SEARCH

SEARCH BY CITATION

Keywords:

  • Birth order;
  • caesarean section;
  • multiple pregnancies;
  • twin delivery;
  • twins

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

Please cite this paper as: Rossi A, Mullin P, Chmait R. Neonatal outcomes of twins according to birth order, presentation and mode of delivery: a systematic review and meta-analysis. BJOG 2011;118:523–532.

Background  The optimal mode of delivery for twins is undetermined.

Objective  To review literature regarding the neonatal outcomes following twin delivery.

Data sources  Searches were conducted in PubMed, Medline, Embase, Cochrane library and reference lists.

Selection criteria  Studies selection criteria were: both twins alive at labour, outcomes stratified for birth order, presentation, planned and actual delivery mode. Eighteen articles were included in the meta-analysis (39 571 twin sets).

Data collection and analysis  The Meta-analysis of Observational Studies in Epidemiology guidelines were followed. Interstudy heterogeneity (I2) was tested. A fixed model was generated whenever I2 < 25%. Pooled odds ratios (OR) with 95% CI were computed. Intergroup comparison was significant if 95% CI did not encompass 1. The first and second twins were indicated as Twin A (TA) and Twin B (TB), respectively.

Main results  Neonatal morbidity was lower in TA than TB (3.0 versus 4.6%; OR 0.53; 95% CI 0.39–0.70). TA experienced neonatal death less often than TB (0.3 versus 0.6%; OR 0.55; 95% CI 0.38–0.81). No differences were noted between vertex and non-vertex and attempted vaginal delivery versus planned caesarean section in either TA or TB. In TA, neonatal morbidity was lower after vaginal delivery (1.1%) than caesarean section (2.2%; OR 0.47; 95% CI 0.27–0.82). Neonatal death was not associated with actual delivery mode. In TB, morbidity following combined delivery (19.8%) was higher than after vaginal delivery (9.5%; OR 0.55; 95% CI 0.41–0.74) or caesarean section (9.8%; OR 0.47; 95% CI 0.43–0.53). When outcomes were stratified for both presentation and delivery mode, mortality rate was lower after vaginal delivery than caesarean section for both vertex and nonvertex TB.

Author’s conclusion  An attempt at vaginal delivery should be considered in twin pregnancies with vertex/vertex presentation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

Twin pregnancies account for approximately 3% of all gestations.1 The influence of birth order on neonatal outcomes is still unclear. The second twin is generally considered at higher risk of severe morbidity and mortality because of obstetric complications that may occur after delivery of the first twin, including placental separation, cord prolapse, uterine atony, long interval delivery and cervical spasm.2,3 Perinatal outcomes of the nonpresenting twins also seem to be associated with inter-twin birthweight discordance4 and very low birthweight.5

There is general consensus that vaginal delivery for twins is safe when both are in vertex presentation, whereas planned caesarean section is typically indicated for breech presentation of the first twin.6,7 This consensus is based on expert opinion rather than randomised clinical trials.8 In fact, studies on the effect of presentation, mode of delivery and birth order have produced conflicting results. Whereas planned vaginal delivery has been associated with an increased risk of perinatal mortality and morbidity of the second twin compared with the first twin,8–11 data from other series did not demonstrate any benefit if caesarean delivery was planned.12–14 The only randomised study of mode of delivery in twin pregnancy was performed towards the end of the 1980s and demonstrated that there was little difference in neonatal morbidity between twins delivered vaginally and those delivered by caesarean section.14 Hence, the optimal delivery mode for twins remains controversial. Therefore, as twin delivery remains a challenge in obstetric practice, we performed a review of the current literature in an attempt to define whether birth order, mode of delivery and presentation are associated with adverse neonatal outcomes.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

A systematic review protocol with prespecified criteria for study selection, outcome measurements and analysis was developed. A search in PubMed, Medline, EMBASE and the Cochrane database was performed to identify relevant articles that reported neonatal outcomes with regard to birth order, presentation and mode of delivery. A manual search of reference lists from included studies and review articles was also performed. Because of the advances in obstetric, anaesthestic and neonatal care in the last 10 years, the search was performed from January 2001 to May 2010. Key words were: twins, vaginal delivery, c(a)esarean section, vertex presentation, nonvertex presentation, breech presentation, transverse presentation, birth order, neonatal morbidity and neonatal mortality. Articles were included if the study population was unselected and represented by both twins alive at labour, outcomes were described according to birth order, presentation, planned and actual mode of delivery, and data were reported as proportional rates. Articles were excluded if they did not meet at least one of the inclusion criteria, were aimed to analyse inter-twin delivery interval, were performed in a selected population (i.e. twins with low birthweight, previous caesarean section, short cervix), included higher multiple-order pregnancies, or reported data in graphs or percentages. Studies based on national registries were included if no overlap of population with original articles was determined by examining the aim of the study, the period of time, the study sample and size, and the country. Non-English language publications, letters and personal communications were also excluded.

