Article Text
PDF
Abstract
Objective To know the association of birth order with the risk of morbidity and mortality in very low-birthweight (VLBW) twin infants less than 32 weeks’ gestational age (GA).
Design Retrospective cohort study.
Setting Infants admitted to the collaborating centres of the Spanish SEN1500 neonatal network.
Patients Liveborn VLBW twin infants, with GA from 23+0 weeks to 31+6 weeks, without congenital anomalies, admitted from 2011 to 2020. Outborn patients were excluded.
Main outcome measures Respiratory distress syndrome (RDS), patent ductus arteriosus, bronchopulmonary dysplasia (BPD), necrotising enterocolitis, major brain damage (MBD), late-onset neonatal sepsis, severe retinopathy of prematurity, survival and survival without morbidity. Crude and adjusted incidence rate ratios were calculated.
Results Among 2111 twin pairs included, the second twin had higher risk (adjusted risk ratio (aRR) of RDS (aRR 1.08, 95% CI 1.03 to 1.12) and need for surfactant (aRR1.10, 95% CI 1.05 to 1.16). No other significant differences were found, neither in survival (aRR 1.01, 95% CI 0.99 to 1.03) nor in survival without BPD (aRR 1.02, 95% CI 0.99 to 1.05), survival without MBD (aRR 1.02, 95% CI 0.99 to 1.06) nor in survival without major morbidity (aRR 0.97, 95% CI 0.92 to 1.03). However, second twins born by caesarean section (C-section) after a vaginally delivered first twin had less overall survival and survival without MBD.
Conclusion In modern perinatology, second twins are still more unstable immediately after birth and require more resuscitation. After admission to the neonatal intensive care unit, they are at increased risk of RDS, but not other conditions, except for second twins delivered by C-section after a first twin delivered vaginally, who have decreased overall survival and survival without major brain injury.
- mortality
- neonatology
- twins
- intensive care units, neonatal
- epidemiology
Data availability statement
Data are available upon reasonable request. The data supporting the findings of this study (deidentified participant data) are available upon request to the corresponding author.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Classically, second twins have been considered to be at higher risk of perinatal morbidity and mortality. The improvement of obstetric and neonatal care could have changed this premise in the context of modern perinatology.
WHAT THIS STUDY ADDS
Advances in perinatal care have reduced the differences in morbidity and mortality related to twin order at birth seen in the past. Second twins are at increased risk of maladaptation immediately after birth, respiratory distress syndrome and need of supplemental oxygen and ventilatory support. After recovery of these acute events, survival and survival without major morbidity are similar between twins regardless of birth order, with the exception of second twins delivered by caesarean section after vaginal delivery of the first twin, who are at increased risk of major brain damage (MBD) and less likely to survive and survive without MBD.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This study, carried out with a large, almost population-based cohort, shows that perinatal management of twin pregnancies has currently significantly reduced the higher morbidity and mortality of second twins. However, special care should be taken in the case of second twins delivered by urgent caesarean section after a first twin delivered vaginally, due to the increased risk of brain injury and mortality.
Background
Every year, around 15 million babies are born prematurely worldwide, representing a global rate of 10.6% of all births.1 This group represents a small but highly relevant population in terms of public health and economic and social costs.2–4 The incidence of prematurity is three to five times higher in twins than singletons.5 With the increased maternal age at conception and the use of assisted reproductive technologies, nowadays, the incidence of multiple pregnancy in most Western countries has increased to 1.5%–3.4%.6 7 On the other hand, compared with singletons, infant mortality in twins has been estimated to be more than six times higher.8
Historically, the second twin has experienced a higher risk of perinatal morbidity and mortality.9 This disadvantage has been attributed to differences in sex, birth weight, mode of delivery, Apgar score or time interval between births.9–11 Although this observation has been acknowledged for years, improved obstetric and neonatal care, including intrapartum monitoring and operative interventions,12 antenatal corticosteroids administration13–15 and postnatal surfactant,16 has improved the outcomes of very low-birthweight (VLBW) infants, and the traditional observation of an increased risk of morbidity and mortality in the second twin may have changed in the context of modern perinatology.
The aim of our study was to determine the association of birth order with the risk of morbidity and mortality in VLBW preterm twins, less than 32 weeks’ gestational age (GA), in the context of modern perinatal care. We hypothesise that worse outcomes for VLBW second twins are no longer present in current practice.
