Neonatal complications in late‐preterm and early‐term infants: a systematic review of observational studies

Objectives

This is a protocol for a Cochrane review (flexible). The objectives are as follows:

To assess neonatal complications, specifically neonatal mortality and morbidities, in late‐preterm and early‐term infants compared to full‐term newborns.

Background

Description of the condition

Preterm birth, defined as birth before 37 completed weeks of gestation, constitutes a significant public health challenge, with an estimated annual rate of 13.4 million newborns worldwide [1]. Prematurity remains the leading cause of death and complications in the neonatal period and for children under five years of age [2, 3]. However, there is considerable variation across the gestational age (GA) spectrum, with the rates of complications increasing exponentially as GA decreases [4, 5]. Consequently, until recently, healthcare providers have primarily focused on newborns delivered before 34 weeks' gestation, whereas infants born later were often perceived as clinically mature and less vulnerable to adverse outcomes. However, emerging evidence has contradicted this assumption and identified late‐preterm delivery (between 34^0/7 and 36^6/7 GA) (numbers in superscript refer to days) as a significant risk factor for neonatal morbidities [6].

Recent research shows that even term newborns born between 37^0/7 and 38^6/7 weeks of gestation, defined as early‐term infants, are more likely to experience adverse neonatal outcomes compared to full‐term births (39^0/7 to 40^6/7 GA) [7]. This strengthens the concept that gestation is a biological continuum, and the inverse dose‐response relationship between GA and neonatal morbidity extends well beyond the 37‐week hallmark, which traditionally denoted infant maturity [8, 9].

Compared to infants born before 34 weeks of GA, the majority of late‐preterm and early‐term newborns will survive and not require intensive care, given they are provided with appropriate postnatal support. However, they are still at a higher risk of neonatal and infant morbidity and mortality than infants born after 39 weeks of GA, contributing substantially to the overall morbidity burden [6, 8, 9, 10]. Additionally, the magnitude of late‐preterm and early‐term deliveries exacerbates the public health burden these newborns exert. More than one‐third of all births occur in the late‐preterm and early‐term periods, and approximately 75% of all preterm newborns are late preterm [11]. In Western countries, late‐preterm and early‐term birth rates range from 3% to 6% and from 15% to 30%, respectively [12, 13]. Over the past decade, these rates have increased by around 10% [1, 11].

Almost two‐thirds of preterm and early‐term deliveries occur spontaneously, while the remaining one‐third are due to obstetric interventions (induced labor or elective cesarean section) [14]. A myriad of maternal, fetal, obstetric, and sociodemographic risk factors associated with late‐preterm and early‐term birth have been identified. Advanced maternal age, maternal medical conditions (diabetes mellitus, hypertension, eclampsia/preeclampsia, placental pathologies, poly‐/oligohydramnios), history of preterm birth, and assisted pregnancy increase the risk of late‐preterm and early‐term delivery [15]. Other factors, such as low maternal education, infection and inflammation, and fetal chromosomal or congenital abnormalities, might be related only to late‐preterm birth [15, 16]. On the other hand, a history of one or more previous childbirths and newborn male sex have been linked to early‐term birth [16]. Understanding these risk factors is crucial for planning mitigation strategies to rein in the rising birth rates of these newborns and to potentially alleviate the overall morbidity burden.

The short‐term outcomes of late‐preterm infants have been extensively described in the literature, while studies focusing on early‐term births are relatively fewer in number [12, 17]. It has been shown that infant mortality is about four times higher for late‐preterm births and about 50% higher for early‐term births compared to full‐term births [18]. An extensive population‐based study showed that infants born late preterm are significantly more likely to require resuscitation at delivery, admission to a neonatal intensive care unit (NICU), and respiratory support than those born at ≥ 37 weeks' gestation [19]. Neonatal morbidity rates are also higher for late‐preterm newborns than for term births due to increased risks of respiratory distress syndrome (RDS), sleep apnea, necrotizing enterocolitis, and intraventricular hemorrhage [2, 20, 21]. Likewise, early‐term infants are at a higher risk of feeding difficulties, hypoglycemia, jaundice, and infections compared to full‐term newborns; however, the risk for serious morbidities, though relatively small, is still present and cannot be ignored [15, 22, 23].