Sample size, gestational age at delivery, birth order, presentation of twins at birth, mode of delivery, and neonatal morbidity and mortality were extracted from each study. Birth order was defined as Twin A and Twin B to indicate the presenting first and second twins, respectively. Presentation was classified as vertex or nonvertex, which included breech and transverse positions. Mode of delivery referred to planned and actual delivery by vaginal route of both twins, caesarean section of both twins, and combined delivery, which was defined as vaginal delivery of Twin A and caesarean section of the co-Twin B. Neonatal morbidity was defined as pH < 7.0, Apgar score < 7 at 5 minutes and any neonatal birth trauma. Neonatal mortality was defined as death within 28 days. These definitions were specified according to a predefined protocol.

Neonatal outcomes were compared for Twin A versus Twin B, vertex versus nonvertex presentation, attempted vaginal delivery versus planned caesarean section, and vaginal delivery versus caesarean section versus combined delivery. When articles stratified neonatal outcomes according to more than one gestational age at delivery, the latest gestational age was used in an attempt to minimise the incidence of poor outcomes related to prematurity rather than the mode of delivery. Outcomes according to presentation and mode of delivery were stratified for birth order. Where data were missing, an effort to contact the corresponding author was made in an attempt to obtain unpublished data.

Study selection and data extraction were performed by the three authors independently by following the Meta-analysis of Observational Studies in Epidemiology guidelines. Discordance was resolved by consensus. Comparative analyses were performed with regard to Twin A versus Twin B, vertex versus nonvertex presentation, planned vaginal delivery versus planned caesarean section (intention to treat), and vaginal delivery versus caesarean section versus combined delivery.

The use of funnel plots to assess publication bias was not considered because asymmetry of the funnel plot, either statistically tested or visually interpreted, does not accurately predict publications bias.15 For this purpose, interstudy heterogeneity, defined according to Higgins et al.16 as the percentage of total variation across studies due to heterogeneity rather than chance (I2), was tested with chi-square test for heterogeneity at a significant level of P = 0.10. A random effect model was generated whenever the I2 statistic was >25%. Intergroup comparison was considered statistically significant at an alpha level of two-tailed P < 0.05. When a statistical difference was determined, the calculation of pooled odds ratios (OR) with 95% confidence interval (CI) was performed. The odds ratio was preferred to the relative risk and the risk difference because the former is the ratio of the probability that a particular event will occur to the probability that it will not occur, whereas the risks describe only the probability with which the event will occur. Meta-analysis was performed with RevMan (Revision Manager, Version 4.2 for Windows, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration 2003).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

The review process is summarised in Supplementary material, see Figure S1. Eighteen articles were included in the meta-analysis (Table 1).8,10,11,17–31 Twelve articles17,19–29 reported neonatal outcomes for both twins, of which all but four19–21,25 described better outcomes for Twin A than Twin B. Overall, neonatal morbidity was lower in Twin A (1224 of 39 571; 3.0%) than Twin B (1842 of 39 571, 4.6%; Z = 4.34; P < 0.001; OR 0.53; 95% CI 0.39–0.70) (Figure 1A) (not reported in one article28). Similarly, Twin A was less likely than Twin B to experience neonatal death (108 of 31 854, 0.3% versus 193 of 31 854, 0.6%; Z = 3.06; P = 0.002; OR 0.55; 95% CI 0.38–0.81) (Figure 1B) (not reported in three articles19,20,22).