Patients and methods
We conducted a retrospective analysis of prospectively collected data from a cohort including all pairs of VLBW twins delivered in the collaborating centres of the Spanish SEN1500 neonatal network over a 10-year period (January 2011–December 2020). Live, inborn patients with 23+0–31+6 weeks’ GA were included. Outborn patients were excluded to avoid selection bias and morbidity and mortality related to the transfer itself. Patients who died in the delivery room and those with severe congenital anomalies were excluded. Triplets or higher-order multiples were also excluded. Only pairs of patients in which both twins were admitted to the neonatal intensive care unit (NICU) were included in the study. Perinatal history, demographic data, diagnostic and therapeutic interventions, and outcomes were systematically collected. All variables were predefined in the SEN1500 network manual and were entered in the electronic system by the principal investigators of the collaborating centres.17 18 The database was fully accessible to the researchers.
The GA was estimated in completed weeks based on the date of the last menstrual period, obstetric parameters and/or a prenatal ultrasound registered in the maternal clinical record. Mortality was defined as death after admission and before hospital discharge. Major morbidity was considered the presence of major brain damage, including severe intraventricular haemorrhage (IVH) (grade 3 and/or periventricular haemorrhagic infarction) and/or echogenic or cystic periventricular leucomalacia19; moderate or severe bronchopulmonary dysplasia (BPD), defined as oxygen dependence or invasive or non-invasive respiratory support at 36 weeks’ postmenstrual age20; necrotisng enterocolitis ≥Bell’s stage 221; retinopathy of prematurity ≥stage 3 or need for surgical treatment,22 and/or late-onset neonatal sepsis (LONS), defined as suggestive clinical symptoms along with a positive blood culture after 72 hours of life.
Continuous variables are expressed as mean and SD or median and IQR, and differences between groups were examined with the paired Student t-test or the Wilcoxon test, as appropriate. Qualitative variables are expressed as n/N ratio (%) and compared using the McNemar’s test. Results are shown as within-pairs differences and 95% CI. All comparisons were two-tail tested, and a p value of <0.05 was considered significant.
In the causal inference study, to minimise potential biases, causal diagrams were analysed using directed acyclic graphs or causal Bayesian networks.23 Birth order was the exposure variable and the composite ‘morbidity and/or mortality’ was the outcome variable. Different variables were tested as potential confounders or intermediate variables, and only birth weight was found to be a true confounder (online supplemental figure 1).24 The crude and adjusted association of birth order with the outcomes of interest was studied by Poisson regression, with robust variance estimation. Results are expressed as risk ratios (RRs) and adjusted risk ratio (aRR) and 95% CI. Missing values for most variables of interest were less than 5% (online supplemental table 1), and a complete case analysis was carried out. For the sensitivity analysis, we carried out the determination of the E-values for the aRR and for limit of the CI closest to the null.25 The E-value is defined as the minimum strength of association on the RR scale that an unmeasured confounder would need to have with both the exposure and the outcome, conditional on the measured covariates, to fully explain away a specific exposure–outcome association.26 What constitutes a large E-value depends on the context but, in general, the higher the E-value, the less likely it is that there is an unknown, unmeasured factor that could override our best guess, given the data.27
Supplemental material
Supplemental material
SPSS V.25 software was used for statistical analysis. The study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) and Reporting of Studies Conducted using Observational Routinely-collected Data (RECORD) reporting guidelines (checklist as online supplemental appendix 1).
Supplemental material
Results
During the study period, 26 146 newborns were registered in the national SEN1500 network. After applying the inclusion and exclusion criteria, 4222 infants were included in the study (figure 1). Among the 166 patients who died in the delivery room, 45 were twins, 25 (55.6%) first twins and 20 (44.4%) second twins. The demographic characteristics and perinatal data of the population are shown in table 1. Most gestations were controlled, and mothers received at least one dose of antenatal steroids.
Demographic characteristics and perinatal aspects of the study population (N=2111 twin gestations)
Flowchart with the inclusion and exclusion criteria of the study population. BW, birth weight; DR, delivery room; GA, gestational age; NICU, neonatal intensive care unit; VLBW, very low birth weight.
Table 2 shows the infants’ characteristics and interventions according to birth order. Second twins had a lower birth weight and were more frequently small for gestational age (SGA). No significant differences were found for other anthropometric data. First twins exhibited more frequently a premature rupture of membranes and a longer time since rupture of membranes to birth. The proportion of caesarean section (C-section) was higher among second twins. They showed lower Apgar scores and needed more oxygen and resuscitation manoeuvres.