Late‐preterm delivery disrupts the physiological development of the fetal respiratory system. Structurally and functionally immature lungs increase the risk of various pulmonary diseases [24, 25]. Irrespective of lung maturity status, the relative overall immaturity of early‐term infants predisposes them to various respiratory problems [26]. Compared to their full‐term counterparts, late‐preterm and early‐term newborns are at an increased risk of respiratory morbidities, including RDS, transient tachypnea of the newborn (TTN), and respiratory failure [27, 28, 29]. The risk of RDS, the most frequent respiratory morbidity among late‐preterm and early‐term neonates, decreases gradually from 34 to 38 weeks compared to newborns delivered at 39 to 40 weeks [27]. A similar pattern is observed for TTN, pneumonia, respiratory failure, and the need for surfactant therapy and respiratory support, substantiating a dose‐response relationship between GA and respiratory complications [27]. Due to respiratory morbidity, late‐preterm and, to a lesser extent, early‐term newborns more frequently require non‐invasive and invasive ventilation than full‐term neonates [30].

Late‐preterm and early‐term infants are also more likely to present with hyperbilirubinemia and more frequently require phototherapy than full‐term newborns [31, 32, 33, 34]. Infants born late preterm tend to have a longer course of hyperbilirubinemia and are more likely to develop kernicterus [35]. Other metabolic issues frequently encountered in late‐preterm newborns include hypoglycemia and hypothermia [36, 37]. These complications are due to a deficiency of glycogen stores, subcutaneous fat, and brown adipose tissue [38]. Early‐term infants are also more prone to hypoglycemia, especially in the first 48 hours of life [32, 34], reflecting the relative immaturity of glucose homeostasis even at 37 and 38 weeks of GA [34]. Additionally, both late‐preterm and early‐term infants exhibit feeding difficulties [33, 37]. Compared to full‐term infants, they are more likely to experience problems with breastfeeding initiation, duration, and exclusivity and to be discharged late from the hospital because of poor feeding and insufficient weight gain [34, 39].

Moreover, in later life, children born late preterm are more likely to suffer from long‐term motor and cognitive impairments, chronic diseases, and premature death [40, 41, 42, 43]. They are more likely to develop cerebral palsy and have poorer neurodevelopmental and academic performance outcomes [44, 45]. Early‐term birth is also associated with poorer school performance and an increased need for hospital care during childhood, mainly due to obstructive airway diseases and ophthalmological and motor problems [44, 46]. A recent Japanese study highlighted that at three years of age, infants born late preterm and early term are at a higher risk of growth failure and respiratory symptoms than those delivered full term [47].

Beyond their clinical consequences, neonatal complications associated with late‐preterm and early‐term birth may also impose considerable emotional and practical burdens on families. Feeding difficulties, NICU admission, and prolonged hospitalization can increase caregiver stress, disrupt early parent‐infant bonding, and require additional healthcare support during the neonatal period [48].

Several maternal and perinatal factors are known to influence neonatal outcomes. Cesarean section, compared with vaginal birth, has been associated with a higher risk of impaired neonatal respiratory adaptation, altered microbiome development, delayed initiation of feeding, and increased early morbidity [49, 50]. Multiple pregnancies also carry increased risks of prematurity and neonatal complications, which makes them clinically distinct from singleton pregnancies [51]. Maternal morbidities, including diabetes and hypertensive disorders of pregnancy, are important contributors to altered fetal growth, metabolic dysregulation, and increased neonatal morbidity [52, 53]. In contrast, administration of antenatal corticosteroids is a well‐established intervention that improves neonatal respiratory stability and metabolic adaptation [54]. Given their potential to modify neonatal outcomes, these factors should be considered important confounders and included in special analyses to allow more accurate interpretation of the associations between risk factors and neonatal outcomes.

Why is it important to do this review

Late‐preterm and early‐term infants are the fastest‐growing subgroups of neonates [2, 55]. Although visually they can easily be conflated with full‐term infants, they are at a significantly higher risk of mortality, a wide range of neonatal morbidities, and adverse long‐term complications extending well into adulthood [2, 42]. Unfortunately, despite the available evidence, many healthcare professionals continue to overlook the risks associated with late‐preterm and early‐term deliveries, prioritizing preterm newborns delivered at less than 34 weeks' GA. However, considering that the number of infants delivered between 34 and 38 weeks is significantly higher than those delivered at less than 34 weeks, and that these rates are increasing continuously, the late‐preterm and early‐term populations represent an important and growing public health challenge and therefore deserve special attention. This warrants a systematic review of the literature to underscore the gravity of the situation.