Table 1.   Characteristics of the included study
Author and ref. no.YearObjectiveSample size (n)Study sampleMorbidity n (%)Mortality n (%)
Caukwell182002Mode of delivery and presentation    
 Vertex vaginal delivery183Twin B56 (30.6)5 (2.7)
 Vertex caesarean section54Twin B42 (77.7)2 (3.7)
 Nonvertex vaginal delivery117Twin B62 (53)7 (5.9)
 Nonvertex caesarean section68Twin B40 (58.8)1 (1.4)
Usta292002Birth order461Twin A367 (79.6)28 (6.0)
Twin B427 (92.6)39 (8.4)
Wen3,102004Mode of delivery    
 Vaginal delivery5970Twin B56 (1)582 (1.0)
 Caesarean section7471Twin B66 (0.8)717 (1.0)
 Combined delivery5842Twin B1283 (22)79 (1.3)
Hartley222005Birth order5138Twin A236 (4.6)NA
Twin B501 (9.7)NA
Ginsberg202005Birth order10 365Twin A196 (1.8)NA
Twin B180 (1.7)NA
Mode of delivery    
 Vaginal delivery4599Twin A148 (3.2)NA
Twin B143 (3.1)NA
 Caesarean section5247Twin A40 (0.7)NA
Twin B28 (0.5)NA
 Combined delivery518Twin B9 (1.7)NA
Presentation    
 Vertex6940Twin A137 (2)NA
7366Twin B145 (2)NA
 Nonvertex2142Twin A37 (1.7)NA
2999Twin B42 (1.4)NA
Usta302005Mode of delivery    
 Nonvertex vaginal delivery138Twin B17 (12.3)15 (10.8)
 Nonvertex caesarean section79Twin B4 (5)3 (3.8)
Haest212005Planned vaginal delivery135Twin A4 (3)0
Twin B8 (6)1 (0.7)
Planned caesarean section29Twin A00
Twin B3 (10)0
Smith282005Neonatal death and birth order8073Twin ANA6 (0.1)
Twin BNA30 (0.3)
Yang31,322005Presentation and mode of delivery    
 Nonvertex vaginal delivery7113Twin B1381 (19.4)107 (1.5)
 Nonvertex caesarean section5723Twin B801 (14)57 (1)
 Nonvertex combined delivery2349Twin B609 (26)27 (1.1)
Yang82006 Vertex vaginal delivery46 071Twin B3368 (7.3)287 (0.6)
 Vertex caesarean section36 977Twin B296 (0.8)288 (0.7)
 Vertex combined delivery2993Twin B423 (14.1)41 (1.4)
Sibony262006Mode of delivery    
 Vertex vaginal delivery426Twin A3 (0.7)1 (0.2)
 Nonvertex vaginal delivery103Twin A01 (1)
 Vertex vaginal delivery303Twin B10 (3.3)1 (0.3)
 Nonvertex vaginal delivery226Twin B3 (1.3)0
Sentilhes252007Planned vaginal delivery124Twin A8 (6.4)1 (0.8)
Twin B4 (3.2)0
Planned caesarean section71Twin A7 (9.8)0
Twin B4 (5.6)1 (1.4)
Smith272007Birth order1377Twin A19 (1.4)41 (3)
Twin B50 (3.6)80 (5.8)
Mode of delivery    
 Vaginal delivery457Twin A7 (1.5)18 (4)
Twin B12 (2.6)38 (8.3)
 Caesarean section416Twin A29 (7)23 (5.5)
Twin B21 (5)42 (10.1)
Bjelic-Radisic172007Mode of delivery    
 Vaginal delivery219Twin A5 (2.2)0
Twin B23 (10.5)0
 Caesarean section48Twin A7 (14.5)0
Twin B6 (12.5)0
 Combined delivery14Twin B7 (50)0
Presentation    
 Vertex171Twin B5 (3)0
 Non vertex48Twin B6 (12.5)0
Schmitz242008Planned vaginal delivery516Twin A75 (14.5)0
Twin B118 (22.8)0
Planned caesarean section40Twin A6 (15)0
Twin B10 (25)0
Herbst232008Birth order20 232Twin A286 (1.4)30 (0.1)
Twin B483 (2.3)42 (0.2)
Mode of delivery    
 Vaginal delivery12 572Twin A150 (1.2)20 (0.1)
Twin B326 (2.6)27 (0.2)
 Caesarean section7660Twin A136 (1.7)10 (0.1)
Twin B157 (2)15 (0.1)
Presentation    
 Vertex17 406Twin A248 (1.4)30 (0.2)
 Nonvertex2826Twin A38 (1.3)0
Fox192010Planned vaginal delivery130Twin A2 (1.5)NA
Twin B7 (5.3)NA
Planned caesarean section157Twin A3 (2)NA
Twin B5 (3.2)NA
image

Figure 1.  (A) Morbidity of Twin A versus Twin B; (B) mortality of Twin A versus Twin B.

Download figure to PowerPoint

Meta-analysis of twin A

Three articles analysed the outcomes of Twin A with regard to presentation.20,23,26 Overall, vertex presentation of Twin A accounted for 12 723 of 15 491 (82.0%) and nonvertex presentation accounted for 2768 of 15 491 (18.0%). Neonatal morbidity affected 250 of 12 723 (2.0%) in the former and 49 of 2768 (1.7%) in the latter group (Z = 0.49; P = 0.62; OR 1.8; 95% CI 0.79–1.48). Neonatal mortality was seen in 11 of 5783 (0.2%) vertex Twin A and one of 626 (<1%) nonvertex Twin A without a statistical difference (Z = 0.12; P = 0.90; OR 0.90; 95% CI 0.15–5.21) (not specified in one article20).

Four articles analysed outcomes following attempted vaginal delivery versus planned caesarean section.19,21,24,25 In 905 of 1202 (75.3%) Twin A, a trial of labour was attempted, whereas in 297 of 1202 (24.7%), a caesarean section was planned. Neonatal morbidity rate was similar between the two groups (attempted vaginal delivery: 89 of 905, 9.8% versus caesarean section: 16 of 297, 5.4% respectively; Z = 0.51; P = 0.61; OR 1.17; 95% CI 0.63–2.17). No mortality rate was reported in any study.

