFR)
10. Athanasiadis et al. [47]	2005	No data on expectant management of triplets
11. Okyay et al. [48]	2014	No data on expectant management of triplets
12. van de Mheen et al. [7]	2014	Difference in gestational age at recruitment between expectant and MFR groups (80/7–146/7 and 100/7–156/7 weeks, respectively)
13. Li et al. [49]	2013	No data on expectant management of triplets Different MFR technique
14. Geipel et al. [50]	2005	No data on MFR Not possible to calculate the numbers of early preterm deliveries
15. De Catte et al. [52]	2002	No data on expectant management of triplets Different outcome measures

MFR, multifetal reduction.

).

3.2. Study Range and Characteristics

From the six articles included, 12 datasets were retrieved [8,17,29,30,31,32]. Expectant management was evaluated in six studies with six datasets [8,17,29,30,31,32]. MFR (of the MC pair) to singleton pregnancy was evaluated in six articles with six datasets (

Table 2. Studies reporting on outcome of DCTA triplet pregnancies managed either expectantly or with ER to singletons [8,17,29,30,31,32].

Management	GA (wks)	Participants (N)	Miscarriage (%)	Cut-Off (wks)	Preterm Birth (%) *
DCTA expectant
Abel et al., 2016	10–14	16	6.3 (1/16)	<34	93.3 (14/15)
Cai et al., 2020	11–13	65	10.8 (7/65)	<32	22.4 (13/58)
Chaveeva et al., 2013	10–14	123	8.1 (10/123)	<34	69.0 (78/113)
Morlando et al., 2015	10–14	77	9.1 (7/77)	<33	24.1 (17/70)
Skiadas et al., 2010	11–13	13	15.4 (2/13)	<34	81.8 (9/11)
Van de Mheen et al., 2016	8–14	42	9.5 (4/42)	<32	44.7 (17/38)
Total			9.2 (31/336)		48.5 (148/305)
DCTA MFR (of the MC pair) to 1
Abel et al., 2016	10–14	30	20.0 (6/30)	<34	4.2 (1/24)
Cai et al., 2020	11–13	84	8.3 (7/84)	<32	2.6 (2/77)
Chaveeva et al., 2013	10–14	29	13.8 (4/29)	<34	1.3 (2/25)
Morlando et al., 2015	10–14	10	20.0 (2/10)	<33	1.3 (1/8)
Skiadas et al., 2010	11–13	13	23.1 (3/13)	<34	10.0 (1/10)
van de Mheen et al., 2016	8–14	46	19.6 (9/46)	<32	27.0 (10/37)
Total			14.6 (31/212)		9.4 (17/181)

*, excluding miscarriages. GA, gestational age. DCTA, dichorionic triamniotic. MFR, multifetal reduction. MC, monochorionic pair.

) [8,17,29,30,31,32]. The procedure was performed transabdominally by using a 20G or 22G needle and intracardiac or intrathoracic injection of potassium chloride [8,17,29,30,31,32]. No other MFR techniques were used in the included studies.

From the six articles included, the cutt-off for preterm birth was defined as <34 weeks in 3 studies (

Table 3. Studies reporting on outcome of DCTA triplet pregnancies managed either expectantly or with ER to singletons using a cut-off for preterm birth < 34 weeks [29,30,31].

Management	GA (wks)	Participants (N)	Miscarriage (%)	Cut-Off (wks)	Preterm Birth (%) *
DCTA expectant
Abel et al., 2016	10–14	16	6.3 (1/16)	<34	93.3 (14/15)
Chaveeva et al., 2013	10–14	123	8.1 (10/123)	<34	69.0 (78/113)
Skiadas et al., 2010	11–13	13	15.4 (2/13)	<34	81.8 (9/11)
Total			8.6(13/152)		72.7(101/139)
DCTA MFR (of the MC pair) to 1
Abel et al., 2016	10–14	30	20.0 (6/30)	<34	4.2 (1/24)
Chaveeva et al., 2013	10–14	29	13.8 (4/29)	<34	1.3 (2/25)
Skiadas et al., 2010	11–13	13	23.1 (3/13)	<34	10.0 (1/10)
Total			18.1(13/72)		6.8 (4/59)

*, excluding miscarriages. GA, gestational age. DCTA, dichorionic triamniotic. MFR, multifetal reduction. MC, monochorionic pair.