Given the heterogeneity of reported findings and a lack of overall consensus on the public health impact of late‐preterm and early‐term deliveries, it is essential to delineate systematically the neonatal morbidities that are most common and concerning in this population. This could promote broader recognition of the risks associated with these deliveries and inform changes in clinical practice. The findings of this systematic review could equip healthcare professionals with the knowledge needed to anticipate risks associated with late‐preterm and early‐term deliveries and implement timely interventions tailored to these newborns, potentially improving the standard of healthcare delivery for these infants. Moreover, it can also trigger the generation of new obstetric guidelines in favor of delaying childbirth until 39 weeks unless the risks of continuing the pregnancy outweigh the risks of an earlier delivery.

Randomized controlled trials are not feasible or ethical for investigating the effects of GA at birth, as GA and timing of delivery cannot be randomly assigned for research purposes. Consequently, evidence on the association between late‐preterm or early‐term birth and neonatal outcomes is necessarily derived from observational studies. A systematic synthesis of this evidence is therefore required to provide a comprehensive assessment of the risks associated with these births and to inform clinical practice and policy.

On a global level, this systematic review could provide a framework for future medical care and service planning for these infants, such as training community health workers in kangaroo mother care (KMC), essential newborn care, and special care for late‐preterm newborns [56]. Additionally, it could support the incorporation of long‐term evaluation, monitoring, and follow‐up of these infants into routine practice, optimizing the medical care offered to these newborns [55]. The findings of this review may also support parental counseling, informed decision‐making, and evidence‐based clinical care for late‐preterm and early‐term infants and their families.

Objectives

To assess neonatal complications, specifically neonatal mortality and morbidities, in late‐preterm and early‐term infants compared to full‐term newborns.

Methods

Criteria for considering studies for this review

Types of studies

We will include non‐randomized prospective or retrospective studies, including cohort, case‐control, cross‐sectional or registry‐based studies, that compare outcomes between late‐preterm and/or early‐term infants (exposure groups) and full‐term infants (comparison group). We will consider both population‐based and facility‐based studies, conducted in single or multiple centers. We will apply no restrictions based on sample size.

Population and exposure

We will include late‐preterm and early‐term infants, which constitute the exposure groups of interest in this PECO‐based review (Population, Exposure, Comparator, Outcome). In this context, the exposure is GA at birth rather than a traditional environmental factor. Specifically, the exposure groups are as follows.

• Late‐preterm infants born between 34^0/7 and 36^6/7 weeks of gestation

• Early‐term infants born between 37^0/7 and 38^6/7 weeks of gestation

We will compare late‐preterm and early‐term infants with full‐term infants (born between 39^0/7 and 40^6/7 weeks of gestation) [7].

We will exclude preterm infants born before 34 weeks of GA and late‐term and post‐term infants (born ≥ 41 weeks of gestation). We will place no restrictions on birth weight and postnatal age.

We will exclude studies involving infants with congenital anomalies because these conditions may independently affect neonatal outcomes.

We will include studies that report outcomes for late‐preterm or early‐term infants, or both, according to our prespecified GA categories. We will consider studies using different GA definitions if data can be extracted for categories that closely correspond to our definitions; any deviations will be documented. Where substantial variation in GA definitions is identified, we will explore its impact through sensitivity analyses, where feasible, or address it in the narrative synthesis.

Studies including broader GA categories will be eligible if data for the relevant GA groups can be extracted separately. Where subgroup‐specific data are not reported, we will attempt to contact study authors to obtain the relevant data. If the eligible populations cannot be adequately separated, we will exclude the study from meta‐analysis and report its findings narratively.

Comparison

We will compare mortality and short‐term morbidities of late‐preterm and early‐term infants with those of full‐term infants. We will analyze late‐preterm and early‐term cohorts separately in comparison with term infants. We define short‐term morbidity as that experienced at birth and within the neonatal period, up to the first 28 days of life.

The comparator group in this review comprises full‐term infants, defined as those born between 39^0/7 and 40^6/7 weeks’ gestation. Studies that also include late‐term infants (> 40^6/7 weeks' gestation) will be eligible if outcome data for full‐term infants can be extracted separately. Where subgroup‐specific data are unavailable, we will attempt to contact study authors to obtain the relevant data. If the comparator population cannot be adequately separated, we will exclude the study from meta‐analysis and report its findings narratively.