Three articles that studied the actual mode of delivery had conflicting results. Whereas one article did not record a significant difference,27 the series by Bjelic-Radisic et al.17 and Herbst and Kallen23 noted that better outcomes were associated with vaginal delivery. When pooled, the delivery mode was 13 248 of 21 372 (61.9%) vaginal deliveries and 8124 of 21 372 (38.1%) caesarean sections. Neonatal morbidity of Twin A was significantly less frequent after vaginal delivery (148 of 13 248, 1.1%) than caesarean section (169 of 8124, 2.1%; Z = 2.55; P = 0.01; OR 0.40; 95% CI 0.20–0.81) (see Supplementary material, Figure S2). All the studies reported that neonatal death was not associated with mode of delivery (38 of 13 248, 0.3% versus 33 of 8124, 0.4%; Z = 0.58; P = 0.63).

The analysis of neonatal outcomes according to both presentation and mode of delivery could not be performed because data on this topic were not reported. Outcomes of Twin A deliveries are summarised in Table 2.

Table 2.   Outcomes of Twin A
 MorbidityMortality
  1. Data expressed as %.

Vertex20.2
Nonvertex1.7<1
P value0.620.90
Planned vaginal delivery9.80
Planned caesarean section5.40
P value0.61 
Vaginal delivery1.10.3
Caesarean section2.20.4
P value0.0080.61
 OR 0.47 (95% CI 0.27–0.82) 

Meta-analysis of twin B

Four articles stratified neonatal morbidity and mortality in vertex and nonvertex Twin B presentations17,18,20,26 (vertex: 94 518, 83.5%; nonvertex: 18 643, 16.5%). Although two studies observed a lower morbidity rate with vertex presentation of Twin B, this difference did not become statistically significant when the articles were pooled with a morbidity rate of 4345 (4.6%) and 2944 (15.7%) in vertex and nonvertex Twin B, respectively (Z = 1.36; P = 0.17; OR 0.49; 95% CI 0.18–1.36). All the studies found a similar mortality rate between vertex (eight of 711, 1.1%) and nonvertex (eight of 459, 1.7%) presentation of Twin B (Z = 0.53; P = 0.60; OR 0.77; 95% CI 0.29–2.02).

Four articles analysed outcomes following attempted vaginal delivery versus planned caesarean section.19,21,24,25 Overall, 905 of 1202 (75.3%) entered a trial of labour and 297 of 1202 (24.7%) opted for planned caesarean section. Neonatal morbidity rate was similar between the two groups (attempting vaginal delivery 137 of 905, 15.1% versus planned caesarean section: 22 of 297, 7.4%; Z = 0.38; P = 0.71; OR 1.11; 95% CI 0.65–1.88). Similarly, neonatal mortality rate of both twins was similar between attempted vaginal delivery (one of 140; 0.7%) versus planned caesarean section (Twin B: one of 775; 0.1%; Z = 1.03; P = 0.30; OR 3.04; 95% CI 0.37–2.52) (unreported in one article19).

The outcomes for Twin B with regard to the actual mode of delivery were investigated in ten articles,8,10,11,17,18,20,23,27,30,32, of which five did not detect a significant difference of morbidity and mortality rate between vaginal delivery and caesarean section. When pooled, there were 127 472 of 250 118 (51%) vaginal deliveries and 122 646 of 250 118 (49%) caesarean sections. Neonatal morbidity occurred equally between Twin B delivered vaginally (11 358 of 127 472, 9.0%) and those delivered by caesarean section (8866 of 122 646, 7.2%; Z = 0.83; P = 0.41; OR 1.33; 95% CI 0.68–2.61). Neonatal death did not differ according to the mode of delivery (vaginal delivery: 1019 of 72 840, 1.4% versus caesarean section: 1125 of 58 496, 2.0%; Z = 0.55; P = 0.58; OR 1.06; 95% CI 0.87–1.29) (unreported in one article20).

Five articles described morbidity8,11,17,20,32 and mortality8,10,17,20,32 of Twin B after combined delivery. In this subgroup, combined delivery of Twin B was performed in 11 716 of 240 089 (4.9%). Morbidity of Twin B born by combined delivery (2 331 of 11 716, 19.8%) was significantly higher than for Twin B delivered vaginally (10 873 of 114 005, 9.5%; Z = 3.94; P < 0.0001; OR 0.55; 95% CI 0.41–0.74) (see Supplementary material, Figure S3A) or by caesarean section (11 278 of 114 369, 9.8%; Z = 14.24; P < 0.00001; OR: 0.47; 95% CI: 0.43–0.53) (see Supplementary material, Figure S3B). Conversely, mortality rates of Twin B after combined delivery (147 of 11 198, 1.3%) was similar to those after vaginal delivery (976 of 109 406, 0.9%; Z = 1.03; P = 0.30; OR 0.76; 95% CI 0.45–1.28) and caesarean section (1119 of 109 122; 1.0%; Z = 0.35; P = 0.73; OR 0.91; 95% CI 0.52–1.57).