) [29,30,31], <32 weeks in 2 studies (

Table 4. Studies reporting on outcome of DCTA triplet pregnancies managed either expectantly or with ER to singletons using a cut-off for preterm birth < 32 weeks [17,32].

Management	GA (wks)	Participants (N)	Miscarriage (%)	Cut-Off (wks)	Preterm Birth (%) *
DCTA expectant
Cai et al., 2020	11–13	65	10.8 (7/65)	<32	22.4 (13/58)
van de Mheen et al., 2016	8–14	42	9.5 (4/42)	<32	44.7 (17/38)
Total			10.3(11/107)		31.3(30/96)
DCTA MFR (of the MC pair) to 1
Cai et al., 2020	11–13	84	8.3 (7/84)	<32	2.6 (2/77)
van de Mheen et al., 2016	8–14	46	19.6 (9/46)	<32	27.0 (10/37)
Total			12.3(16/130)		10.5(12/114)

*, excluding miscarriages. GA, gestational age. DCTA, dichorionic triamniotic. MFR, multifetal reduction. MC, monochorionic pair.

) [17,32], and <33 weeks in 1 study [8].

3.3. Study Quality and Potential Sources of Bias

Study quality was assessed in six articles, and all six retrospective studies were included in the systematic review (

Table A2. GoodReseArch for Comparative Effectiveness (GRACE) and Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (QATOCS/QATCSS) scores of the included studies [8,17,29,30,31,32].

Author	GRACE (GoodReseArch for Comparative Effectiveness) Checklist	Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies
Abel et al., 2016	Pass, D: Y5N1O0/M: Y4N1O0	9/14, Fair (Y9N2O3)
Cai et al., 2020	Pass, D: Y6N0O0/M: Y4N1O0	9/14, Fair (Y9N2O3)
Chaveeva et al., 2013	Pass, D: Y6N0O0/M: Y4N1O0	9/14, Fair (Y9N2O3)
Morlando et al., 2015	Pass, D: Y5N1O0/M: Y4N1O0	9/14, Fair (Y9N2O3)
Skiadas et al., 2010	Pass, D: Y5N1O0/M: Y4N1O0	8/14, Fair (Y8N3O3)
Van de Mheen., 2015	Pass, D: Y5N1O0/M: Y4N1O0	9/14, Fair (Y9N2O3)

Y, yes. N, no. O, other (i.e., cannot determine, not applicable, not reported). Good (4 ≥ Y ≥ 0), Fair (9 ≥ Y ≥ 5), Poor (14 ≥ Y ≥ 10). D, data. M, methods. O, other (i.e., not applicable).

) (Figure A1). No RCTs or prospective cohort studies were identified. Consequently, the overall certainty of evidence derived from this review is inherently limited, with most evidence likely rated as low or very low certainty if formally assessed.

This review identified several imperfections in study characteristics among the included studies: for instance, the method of data collection was retrospective in all six (100%); patient selection was non-consecutive or unknown in all six articles (100%); the outcome assessors were not blinded to the exposure status of the participants in all six (100%); reporting of study population was insufficient in two (33.3%) [8,29]; although there was sufficient description of MFR procedures in all six (100%).

The potential for reporting biases was carefully assessed qualitatively, given the constrictions of only a few articles analyzed and subsequently a limited amount of data sets available. The visual inspection of the derived funnel plots, both of standard error by Logit risk ratio and of precision (1/Std Err) by Logit risk ratio, appears asymmetrical (Comprehensive Meta-analysis Version 3.3.070, 2014, Biostat Inc., 14 North Dean Street, Englewood, NJ, USA) (Figure A2, Figure A3, Figure A4 and Figure A5).

The classical measure of heterogeneity is Cochran’s Q, which is included in each meta-analysis function because it forms part of the Der Simonian–Laird random effects pooling method [33]. Gavanghan et al., in 2000, indicated that Q has low power as a comprehensive test of heterogeneity, especially when the number of studies is small, i.e., most meta-analyses [34]. Conversely, Q has too much power as a test of heterogeneity if the number of studies is large as demonstrated by Higgins et al. in 2003 [35]. Due to the inherent heterogeneity of study designs and outcomes in the included studies, there is a tendency to advise using random effect models for the pooling of all observational studies [36].