Outcomes

Outcome measures are detailed below. We will include neonatal outcomes (i.e. occurring within the first 28 days of life or until discharge from birth hospitalization, if it happens later). We will include studies if they measure at least one of the prespecified outcomes of interest, regardless of whether data are reported; we will exclude studies that do not measure any outcomes of interest.

Critical outcomes

• Neonatal death.

• Resuscitation in the delivery room, defined as the need for the use of positive airway pressure or chest compressions or both, immediately after birth.

• Use ofsurfactant treatment within the first 48 hours of life.

• Need for mechanical ventilation, defined as ventilation through an endotracheal tube, during the hospital stay after birth.

• Need for non‐invasive respiratory support, defined as respiratory support without the need for an endotracheal tube (including but not limited to nasal continuous positive airway pressure [nCPAP], nasal intermittent positive pressure ventilation [NIPPV], high flow nasal cannula [HFNC]), during the hospital stay after birth.

Important outcomes

• Duration of hospital stay in days.

• Admission to the NICU.

• Early‐onset sepsis, defined as a positive blood culture with signs of infection, diagnosed within the first 72 hours of life.

• Late‐onset sepsis, defined as a positive blood culture with symptoms of infection, diagnosed in a newborn after 72 hours of life.

• Necrotizing enterocolitis (NEC), defined as inflammation of the gastrointestinal tract in newborns, diagnosed based on Bell criteria, stage ≥ II [57].

• Hypoglycemia requiring treatment (intravenous infusion, sucrose gel).

• Hyperbilirubinemia receiving treatment, according to the American Academy of Pediatrics Guidelines [58, 59].

• Number of days until independent feeding (breast or bottle).

• Rate of exclusive breastfeeding (only feeding mother's breast milk).

Search methods for identification of studies

Electronic searches

A draft search strategy was written by an Information Specialist (MBj) and is provided in Supplementary material 1. We will conduct searches without date, publication type, or language limits. Search strategies will be peer‐reviewed by an Information Specialist based on the Peer Review of Electronic Search Strategies checklist [60, 61]. We will search the following databases:

• PubMed, National Library of Medicine, 1946 forward;

• Embase, Elsevier via Embase.com, 1947 forward;

• CINAHL (Cumulative Index to Nursing and Allied Health Literature), EBSCOhost, 1982 forward;

• Global Health via CABI.

Searching other resources

We will search the following trial registries:

• National Library of Medicine trial registry ClinicalTrials.gov (clinicaltrials.gov/);

• World Health Organization International Clinical Trials Registry Platform (ICTRP) (trialsearch.who.int/Default.aspx).

We will search for conference abstracts published during the past five years, as available, for:

• Perinatal Society of Australia and New Zealand (PSANZ);

• Pediatric Academic Societies (PAS);

• European Academy of Paediatric Societies (EAPS).

We will search for errata or retractions for studies selected for inclusion via PubMed and Retraction Watch. We will check the reference lists of systematic reviews, papers related to the topic of this review, and studies selected for inclusion to identify studies not found through our searches.

Data collection and analysis

Selection of studies

We will manage search results in EndNote [62] and screen results in Covidence [63].

Two review authors (of KWS, MA, MG, FB, or GS) will independently screen the titles/abstracts and the full texts retained following title/abstract review. At any point in the screening process, we will resolve disagreements by discussion. We will document the reasons for excluding studies during full‐text review in a 'Characteristics of excluded studies' table. We will exclude studies if the population or study design do not meet our eligibility criteria. We will collate multiple reports of the same study so that each study, rather than each report, is the unit of interest in the review. We will record the selection process in sufficient detail to complete a PRISMA flow diagram [64].

Data extraction and management

Two review authors (of KWS, MA, MG, FB, or GS) will independently extract data in Covidence [63]; the characteristics will be based on the Cochrane Effective Practice and Organisation of Care Group (EPOC) data collection checklist [65]. We will pilot the form within the review team, using a sample of three included studies. We will extract the following characteristics from each included study.

• Administrative details: study author(s), published or unpublished, year of publication, year in which the study was conducted, study authors’ conflicts of interest and study funding sources.

• Methodological characteristics of each study: study design, population‐based or facility‐based study, study setting, number of study centers and location, ethical approval, completeness of follow‐up (e.g. greater than 80%).