Five articles could be pooled with regard to both presentation and mode of delivery,8,18,20,30,32, which collectively included 90 347 (85%) vertex and 16 022 (15%) nonvertex presentation of Twin B. Of the former, 50 168 (55%) were delivered vaginally and 40 179 (45%) by caesarean section. Of the latter, 8053 (50%) and 7969 (50%) underwent vaginal and abdominal delivery, respectively. No differences were noted in morbidity rates between vertex presentation of Twin B delivered by caesarean section (356 of 40 179, 0.8%) and vaginally (3537 of 50 168, 0.7%, Z = 0.61; P = 0.54; OR 0.07; 95% CI 0.07–4.14). In contrast, the mortality rate was modestly reduced after vaginal delivery (292 of 46 174, 0.6%) compared with delivery by caesarean section (290 of 37 031, 0.8%; Z = 2.64; P = 0.008; OR 1.25; 95% CI 1.06–1.47) (not reported in one article20). Similar outcomes were obtained for nonvertex presentation of Twin B (morbidity: vaginal delivery: 855 of 7969, 10.7%, versus caesarean section: 1490 of 8053, 18.5%; Z = 1.79; P = 0.07; OR 0.45; 95% CI 0.19–1.08; mortality: vaginal delivery: 61 of 5870, 1%, versus caesarean section: 129 of 7368, 1.7%; Z = 3.18; P = 0.001; OR 0.61; 95% CI 0.45–0.83). Outcomes of Twin B deliveries are summarised in Table 3.

Table 3.   Outcomes of Twin B
 MorbidityMortality
  1. Data expressed in %.

  2. *Vaginal delivery versus combined delivery: OR 0.55 (95% CI 0.41–0.74); caesarean section versus combined delivery: OR 0.47 (95% CI 0.43–0.53).

Vertex4.61.1
Nonvertex15.71.7
P value0.170.60
Planned vaginal delivery15.10.7
Planned caesarean section7.40.1
P value0.710.30
Vaginal delivery91.4
Caesarean section7.22
Combined delivery19.81.3
P value<0.0001*0.73
Vertex vaginal delivery0.70.6
Vertex caesarean section0.80.8
P value0.540.008
  OR 1.25 (95% CI 1.06–1.47)
Nonvertex vaginal delivery10.71
Nonvertex caesarean section18.51.7
P value0.070.001
  OR 0.61 (95% CI 0.45–0.83)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

This systematic review describes perinatal outcomes of twin pregnancies at delivery. When the outcomes of the twins were compared for birth order, the second twin was at greater risk of serious perinatal morbidity than the first twin. This finding confirms the prevailing evidence that the second twin is more likely to experience adverse events than its co-twin.33,34

With regard to the first twin, neonatal morbidity and mortality did not differ between vertex and non-vertex infants. We also observed a similar rate of neonatal morbidity in fetuses delivered by planned caesarean section compared with those twins delivered after attempting a trial of labour. However, when the mode of delivery was analysed according to the actual delivery, adverse outcomes were more frequent in twins born by caesarean section than in infants born vaginally. The risk of neonatal death appeared to be independent of presentation and type of delivery.

With regard to the second twin, no significant differences were noted according to presentation. Although planned caesarean section was associated with a lower rate of neonatal morbidity compared with planned vaginal delivery, this difference did not reach statistical significance. When neonatal outcomes were calculated according to both mode of delivery and presentation we observed a trend for less morbidity in vertex and nonvertex fetuses delivered by vaginal delivery than caesarean section. In addition, a mild reduction of neonatal death occurred after vaginal delivery in both vertex and nonvertex presentation as compared with caesarean section. In our opinion, these findings do not justify an absolute policy of considering caesarean delivery as the optimal mode for delivery of twins.