Quantity I² describes the percentage of total variation across studies that is due to true heterogeneity rather than chance, thus quantifying the effect of heterogeneity and providing a measure of the degree of inconsistency in the studies’ results. A value of 0% indicates no observed heterogeneity and larger values show increasing heterogeneity. The high I² values show that most of the variability across studies is due to heterogeneity rather than chance. Statistical analyses were performed using MedCalc for Windows, version 12.7 (MedCalc Software, Mariakerke, Belgium), Review Manager (RevMan), Version 5.3, 2014 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration) and MetaAnalysisOnline.com assessed on 15 January 2026: Web-Based Tool for the Rapid Meta-Analysis of Clinical and Epidemiological Studies [37].

3.4. Combined Data

In DCTA triplet pregnancies managed expectantly (n = 336), the rates of miscarriage and preterm birth < 34 weeks were 9.2% and 48.5% (excluding miscarriages), respectively. MFR to singleton pregnancy in DCTA triplets (n = 212), is associated with a statistically non-significant increase in miscarriage rate (14.6% vs. 9.2%, p = 0.07 and RR = 1.53, 95% 0.91–2.55) (Figure 3) (Figure A6) and a statistically significant decrease in preterm birth rate (9.4% vs. 48.5%, p < 0.0001 and RR = 0.19, 95%CI 0.07–0.51) (Figure 4 and Figure 5) (Figure A7). In DCTA triplets, it was calculated, based on percentages (i.e., 9.4% vs. 48.5%), that 2.56 MFRs to singleton pregnancy need to be performed to prevent 1 preterm birth [38,39].

From the six articles included, the cut-off for preterm birth was defined as <34 weeks in three studies (Figure 6 and Figure 7) [29,30,31] and <32 weeks in two studies (Figure 8 and Figure 9) [17,32].

A subgroup of three studies were the cut-off for preterm birth was defined as <33 weeks was further evaluated (Figure 10 and Figure 11) [8,17,32].

4. Discussion

The primary result of our study is that MFR to singleton pregnancy was associated with a reduced rate of preterm birth but not associated with an increased rate of miscarriage in DCTA triplets.

The greatest strength of our systematic review is that contrary to current literature, it only included studies with both intervention and expectant arm. This increases the strength of the review by directly comparing outcomes within the same studies, reducing confounding variables and enhancing reliability. Therefore, the possibility of selective reporting bias is reduced but cannot be entirely ruled out. The number of patients included in the analysis is high, thus allowing some quantitative synthesis.

Nevertheless, this systematic review has several limitations, mainly due to significant heterogeneity among studies. Common limitations were lack of data on parity, rates of death beyond neonatal period, and neurodevelopmental impairment in survivors. Moreover, none of the included studies conducted a subgroup analysis of different time periods concerning pregnancy outcomes in relation to recent advances in neonatal intensive care and obstetric care, which have improved the outcome for younger and lighter neonates (

Table A3. Data on maternal age, method of conception, gestation at recruitment or multifetal embryo reduction and pregnancy outcome in dichorionic triamniotic triplet pregnancies [8,17,29,30,31,32].