• Participants: baseline characteristics of study participants, number, the number lost to follow‐up/withdrawn, number analyzed, mean GA, GA range, method used to assess GA (ultrasound versus last menstrual period), sex, single versus multiple births, inclusion criteria, exclusion criteria, maternal and antenatal characteristics (antibiotic exposure during pregnancy, antenatal corticosteroid use, gestational diabetes, hypertensive disorders, maternal infection, hospitalization prior to delivery, smoking, alcohol, or substance use).

• Outcomes: as outlined above under 'Outcomes'; the number of participants with outcomes.

• Estimates of effect or association: names of the metrics used to measure effect or association (e.g. risk ratio, odds ratio, mean difference, correlation); adjusted or unadjusted point estimates; and confidence intervals, and their significance levels or standard errors, or P values for all risk factors meeting our inclusion criteria; and the direction of effect or association (e.g. whether presence or absence of a risk factor is associated with increased or decreased outcome risk). This list specifies the full set of statistical effect measures that will be extracted to ensure transparent and comparable reporting of the magnitude, direction, and uncertainty of associations across all included studies.

• Equity characteristics as outlined in the 'Equity‐related assessment' section.

We will extract the characteristics of study participants and outcomes separately for each study group (late‐preterm, early‐term, and full‐term infants [cohort studies] or cases and controls [case‐control studies]). We will resolve any disagreements by discussion.

We will describe any ongoing studies and document the available information, such as the primary author, research question(s), methods, outcome measures, and an estimated reporting date, in the 'Characteristics of the ongoing studies' table.

Should any queries arise, or in cases in which additional data are required, we will contact study investigators/authors for clarification. Two review authors will use RevMan software for data entry [66].

Risk of bias assessment in included studies

We will use the standard methods of Cochrane and Cochrane Neonatal to assess the methodological quality of the studies. For each outcome, two review authors (of KWS, MA, MG, FB, GS, or MBr) will independently assess the risk of bias. We will resolve disagreements with a third review author (KWS or MBr). We will perform the assessments using the Risk Of Bias In Non‐randomized Studies – of Exposure (ROBINS‐E) tool [67]. We will seek information regarding the method of reporting outcomes for all the infants enrolled in the study. We will assess each criterion as having a low, some concerns, high, or very high risk of bias [67]. We will add this information to the 'Characteristics of included studies' table. We will evaluate the following issues and enter the findings into the risk of bias table.

• Bias due to confounding

• Bias arising from the measurement of the exposure

• Bias in the selection of participants for the study

• Bias due to post‐exposure interventions

• Bias due to missing data

• Bias in the measurement of the outcome

• Bias in the selection of the reported result

The confounding factors taken into consideration will be maternal morbidities (maternal diabetes, hypertensive disorders, eclampsia/preeclampsia), small for gestational age (SGA), mode of delivery, and multiple gestation.

We will assess overall risk of bias for each outcome. The overall risk of bias will be set as high as the highest‐rated domain. Studies judged to have a very high risk of bias will not undergo a detailed domain‑level assessment and will not be included in any quantitative synthesis, as their results will be considered too unreliable for pooling. They will, however, be retained in the review and described narratively.

A more detailed description of the risk of bias for each domain is provided in Supplementary material 2.

We will conduct a risk of bias assessment for all critical outcomes (i.e. neonatal death, resuscitation in the delivery room, use of surfactant treatment, need for mechanical ventilation, need for non‐invasive respiratory support) and for two important outcomes (duration of hospital stay and admission to the NICU). These outcomes will be presented in the summary of findings tables [68].

Measures of treatment effect

Dichotomous data

For dichotomous outcomes, we will extract risk ratios (RR), odds ratios (OR), or hazard ratios (HR) depending on how the data are reported in the included studies. We may calculate risk difference (RD) for sensitivity or supplementary analyses if they provide additional insight (e.g. when event rates are very low or when absolute effects are needed). We will report all effect estimates—both extracted and calculated—with corresponding 95% confidence intervals (CIs).

Continuous data

For continuous outcomes, we will extract and report mean differences (MDs) or standardized mean differences (SMDs), depending on the scales used across studies. Where trials report continuous data as median and interquartile range (IQR), and data pass the test of skewness, we will convert the median to mean and estimate the standard deviation (SD) as IQR/1.35 [69].