Demographic characteristics and perinatal interventions according to birth order
Table 3 shows the morbidity exhibited and interventions needed by the patients according to birth order. Data are shown in proportions, as well as crude RR and aRR with 95% CI. Second twins developed more frequently respiratory distress syndrome (RDS) (E-value for the point estimate and for the limit of the CI closest to the null 1.37 and 1.21, respectively) and required more frequently surfactant administration after adjusting for birth weight (E-value for the point estimate and for the limit of the CI closest to the null 1.51 and 1.31, respectively). Second twins also needed more oxygen and invasive mechanical ventilation after admission to NICU and pharmacological closure of a patent ductus arteriosus (PDA) and exhibited more frequently LONS, but all these differences disappeared after adjusting for birth weight. There were no other differences in morbidity or mortality (tables 3 and 4). The total length of stay was similar between first and second twins: mean 58.9 (SD 37.5) vs mean 58.3 (SD 36.0) days, respectively. Paired mean difference was 0.65 (95% CI −1.13 to 2.42) days (p=0.475).
Morbidity and interventions by twin order
Survival and survival without morbidity according to the order of the twins
The time interval between the first and second births was analysed in a total of 2054/2111 (97.3%) twin pairs. The time interval was longer for vaginal deliveries of the first twin followed by C-section of the second, with a median of 14 (IQR 4–26) min, and shortest when both twins were delivered by C-section, with a median of 2 (IQR 1–4) min. When both twins were vaginally delivered, the time interval was 6 (IQR 3–12) min (p<0.001). Overall, we did not find significant differences regarding morbidity or mortality among second twins according to the difference in time to the birth with respect to the first twin. However, second twins born by C-section after a vaginal delivery of the first twin tended to have worse results, although they were only statistically significant for major brain damage, survival and survival without major brain damage (online supplemental table 2).
Supplemental material
Discussion
Our study confirms previous observations that second twins are at higher risk of postnatal immediate maladaptation and RDS, needing more intervention during resuscitation, oxygen supplements and surfactant administration. However, after recovering from these initial acute morbidities, they evolved favourably with similar rates of survival and survival without morbidity.
RDS is a consequence of lung immaturity and surfactant deficiency in preterm newborns, with higher incidence at lower GA.28 In our study, most potential sources of variability, such as maternal conditions, antenatal steroid exposure and GA, were controlled by matched twins’ pairs. For this reason, a priori, a similar incidence of RDS in each pair of patients should be expected. However, we found a higher rate of acute respiratory morbidity that remained statistically significant after adjusting for birth weight. This acute morbidity has been attributed, at least in part, to the difference in time at birth, with the second twin being exposed to the stress of birth for a longer period. However, in our population, we could not demonstrate any correlation between the difference in time at birth and any morbidity or mortality. Nevertheless, we did observe a higher rate of RDS when both twins were born by C-section, which is a known independent risk factor for RDS.29 It is noteworthy that, after the initial adaptation and support, we did not find long-term differences in respiratory or other clinical outcomes. This is in contrast with the study of Shinwell et al,11 who, in a similar cohort of VLBW infants, found an increased risk of RDS in the second twin, followed by an increased risk of chronic lung disease (CLD) and mortality. However, in that study, CLD was defined as oxygen dependency at 28 days. Other studies, like ours, found an increased risk of RDS in the second twin, but not CLD, defined as oxygen requirement at 36 weeks’ PMA.30 31 In Donovan et al’s study,31 second twins also developed more PDA and required more frequently pharmacological or surgical closure. Recently, we have shown that during the last decade in Spain, in spite of an increased use of non-invasive ventilatory support, overall, neither survival nor survival without BPD has improved in very preterm infants.32
We also found a higher frequency of LONS in the second twin, although statistical significance disappeared after adjusting for birth weight. Total hospital stay was similar in both twins, and the proportion of infants undergoing invasive mechanical ventilation or central parenteral nutrition at 28 days was also similar. The association between birth route and LONS is not consistent in the literature. In our population, second twins were born more frequently by C-section. In a post hoc analysis, we studied the influence of the type of delivery on nosocomial sepsis, separating coagulase-negative Staphylococcus (CONS) infections from those caused by other germs. We found no differences for other bacteria in vaginal versus caesarean deliveries: 16.5% vs 16.1% (p=0.932), respectively. However, the incidence of CONS infections was significantly higher in patients born by C-section: 17.6% vs 22.4% (p=0.01). Similar findings were reported by Olivier et al.33 Alterations in the microbiota and their consequences on a developing immune system have been postulated as possible causes of changes in the colonisation of the newborn intestine and various associated complications.34
The largest difference in time at birth occurred when the first twin was born vaginally and the second by C-section. In these cases, the second twin exhibited a higher incidence of major brain damage, a lower survival and survival without major brain damage. Although the pathophysiology of IVH and PVL is multifactorial,19 fluctuations of cerebral blood flow due to impaired autoregulation and ischaemia/reperfusion abnormalities are among the underlying causes.19 We speculate that second twins, who remained precariously in utero during labour after the first twin was born, might suffer from haemodynamic instability and potentially birth asphyxia. In a previous study several decades ago, Müller-Holve et al found that when the time interval exceeded 7 and 10 min, the second twin exhibited worse pH values in the umbilical artery blood, without worse clinical outcomes.35 However, in this study, the included patients were full-term newborns with >2000 g, whose maturity could favour the adaptation and correction of metabolic alterations after birth, unlike our population of more immature patients.