	DCTA Triplet Pregnancies
	Expectant Management	MFR (of the MC Pair) to 1
Maternal age, years
Abel et al., 2016	34.0 (29.3–38.7)	34.1 (28.8–39.4)
Cai et al., 2020	29.6 (25.4–33.8)	28.4 (24.7–32.1))
Chaveeva et al., 2013	34.3 (21.7–46.8) a	35.3 (23.2–46.0) a
Morlando et al., 2015	-	-
Skiadas et al., 2010	-	-
van de Mheen et al., 2016	32.5 (28.7–35.5) b	33.8 (30.3–36.1) b
Total	32.6	32.9
Recruitment gestation, wks
Abel et al., 2016	-	-
Cai et al., 2020	-	-
Chaveeva et al., 2013	12.3 (9.3–14.0) a	12.1 (10.0–14.0) a
Morlando et al., 2015	-	-
Skiadas et al., 2010	-	-
van de Mheen et al., 2016	-	-
Total	12.3	12.1
Delivery gestation, wks
Abel et al., 2016	30.9 (27.7–34.1)	37.7 (36.1–39.3)
Cai et al., 2020	33.4 (30.4–36.4)	38.5 (36.4–40.6)
Chaveeva et al., 2013	33.0 (25.0–38.5) a,	39.1 (30.0–41.7) a,
Morlando et al., 2015	-	-
Skiadas et al., 2010	-	-
van de Mheen et al., 2016	32.8 (29.6–34.3)	38.7 (35.2–40.3)
Total	32.5	38.5
Conception
• Spontaneous
Abel et al., 2016	12.5 (2/16)	16.7 (5/30)
Cai et al., 2020	-	-
Chaveeva et al., 2013	18.7 (23/123)	0.0 (0/29)
Morlando et al., 2015	19.5 (15/77)	40.0 (4/10)
Skiadas et al., 2010	0.0 (0/13)	0.0 (0/13)
van de Mheen et al., 2016	61.9 (26/42)	43.5 (20/46)
Total	24.4 (66/271)	22.7 (29/128)
• Ovulation induction
Abel et al., 2016	0.0 (0/16)	6.7 (2/30)
Cai et al., 2020	-	-
Chaveeva et al., 2013	5.7 (7/123)	0.0 (0/29)
Morlando et al., 2015	1.3 (1/77)	0.0 (0/10)
Skiadas et al., 2010	0.0 (0/13)	0.0 (0/13)
van de Mheen et al., 2016	-	-
Total	3.5 (8/229)	2.4 (2/82)
• IVF/ICSI/IUI (ART)
Abel et al., 2016	87.5 (14/16)	76.7 (23/30)
Cai et al., 2020	100 (65/65)	100 (84/84)
Chaveeva et al., 2013	75.6 (93/123)	100.0 (29/29)
Morlando et al., 2015	62.3 (48/77)	60.0 (6/10)
Skiadas et al., 2010	100.0 (13/13)	100.0 (13/13)
van de Mheen et al., 2016	-	-
Total	79.3 (233/294)	93.4 (155/166)
• Unknown
Abel et al., 2016	0.0 (0/16)	0.0 (0/30)
Cai et al., 2020	-	-
Chaveeva et al., 2013	0.0 (0/123)	0.0 (0/29)
Morlando et al., 2015	16.9 (13/77)	0.0 (0/10)
Skiadas et al., 2010	0.0 (0/13)	0.0 (0/13)
van de Mheen et al., 2016	38.1 (16/42)	56.5 (26/46)
Total	10.7 (29/271)	20.3 (26/128)
Pregnancies with no survivors
Abel et al., 2016	6.3 (1/16)	20.0 (6/30)
Cai et al., 2020	6.9 (4/58)	1.3 (1/77)
Chaveeva et al., 2013	8.9 (11/123)	17.2 (5/29)
Morlando et al., 2015	9.1 (7/77)	20.0 (2/10)
Skiadas et al., 2010	15.4 (2/13)	23.1 (3/13)
van de Mheen et al., 2016	14.3 (6/42)	19.6 (9/46)
Total	9.4 (31/329)	12.7 (26/205)
Pregnancies with 1 survivor
Abel et al., 2016	6.3 (1/16)	80.0 (24/30)
Cai et al., 2020	-	98.7 (76/77)
Chaveeva et al., 2013	5.7 (7/123)	82.8 (24/29)
Morlando et al., 2015	15.6 (12/77)	80.0 (8/10)
Skiadas et al., 2010	23.1 (3/13)	76.9 (10/13)
van de Mheen et al., 2016	-	-
Total	10.0 (23/229)	89.3 (142/159)
Pregnancies with 2 survivors
Abel et al., 2016	18.8 (3/16)	-
Cai et al., 2020	-	-
Chaveeva et al., 2013	12.2 (15/123)	-
Morlando et al., 2015	19.5 (15/77)	-
Skiadas et al., 2010	23.1 (3/13)	-
van de Mheen et al., 2016	-	-
Total	15.7 (36/229)	-
Pregnancies with 3 survivors
Abel et al., 2016	68.8 (11/16)	-
Cai et al., 2020	-	-
Chaveeva et al., 2013	73.2 (90/123)	-
Morlando et al., 2015	55.8 (43/77)	-
Skiadas et al., 2010	38.5 (5/13)	-
van de Mheen et al., 2016	71.4 (30/42)	-
Total	66.1 (179/271)	-
Pregnancies with at least 1 survivor
Abel et al., 2016	93.8 (15/16)	80.0 (24/30)
Cai et al., 2020	93.1 (54/58)	98.7 (76/77)
Chaveeva et al., 2013	91.1 (112/123)	82.8 (24/29)
Morlando et al., 2015	90.9 (70/77)	80.0 (8/10)
Skiadas et al., 2010	84.6 (11/13)	76.9 (10/13)
van de Mheen et al., 2016	85.7 (36/42)	80.4 (37/46)
Total	90.6 (298/329)	87.3 (179/205)