As recommended for non‑randomized studies, we will prioritize adjusted effect estimates over unadjusted ones [70]. The key confounders we expect primary study authors to adjust for are maternal morbidities (maternal diabetes, hypertensive disorders, eclampsia/preeclampsia), SGA, mode of delivery, and multiple gestation.

Unit of analysis issues

The unit of analysis for all outcomes will be the infant. We will consider an infant only once in the analysis.

For studies that include more than two GA groups (e.g. late preterm, early term, full term), we will extract data for each relevant group separately. When multiple eligible comparator groups are included in a single study, we will combine groups that are clinically similar, or, if groups must remain separate, we will split the shared comparator group evenly to prevent double‑counting in meta‑analysis. If appropriate adjustments cannot be made, we will include the study in narrative synthesis only.

Dealing with missing data

If we identify important missing or unclear data, we will request the needed information through contact with the original investigators. We will make explicit the assumptions of any methods used to deal with missing data. We will address the potential impact of missing data on the findings of the review in the Discussion section of the review.

Reporting bias assessment

We will assess reporting bias by comparing the stated neonatal outcomes with the reported outcomes. Where study protocols are available, we will compare these to the full publications to determine the likelihood of reporting bias.

We will use funnel plots to screen for publication bias where there is a sufficient number of studies (more than 10) reporting the same outcome. If publication bias is suggested by a significant asymmetry of the funnel plot in visual assessment, we will incorporate this into our assessment of the certainty of the evidence [71]. If our review includes only a few studies eligible for meta‐analysis, the ability to detect publication bias will be largely diminished, and we will simply note our inability to rule out possible publication bias or small‐study effects.

Synthesis methods

If we identify multiple studies that we consider to be sufficiently similar, we will perform meta‐analysis using RevMan [66]. We will combine studies in meta‑analysis only when they are sufficiently similar in key methodological and clinical characteristics, including population definition, outcome definitions, timing of follow‐up, and statistical adjustments for important confounders. Our primary analysis will include studies rated as having low, some concerns, or high risk of bias but will exclude studies judged to be at very high risk of bias. We will describe studies assessed as being at very high risk of bias narratively and will not include them in any evidence synthesis.

We will use a random‐effects model as the primary method of synthesis, given the anticipated clinical and methodological heterogeneity across the included studies. We will use the Restricted Maximum Likelihood (REML) estimator to estimate between‐trial variance. We will use the Hartung‐Knapp‐Sidik‐Jonkman method to calculate a CI for the meta‐analysis effect estimate when there are at least three studies and the estimate of heterogeneity is greater than zero. In other scenarios (i.e. in pooled analyses of two studies, or where the estimate of heterogeneity is equal to zero), we will use the Wald‐type method.

If we judge meta‐analysis to be inappropriate, we will analyze and interpret individual studies separately. We will refer to methodological guidance in Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions [72] and Synthesis Without Meta‐analysis (SWiM) reporting guidance [73]. If we find that the included studies differ in terms of underlying risk sets or sampling designs, for example case‑control versus cohort studies, we will restrict meta‑analysis to comparable effect measures or perform it separately by design. If substantial differences remain between study designs or methodological features, we will synthesize the results narratively rather than quantitatively.

We will analyze each cohort (late preterm and early term) separately in comparison with term infants.

Investigation of heterogeneity and subgroup analysis

We will describe the clinical diversity and methodological variability of the evidence narratively and in tables. Tables will include data on study characteristics, such as design features, population characteristics, and intervention details.

We will assess the clinical homogeneity of the results of the included studies based on the similarity of population, outcomes, and follow‐up.

We will assess heterogeneity after evaluating the included studies. We will reach decisions after consulting the entire review team.

To assess statistical heterogeneity, we will visually inspect forest plots and describe the direction and magnitude of effects and the degree of overlap between CIs. We will also consider the statistics generated in forest plots that measure statistical heterogeneity. We will use the I² statistic to quantify inconsistencies between the studies in each analysis. We will also consider the P value from the Chi² test to assess if this heterogeneity is significant (P < 0.1). If we identify substantial heterogeneity, we will report the finding and explore possible explanatory factors using prespecified subgroup analysis.

We will grade the degree of heterogeneity as:

• 0% to 40%: might not be important;

• 30% to 60%: may represent moderate heterogeneity;

• 50% to 90%: may represent substantial heterogeneity;

• 75% to 100%: considerable heterogeneity [74].