Our study has the inherent limitations of retrospective cohort studies, in which certain interventions that are now known to influence outcomes, such as delayed cord clamping or early versus late use of surfactant, were not systematically collected in the database and, therefore, are not available for analysis. We also lacked information on zygosity and chorionicity. In addition, patients who died in the delivery room were not included in the study, which could modify the overall results in terms of morbidity and mortality. However, on the other hand, this prevents the bias of possible decisions made before delivery or by different hospital policies regarding certain groups of patients, such as patients at the limit of viability. Finally, the exclusion of patients where only one of the twins was admitted could introduce a bias in the analysis. However, this could also be a strength of the study as the excluded twins could have weighed more than 1500 g or could have died in the delivery room and only the healthier one could have been admitted.
Our study has also other strengths such as the large number of patients from a large nationally representative database. This large contemporary cohort makes our results potentially generalisable to countries with similar socioeconomic development. The focus on VLBW twins increased the probability of finding real differences between the cohorts, since the prenatal and maternal histories are the same, avoiding possible confounding factors related to them.
In conclusion, in modern perinatology, advances in perinatal care have reduced the differences in morbidity and mortality related to twin order at birth seen in the past. Second twins are at increased risk of maladaptation immediately after birth and develop more frequently RDS, requiring more supplemental oxygen and non-invasive and invasive ventilatory support in the delivery room. After recovery of these acute events, no differences in CLD or in the risk of mortality are observed. Survival and survival without major morbidity are currently similar between twins regardless of birth order, with the exception of second twins delivered by C-section after vaginal delivery of the first twins, who are at increased risk of major brain damage and less likely to survive and survive without major brain damage.
Supplemental material
Data availability statement
Data are available upon reasonable request. The data supporting the findings of this study (deidentified participant data) are available upon request to the corresponding author.
Ethics statements
Patient consent for publication
Ethics approval
The research ethics committees of all the participating centres approved the data collection protocol when they joined the network. Permission for data analysis was obtained from the executive committee of the Spanish Society of Neonatology (SEN1500). The research ethics committee of the principal investigator's centre acted as a reference committee (code 2021-261-1). Given the retrospective design and anonymous nature of the data, it was not deemed necessary to obtain informed consent from the parents or legal representatives of the patients.
Acknowledgments
We are indebted to all the investigators, coordinators and members of the Spanish SEN1500 Neonatal Networks (see online supplemental appendix 2).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @AlexAvila81