It was noteworthy that analysis of the actual delivery showed that the highest morbidity rate occurred when the second twin was born by caesarean delivery following vaginal delivery of the co-twin. This suggests that the management of labour of twin pregnancies should take into account factors that could potentially lead to unsuccessful vaginal delivery of the second twin. Nonvertex presentations, cephalopelvic disproportion, and cord prolapse have been reported as contributory factors for caesarean section of the second twin after vaginal delivery of the first twin.3 In addition, emergency caesarean section of the second twin after vaginal delivery of the first twin increases the risk of serious complications by approximately 50%. It is reasonable to assume that all deliveries in the combined group are unplanned and caesarean section of the second twin occurs under emergency conditions such as placental abruption, umbilical cord prolapse and non-reassuring fetal heart rate; these factors further increase the risk of neonatal morbidity. Additionally, cervical spasm after the delivery of the first twin can complicate the vaginal delivery of the second twin.35

There are several limitations of this review, which are mainly attributed to a lack if specification of the study population. The reviewed articles were mainly performed in Western countries. It is likely that immigration leading to different ethnicity within country might have biased our results. In most of the reviewed articles, selective and emergency caesarean sections were included in a single group and authors did not specify whether elective caesarean delivery was performed for maternal or fetal indications. This might have influenced neonatal morbidity rates. In addition, studies were not randomised, so the decision to perform caesarean section or vaginal delivery was at the discretion of the delivering physician. Information on delivery decision-making was limited and it is quite possible that different physicians may have chosen different modes of delivery, although fetal and maternal conditions were similar. Furthermore, the reviewed articles do not describe neonatal outcomes following operative delivery. Therefore, selection bias could have affected our results. In addition, the outcomes were not categorised based on chorionicity, nor was this reported for twin sets. Information on this topic would be useful for counselling women carrying twin pregnancies. Biased estimates of the effects of mode of delivery could best be avoided by performing prospective, randomised clinical trials, taking into account these specific study population characteristics.

Another potential limitation is that we could not stratify outcomes with regard to birthweight. Neonatal death and morbidity have been associated with birthweight discrepancies of more than 20–40%9,36–38 or birthweight < 1500 g.8 Some studies also focused on perinatal outcomes in twins with a larger second twin. Usta et al.4 reported that vaginal delivery remains feasible when the second twin is over 250 g larger than the first twin. However, the association between birthweight and neonatal outcomes differed among studies; intertwin discordance was expressed by percentage of the larger twin or differences in grams4 in some articles, whereas in others outcomes for twins were categorised according to arbitrary cut-offs of birthweight.8 Such variations made it difficult to reconcile the results from various reports. In addition, the inclusion of very-low-birthweight infants would interfere with the neonatal morbidity secondary to the mode of delivery. Although preterm delivery is associated with an increased risk of adverse outcomes, we could not control for differences in neonatal outcomes by gestational age at birth. However, gestational age at birth is usually the same for both twins and is unlikely to influence the decision on the mode of delivery. Lastly, articles based on birth certificates and national registry might have been completed by nonmedical personnel without identification and correction of data leading to entry mistakes and coding errors.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

In the absence of more definitive data, our systematic review suggests that an attempt at vaginal delivery should be considered in twin pregnancies. Current literature shows that in twins with vertex/vertex presentation, vaginal delivery is safer than caesarean section for the first twin, and no differences are observed for the second twin after vaginal or caesarean section. In pregnancies with vertex presentation of the first twin and nonvertex presentation of the second twin, women should be counselled that trial of labour is a safe option in the absence of risk factors that may increase the risk of a caesarean delivery of the second twin after vaginal delivery of the first twin, although predicting combined delivery in the antenatal period is almost impossible. Multicentre, randomised studies are needed to assess guidelines about the optimal mode of delivery for twin pregnancies. However, randomised control trials might not be feasible to perform because delivery of twins is highly influenced by the operator’s skill and experience.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