We will use a rough guideline to interpret the I² value rather than a simple threshold, and our interpretation will take into account the understanding that measures of heterogeneity (I² and Tau²) will be estimated with high uncertainty when the number of studies is small [74]. We will interpret tests for subgroup differences in effects with caution, given the potential for confounding with other study characteristics and the observational nature of the comparisons; see Section 10.11.2 of the Cochrane Handbook for Systematic Reviews of Interventions [69]. In particular, subgroup analyses with fewer than five studies per category are unlikely to be adequate to ascertain valid differences in effects, and we will not highlight them in our results.

We plan to carry out the following subgroup analyses of factors that may contribute to heterogeneity in the outcomes.

• Modes of delivery (vaginal delivery versus cesarean section)

• Hypertension or preeclampsia, or both, in mother (mothers with hypertension or preeclampsia, or both, versus mothers without such a condition)

• Maternal diabetes (mothers with pregestational or gestational diabetes versus mothers without such conditions)

• Antenatal steroids (antenatal steroids administered versus no antenatal steroids administered)

• Multiple pregnancies (newborns of singleton versus newborns of multiple pregnancies)

We will use the main outcomes (those specified for the summary of findings tables) in subgroup analyses if a sufficient number of studies report the outcomes to support valid subgroup comparisons (at least five studies per subgroup).

We will formally test for differences between subgroups using the test for subgroup differences in RevMan [66], which compares effect estimates across subgroups and evaluates whether these differences are statistically significant. We will interpret the findings cautiously, considering the observational nature of subgroup comparisons.

Equity‐related assessment

We will document equity‐associated characteristics as suggested by PROGRESS‐Plus [75] (place of residence, race/ethnicity/culture/language, occupation, education, socioeconomic status, social capital, age, and disability of the parents) to determine if the populations included in studies would be subject to any health inequity in terms of the outcomes that we will assess. Given there may be differences in outcomes across high‐, middle‐, and low‐income country settings, we will map study locations to income settings using the World Bank Group income classifications. We will describe any findings descriptively in our review. We will highlight and present any baseline risk differences in our population that might indicate disadvantage(s) in our summary of findings table.

Sensitivity analysis

We will assess whether meta‐analysis results are sensitive to methodological quality by re‐running the meta‐analyses excluding studies assessed to be at high risk of bias. We will present forest plots for these analyses in the supplementary material and summarize the conclusions narratively.

In addition, we will examine whether the meta‑analytic outcomes differ according to study setting by re‐running the analyses after excluding studies conducted within a single institution. The forest plots generated from these sensitivity analyses will likewise be included in the supplementary materials, accompanied by a narrative synthesis of the principal findings.

Certainty of the evidence assessment

We will use the GRADE approach, as outlined in the GRADE Handbook, and GRADEpro GDT software [76] to assess the certainty of evidence for the following (clinically relevant) outcomes, with particular guidance from documents related to observational studies [67].

• Neonatal death.

• Resuscitation in the delivery room, defined as the need for the use of positive airway pressure or chest compressions, or both, immediately after birth.

• Use ofsurfactant treatment within the first 48 hours of life.

• Need for mechanical ventilation, defined as ventilation through an endotracheal tube, during the hospital stay after birth.

• Need for non‐invasive respiratory support, defined as respiratory support without the need for an endotracheal tube (including but not limited to nCPAP, NIPPV, HFNC), during the hospital stay after birth.

• Duration of hospital stay in days.

• Admission to the NICU.

Two review authors (of KWS, MA, MG, FB, GS) will independently assess the certainty of the evidence for each of the predefined outcomes. We will resolve disagreements through discussion or by consulting a third review author (KWS or MBr). Evidence from observational studies assessed using ROBINS‑E will start at high certainty and may be downgraded based on identified concerns in the following domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias [67]. We will downgrade the certainty of evidence when there are serious or very serious concerns in any domain, following the structured framework proposed for ROBINS‑E. The certainty of evidence may also be upgraded based on three factors: dose response, large effects, and opposing residual confounding.

The GRADE system specifies four levels of certainty, as follows.

• High certainty: we are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different.

• Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect.

• Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of the effect.

We will present the main findings of the review in summary of findings tables, which will follow the structure described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions [69]. We will include key information concerning the certainty of evidence, the magnitude and CIs of the risks, and the sum of available data on the outcomes.