Collaborators Spanish SEN1500 Network: H. GERMANS TRIAS I PUJOL (Gemma Ginovart); H. CLÍNIC DE BARCELONA (Josep Figueras Aloy); H. U. BASURTO (Alberto Pérez Legorburu); H.G. DE CASTELLÓN (Ramon Aguilera Olmos); H.U. DE SAN CECILIO (J Uberos); H. DE LEON (Sandra Terroba); H. CARLOS HAYA (Tomás Sanchez-Tamayo); H.I. LA PAZ (Mª Dolores Elorza Fernandez); H. CLÍNICO SAN CARLOS (Araceli Corredera Sánchez); H. CENTRAL DE ASTURIAS (Belén Fernández Colomer); COMPLEXO HOSPITALARIO PONTEVEDRA (Mª Angeles Martinez Fernandez); H. UNIVERSITARIO MARQUES DE VALDECILLA (Mª Isabel de las Cuevas Terán); H. DONOSTIA (Miguel Ángel Cortajarena Altuna); H.U.I. VIRGEN DEL ROCÍO (Carmen Macias Díaz); H.U. DE CANARIAS (Pedro Fuster Jorge); H. MIGUEL SERVET (Segundo Rite Gracia); H.C.U. LOZANO BLESA (Mª Purificación Ventura Faci); H.U. LA FE (Mª Isabel Izquierdo Macián); H. VIRGEN DE LA SALUD (Ana Belén Escobar Izquierdo); H.C.U. DE SANTIAGO (Mª Luz Couce Pico); H.U. SALAMANCA (Elena Pilar Gutierrez González); H.G.U. GREGORIO MARAÑÓN (D. Blanco Bravo); H. SAN PEDRO DE LOGROÑO (Mª Yolanda Ruiz del Prado); H.U.M.I. LAS PALMAS (Lourdes Urquía Martí); H. DE CABUEÑES (Rafael Garcia Mozo); H.U. REINA SOFÍA (Mª Pilar Jaraba Caballero); H.U. DE BURGOS (Cristina de Frutos Martínez); SCIAS-H. BARCELONA (Sílvia Martínez-Nadal); H. SAN JOAN DE DEU (Martin Iriondo); H. DE CRUCES (Amaya Rodriguez Serna); H.G.U. DE ALICANTE (María Gonzalez Santacruz); H. VIRGEN DE LAS NIEVES (Maria Fernanda Moreno Galdo); CORPORACIO PARC TAULÍ (Joan Badia Barnusell); H.U. RIO HORTEGA (Mª Mar Montejo Vicente); H. TXAGORRITXU (Aintzane Euba); H. JUAN XXIII (Mar Albújar); H. DE GETAFE (Irene Cuadrado Perez); INSTITUT DEXEUS; COMPLEJO HOSPITALARIO ALBACETE (Andres Martinez Gutierrez); H. DE LA SANTA CREU I SANT PAU (Gemma Ginovart Galiana); H. SAN PEDRO DE ALCÁNTARA CÁCERES (Mª Jesús López Cuesta); H. SEVERO OCHOA (Mª José Santos Muñoz); ESTRUCTURA ORGANIZADA DE XESTIÓN INTEGRADA DE VIGO. HOSPITAL ALVARO CUNQUEIRO (María Suárez Albo); H. VIRGEN DE LA CONCHA - COMPLEJO ASISTENCIAL DE ZAMORA (Víctor Manuel Marugán Isabel); H. DE JEREZ (María Victoria Ramos Ramos); H. MONTEPRINCIPE (Gerardo Romera Modamio); H. DE ELCHE (Carolina Vizcaíno); H. JUAN RAMÓN JIMÉNEZ (David Mora Navarro); H. VALME DE SEVILLA (Laura Acosta Gordillo); H.U. ARNAU DE VILANOVA DE LLEIDA (Eduard Soler Mir); H. VIRGEN DE LA MACARENA DE SEVILLA (Mercedes Granero Asensio); H.C.U. DE VALENCIA (Dr. Javier Estañ Capell); H.G.U. DE CIUDAD REAL (Miguel Angel García Cabezas); H. DE LA ZARZUELA (D. López Gómez); H. U. DE GIRONA DR. JOSEP TRUETA (Alberto Trujillo Fagundo); H. DE GRANOLLERS (Israel Anquela Sanz); H. PUERTA DEL MAR (Antonio Segado Arenas); H. 12 DE OCTUBRE (Carmen Rosa Pallás Alonso); H.U. NUESTRA SEÑORA DE CANDELARIA (Sabina Romero); H. PUERTA DE HIERRO (Carmen González Armengod); H.U. SANTA LUCIA DE CARTAGENA (Jose María Lloreda García); H. FUENLABRADA (Laura Domingo Comeche); H.G. DE CATALUÑA (Laura Castells Vilella); H. VIRGEN DEL CAMINO DE PAMPLONA (Concepción Goñi Orayen); H.U. QUIRÓN; CLÍNICA CORACHÁN (Mª Dolores Muro Sebastian); COMPLEXO HOSPITALARIO UNIVERSITARIO DE A CORUÑA (Alejandro Avila-Alvarez); HOSPITAL HM PUERTA DEL SUR (Mara Fernandez Diaz); HOSPITAL QUIRÓN SAGRADO CORAZÓN (Elena García Victori); HOSPITAL UNIVERSITARIO FUNDACIÓN JIMENEZ DÍAZ (T. Carrizosa).
Contributors FG-MR, LUM and DRS conceptualised the study and carried out the initial design. MJGD and ILDPH curated the database, conducted the literature review and wrote the first draft. FG-MR carried out the statistical analysis. AA-A contributed to the literature review and suggested improvements to the manuscript and tables. All authors reviewed the manuscript, suggested improvement and approved the final version. FG-MR is responsible author for the overall content, had full access to the data, and accept the resposibility of its publication.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.