ACR conceived and designed the manuscript and analysed data; PM analysed data; RC commented on results. All the authors selected the articles included in the meta-analysis, interpreted the results, contributed to writing the manuscript and gave their consensus for submission.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information
  • 1
    Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep 2003;52:1113.
  • 2
    MacKay AP, Berg CJ, King JC, Duran C, Chang J. Pregnancy-related mortality among women with multifetal pregnancies. Obstet Gynecol 2006;107:5638.
  • 3
    Wen SW, Fung KF, Oppenheimer L, Demissie K, Yang Q, Walker M. Occurrence and predictors of cesarean delivery for the second twin after vaginal delivery of the first twin. Obstet Gynecol 2004;103:4139.
  • 4
    Usta IM, Nassar AH, Abu Musa AA, Awwad JT, Yunis KA, Seoud MA. Perinatal outcome of vaginally delivered twin gestations with a larger twin B. J Perinatol 2003;23:40913.
  • 5
    Shinwell ES, Blickstein I, Lusky A, Reichman B. Effect of birth order on neonatal morbidity and mortality among very low birthweight twins: a population based study. Arch Dis Child Fetal Neonatal Ed 2004;89:F1458.
  • 6
    Blickstein I. Cesarean section for all twins? J Perinat Med 2000;28:16974.
  • 7
    Boggess KA, Chisholm CA. Delivery of the nonvertex second twin: a review of the literature. Obstet Gynecol Surv 1997;52:72835.
  • 8
    Yang Q, Wen SW, Chen Y, Krewski D, Fung Kee Fung K, Walker M. Neonatal mortality and morbidity in vertex–vertex second twins according to mode of delivery and birth weight. J Perinatol 2006;26:310.
  • 9
    Armson BA, O’Connell C, Persad V, Joseph KS, Young DC, Baskett TF. Determinants of perinatal mortality and serious neonatal morbidity in the second twin. Obstet Gynecol 2006;108:55664.
  • 10
    Wen SW, Fung Kee Fung K, Oppenheimer L, Demissie K, Yang Q, Walker M. Neonatal mortality in second twin according to cause of death, gestational age, and mode of delivery. Am J Obstet Gynecol 2004;191:77883.
  • 11
    Wen SW, Fung Kee Fung K, Oppenheimer L, Demissie K, Yang Q, Walker M. Neonatal morbidity in second twin according to gestational age at birth and mode of delivery. Am J Obstet Gynecol 2004;191:7737.
  • 12
    Greig PC, Veille JC, Morgan T, Henderson L. The effect of presentation and mode of delivery on neonatal outcome in the second twin. Am J Obstet Gynecol 1992;167:9016.
  • 13
    Hogle KL, Hutton EK, McBrien KA, Barrett JF, Hannah ME. Cesarean delivery for twins: a systematic review and meta-analysis. Am J Obstet Gynecol 2003;188:2207.
  • 14
    Rabinovici J, Barkai G, Reichman B, Serr DM, Mashiach S. Randomized management of the second nonvertex twin: vaginal delivery or cesarean section. Am J Obstet Gynecol 1987;156:526.
  • 15
    Lau J, Ioannidis JP, Terrin N, Schmid CH, Olkin I. The case of the misleading funnel plot. BMJ 2006;333:597600.
  • 16
    Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:55760.
  • 17
    Bjelic-Radisic V, Pristauz G, Haas J, Giuliani A, Tamussino K, Bader A, et al. Neonatal outcome of second twins depending on presentation and mode of delivery. Twin Res Hum Genet 2007;10:5217.
  • 18
    Caukwell S, Murphy DJ. The effect of mode of delivery and gestational age on neonatal outcome of the non-cephalic- presenting second twin. Am J Obstet Gynecol 2002;187:135661.
  • 19
    Fox NS, Silverstein M, Bender S, Klauser CK, Saltzman DH, Rebarber A. Active second-stage management in twin pregnancies undergoing planned vaginal delivery in a U.S. population. Obstet Gynecol 2010;115:22933.
  • 20
    Ginsberg NA, Levine EM. Delivery of the second twin. Int J Gynaecol Obstet 2005;91:21720.
  • 21
    Haest KM, Roumen FJ, Nijhuis JG. Neonatal and maternal outcomes in twin gestations > or =32 weeks according to the planned mode of delivery. Eur J Obstet Gynecol Reprod Biol 2005;123:1721.
  • 22
    Hartley RS, Hitti J. Birth order and delivery interval: analysis of twin pair perinatal outcomes. J Matern Fetal Neonatal Med 2005;17:37580.
  • 23
    Herbst A, Kallen K. Influence of mode of delivery on neonatal mortality in the second twin, at and before term. BJOG 2008;115:15127.
  • 24
    Schmitz T, Carnavalet Cde C, Azria E, Lopez E, Cabrol D, Goffinet F. Neonatal outcomes of twin pregnancy according to the planned mode of delivery. Obstet Gynecol 2008;111:695703.
  • 25
    Sentilhes L, Goffinet F, Talbot A, Diguet A, Verspyck E, Cabrol D, et al. Attempted vaginal versus planned cesarean delivery in 195 breech first twin pregnancies. Acta Obstet Gynecol Scand 2007;86:5560.
  • 26
    Sibony O, Touitou S, Luton D, Oury JF, Blot P. Modes of delivery of first and second twins as a function of their presentation. Study of 614 consecutive patients from 1992 to 2000. Eur J Obstet Gynecol Reprod Biol 2006;126:1805.
  • 27
    Smith GC, Fleming KM, White IR. Birth order of twins and risk of perinatal death related to delivery in England, Northern Ireland, and Wales, 1994–2003: retrospective cohort study. BMJ 2007;334:576.
  • 28
    Smith GC, Shah I, White IR, Pell JP, Dobbie R. Mode of delivery and the risk of delivery-related perinatal death among twins at term: a retrospective cohort study of 8073 births. BJOG 2005;112:113944.
  • 29
    Usta IM, Nassar AH, Awwad JT, Nakad TI, Khalil AM, Karam KS. Comparison of the perinatal morbidity and mortality of the presenting twin and its co-twin. J Perinatol 2002;22:3916.
  • 30
    Usta IM, Rechdan JB, Khalil AM, Nassar AH. Mode of delivery for vertex–nonvertex twin gestations. Int J Gynaecol Obstet 2005;88:914.
  • 31
    Yang Q, Wen SW, Chen Y, Krewski D, Fung Kee Fung K, Walker M. Occurrence and clinical predictors of operative delivery for the vertex second twin after normal vaginal delivery of the first twin. Am J Obstet Gynecol 2005;192:17884.
  • 32
    Yang Q, Wen SW, Chen Y, Krewski D, Fung Kee Fung K, Walker M. Neonatal death and morbidity in vertex–nonvertex second twins according to mode of delivery and birth weight. Am J Obstet Gynecol 2005;192:8407.
  • 33
    Joseph KS, Marcoux S, Ohlsson A, Liu S, Allen AC, Kramer MS, et al. Changes in stillbirth and infant mortality associated with increases in preterm birth among twins. Pediatrics 2001;108:105561.
  • 34
    Sheay W, Ananth CV, Kinzler WL. Perinatal mortality in first- and second-born twins in the United States. Obstet Gynecol 2004;103:6370.
  • 35
    Hirnle P, Franz HB, Sulkarnejewa E, Pfeiffer KH, Kiesel L. Caesarean section for the second twin after vaginal delivery of first. J Obstet Gynaecol 2000;20:3925.
  • 36
    Ananth CV, Demissie K, Hanley ML. Birth weight discordancy and adverse perinatal outcomes among twin gestations in the United States: the effect of placental abruption. Am J Obstet Gynecol 2003;188:95460.
  • 37
    Kontopoulos EV, Ananth CV, Smulian JC, Vintzileos AM. The influence of mode of delivery on twin neonatal mortality in the US: variance by birth weight discordance. Am J Obstet Gynecol 2005;192:2526.
  • 38
    Wen SW, Fung KF, Huang L, Demissie K, Joseph KS, Allen AC, et al. Fetal and neonatal mortality among twin gestations in a Canadian population: the effect of intrapair birthweight discordance. Am J Perinatol 2005;22:27986.