If sufficient data are available to conduct both syntheses, we will create two summary of findings tables, one for late‐preterm infants and one for early‐term infants, compared with full‐term infants.

We will justify all decisions to downgrade the certainty of the evidence using footnotes and make comments to aid the reader's understanding of the review where necessary.

Consumer involvement

Consumers were not involved in the design of this protocol or in the planning of the review because of limited resources.

Supporting Information

Supplementary materials are available with the online version of this article: 10.1002/14651858.CD016188.

Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.

Supplementary material 1 Search strategies

Supplementary material 2 ROBINS‐E Domains

New

Additional information

Contributions of authors

KWS reviewed the literature to inform the protocol, wrote the first draft of the protocol, and reviewed and approved the final version of the protocol.

MA reviewed the literature to inform the protocol, wrote the first draft of the protocol, and reviewed and approved the final version of the protocol.

MG drafted and reviewed the first version of the protocol.

MBj wrote the search strategy and the 'Search methods for identification of studies' section, and reviewed and approved the final version of the protocol.

CJR planned the statistical analyses and reviewed the text.

GS drafted and reviewed the first version of the protocol.

FB drafted and reviewed the first version of the protocol.

RS drafted and reviewed the manuscript.

MF drafted and reviewed the manuscript and peer‐reviewed the search strategy.

MBr reviewed the literature to inform the protocol, wrote the first draft of the protocol, and reviewed and approved the final version of the protocol.

Declarations of interest

KWS: works as an assistant professor (MD, PhD) in the II Department of Neonatology, Poznan University of Medical Sciences, Poland. No commercial or non‐commercial conflicts of interest relevant to this review.

MA: no commercial or non‐commercial conflicts of interest relevant to this review.

MG: no commercial or non‐commercial conflicts of interest relevant to this review.

MBj: no commercial or non‐commercial conflicts of interest relevant to this review.

CJR: no commercial or non‐commercial conflicts of interest relevant to this review.

GS: is employed as a neonatologist (MD, PhD) at the II Department of Neonatology, Poznan University of Medical Sciences, Poland (public employment). No other commercial or non‐commercial conflicts of interest relevant to this review.

FB: works as an MD (resident) in the II Department of Neonatology, Poznan University of Medical Sciences, Poland. No other commercial or non‐commercial conflicts of interest relevant to this review.

RS: is the Co‐ordinating Editor of the Cochrane Neonatal Group, but did not participate in the editorial assessment of this manuscript. He is the Vice President and Director of Clinical Trials of the Vermont Oxford Network, and Professor at the Larner College of Medicine, University of Vermont. He has no commercial or non‐commercial conflicts of interest relevant to this review.

MF: is a Managing Editor and Information Specialist with the Cochrane Neonatal Group, but did not participate in the editorial assessment of this manuscript. She has no commercial or non‐commercial conflicts of interest relevant to this review.

MBr: is the Director of Cochrane Sweden and Associate Editor for the Cochrane Neonatal Group. However, he had no involvement in the editorial processing of this protocol. He has no commercial or non‐commercial conflicts of interest relevant to this review.

Sources of support

Internal sources

• Institute for Clinical Sciences, Lund University, Lund, Sweden

  Matteo Bruschettini is employed by this organization.

External sources

• Vermont Oxford Network, USA

  Cochrane Neonatal Reviews are produced with support from Vermont Oxford Network, a worldwide collaboration of health professionals dedicated to providing evidence‐based care of the highest quality for newborn infants and their families.

• Region Skåne and Skåne University Hospital, Lund University, Sweden

  Cochrane Sweden received support from Region Skåne and Skåne University Hospital Lund University.

Registration and protocol

Cochrane approved the proposal of this review in June 2024.

Data, code and other materials

As part of the published Cochrane protocol, the following is made available for download for users of the Cochrane Library: search strategies (Supplementary material 1). Data sharing is not applicable to this article as it is a protocol, so no datasets were generated or analyzed.

Disclosure of artificial intelligence use

Artificial intelligence (AI) tools (Microsoft Copilot, Grammarly) were used to assist in drafting parts of this manuscript, including language refinement and paraphrasing (December 2025–April 2026). All AI‐generated content was reviewed, edited, and approved by the authors, who take full responsibility for the final version of the work.