DCTA, dichorionic triamniotic. MFR, multifetal embryo reduction, MC, monochorionic. IVF/ICSI/IUI (ART), in vitro fertilization/intra-cytoplasmic sperm injection/intra-uterine insemination (assisted reproduction technologies). Values represent the means ± SD or n (%). ^a Values represent medians and range. ^b Excluding pregnancies with delivery < 24 weeks.

). No RCTs were identified from our literature search. These limitations highlight the necessity for cautious interpretation of synthesized results and advocate for improved transparency in reporting future research.

The abovementioned accurately reflect the weaknesses of our review methods: (1) there were only a small number of papers with small sample sizes and potentially suffering from bias that we were able to include, (2) the cut-off for preterm birth was not consistent in all studies included, and (3) non-English publications were missed. Furthermore, the potential for publication and reporting biases, particularly given the reliance on retrospective studies, should also be acknowledged as a limitation impacting the certainty of conclusions drawn.

Chaveeva et al. calculated that MFR to singleton pregnancy adds 4.5% to the rate of miscarriage noted in DCTA triplets managed expectantly [30]. Despite including intervention-only studies, Morlando et al. reported that in MFR to singleton pregnancy the rate of miscarriage was 14.5% (95% CI 7.6–26.2%) [8]. Although inconclusive, our results agree with current literature. In our study, the rate of miscarriage was 14.6%, which is comparable to that reported by the largest studies included [17,30]. However, concerning MFR to singleton pregnancy, ethical considerations arise both for the parents and the health practitioners including: (1) the parents’ right to decide as well as the psychological impact of their decision, and (2) respecting the principles of beneficence and/or non-maleficence as well as the potential for proportionality between benefits and harms.

Regarding implications for current practice and future research, when the top priority is three live-born infants, expectant management seems to be a reasonable choice. However, if the priority is to decrease the rate of preterm birth, MFR to singleton pregnancy could be discussed (Table A3). Given the fact that in the MFR group the MC pair is reduced only to decrease the rate of preterm birth, appropriate medico-legal and socio-ethical counselling and justification are important. Parental autonomy, cultural and legal variability across countries, and psychological morbidity after MFR should be considered. MFR should not only be viewed as technical intervention but rather as a value-laden, preference-sensitive decision. Outcome metrics alone are insufficient to guide counseling and patient-centered values may outweigh statistical risk reduction.

Included studies span over two decades (2000–2025), during which neonatal survival at <34 weeks and obstetric management of multiples have dramatically evolved. Pooling outcomes across eras risks temporal confounding. Observed benefits may partially reflect era effects rather than intervention effects. In the absence of RCTs, data from systematic reviews appear to be the best existing evidence for counseling on the different options available. Future research should aim for standardized methodologies and clearer reporting practices to enable more robust certainty assessments and provide stronger evidence to guide clinical practice.

5. Conclusions

In DCTA triplet pregnancies, our results suggest the following:

• MFR to singleton pregnancy is associated with a lower rate of preterm birth but not a higher rate of miscarriage. Because of modest numbers, these results should be interpreted with caution.
• If the cut-off for preterm delivery is lowered <33 weeks, MFR to singleton pregnancy could prove more efficient for an equal rate of miscarriage < 24 weeks.
• Regardless of personal values conflict with parents’ request, unbiased information must be available, or the parents should be referred elsewhere.