Journal club

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

Discussion points

1. Background: The authors discuss in the introduction the lack of evidence for benefit of planned caesarean birth for twins, despite the association of vaginal birth with increased morbidity and mortality for the second twin. Discuss the possible reasons.

Describe what the authors mean by ‘combined delivery’ of twins, and discuss the possible risks associated with this mode of delivery.

2. Methods: Define ‘systematic review’ and ‘meta-analysis’.

Critically appraise the methods in this study, with reference to published guidance for the conduct and reporting of systematic reviews (http://www.equator-network.org/resource-centre/library-of-health-research-reporting/reporting- guidelines/systematic-reviews-and-meta-analysis).

In addition to published papers, what other sources of information exist for studies to include in systematic reviews?

The authors excluded studies that reported data as graphs or percentages alone. Discuss possible methods for systematic reviewers to overcome problems associated with incomplete reporting of data in published studies.

Debate the inclusion or exclusion of unpublished data in systematic reviews.

3. Results and implications: This meta-analysis, based on observational studies, suggests worse outcome for twins B who were born by ‘combined delivery’ and for twins A who were born by caesarean section. Discuss possible reasons, and why a randomised controlled trial is needed to evaluate outcomes for twins.

The authors suggest that their study supports an attempt at vaginal birth for twins. Discuss which factors may affect your recommendation for mode of delivery when counselling women with twins, including factors not assessed in this systematic review.

4. Further reading: Leung TY, Lok IH, Tam WH, Leung TN, Lau TK. Deterioration in cord blood gas status during the second stage of labour is more rapid in the second twin than in the first twin. BJOG 2004;111:546–9.

Smith GC, Fleming KM, White IR. Birth order of twins and risk of perinatal death related to delivery in England, Northern Ireland, and Wales, 1994–2003: retrospective cohort study. BMJ 2007;334(7593):576.

Smith GC, Shah I, White IR, Pell JP, Dobbie R. Mode of delivery and the risk of delivery-related perinatal death among twins at term: a retrospective cohort study of 8073 births. BJOG 2005;112:1139–44.

D Siassakos University of Bristol & Southmead Hospital, Bristol, UK Email jsiasakos@gmail.com

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interest
  9. Contribution to authorship
  10. Funding
  11. References
  12. Journal club
  13. Supporting Information

Figure S1. Flow chart of the review process.

Figure S2. Morbidity of Twin A: vaginal delivery versus caesarean section.

Figure S3. Morbidity of Twin B: (A) combined delivery versus vaginal delivery; (B) combined delivery versus caesarean section.

FilenameFormatSizeDescription
BJO_2836_sm_figs1.xls16KSupporting info item
BJO_2836_sm_figs2.jpg32KSupporting info item
BJO_2836_sm_figs3a.jpg39KSupporting info item
BJO_2836_sm_figs3b.jpg38KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.