Evidence for Global Health Care Interventions for Preterm or Low Birth Weight Infants: An Overview of Systematic Reviews

We used standard Cochrane methods for overview of systematic reviews¹⁰  to search for all published systematic reviews of randomized controlled trials (RCTs) and nonrandomized studies of interventions (NRSIs) for the 24 research questions listed in Table 1. We searched for both RCT and observational study systematic reviews. The review was registered in PROSPERO CRD42022309313.¹¹ 

The search strategy included terms for preterm and LBW and systematic review only, to avoid losing data we did not restrict the initial search terms on intervention type or PICO. The appendix lists the search strategies. The search was limited to the last 36 months, ie, from January 1, 2019 to December 31, 2021. We searched the following databases: Medline via Ovid; the Cochrane Database of Systematic Reviews; the Cochrane Database of Systematic Review Protocols; and the PROSPERO International prospective register of systematic reviews. We used the Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia,¹²  to manage all stages.

Reviews had to assess preterm (<37 weeks’ gestation) or LBW infants (<2.5kg) from 0 to 24 months. They could be cared for in both the health facility and the home in all countries (ie, high, middle, and low income countries).

Only studies that examined the impact of the interventions in Table 1 (kangaroo mother care [KMC], feeding, micronutrients, probiotics, emollients, continuous positive airways pressure [CPAP], methyl xanthines, and family involvement and family support) were included. The intervention had to be administered to the infant from birth to 24 months chronological age.

The interventions had to be compared with placebo, no intervention, or usual care as defined by trial authors.

“Critical” outcomes for inclusion in the review were mortality, morbidity, growth, and neurodevelopment.

Reviews were excluded if they: did not address at least 1 of the 24 research questions; did not include preterm or LBW infants; did not assess all of the critical outcomes (ie, mortality, morbidity, growth, and neurodevelopment); if they included small subpopulations only (eg, very preterm or very LBW infants only); or if they did not include a meta-analysis.

Two review authors (K.E. and N.S.) independently screened titles and abstracts, and assessed the full texts of all identified systematic reviews for eligibility. We assessed the reviews’ objectives and methods, including outcomes and participants for relevance and included only those reviews that meet the criteria listed above. We resolved any disagreements through discussion until we reached a consensus.

We generated a data extraction form and pretested it. After verification, 2 review authors (K.E. and N.S.) independently extracted data from each review. We resolved any discrepancies through discussion until we reached a consensus, or, if necessary, by consulting another review author. We collected basic data on: design, year of publication, year of search, number of studies, number of participants, country, health facility type and predefined covariates: population (gestational age and birth weight); intervention characteristics (type, dose, and frequency); comparator characteristics (placebo, no intervention, usual care, and other); and outcomes (mortality, morbidity, growth, and neurodevelopment).¹¹ 

We assessed the methodological quality of the systematic reviews using AMSTAR 2 (A Measurement Tool to Assess Systematic Reviews).¹³  We resolved any discrepancies through discussion until we reached a consensus. AMSTAR 2 assesses the degree to which review methods avoid bias by evaluating the methods against specific criteria. There are 7 critical domains: apriori registration of the study protocol, adequacy of the literature search, justification for excluding studies, assessment of risk of bias including publication bias, appropriateness of meta-analytical methods, assessment of conflicts, and funding source. Each item is rated as high, moderate, low, or critically low.¹³  We also collected data on the tool used to assess risk of bias and the tool used to assess certainty of evidence. We did not assess the quality of studies in network meta-analyses because of their different design.

We described each systematic review by intervention and covariate (as defined above). We also assessed whether each systematic review addressed the research question “fully” (all 4 components of the PICO), “partially” (1 to 3 components of the PICO) or “did not address” (no components of the PICO). We also assessed the methodological quality of the reviews using AMSTAR2 criteria and the overall total quality rating.

We identified 1595 records (Fig 1). After removing duplicates and screening titles, abstracts, and full text articles we included 9 reviews.^14–22  Sixteen reviews were excluded because of wrong population (ie, included very LBW or very preterm infants only), 6 had the wrong study design (eg, were literature reviews without a systematic review or meta-analysis), 2 had the wrong intervention (eg, restricted on feed volume), and 15 had insufficient outcomes assessed (ie, 3 reviews only assessed necrotizing enterocolitis, 3 only assessed bronchopulmonary dysplasia, and 1 only assessed apnea, sepsis, growth, biomarkers). Characteristics of excluded reviews can be found in the appendix.

The characteristics of the 9 included reviews are shown in Table 2. There were a total of 121 studies and 25 465 preterm or LBW infants in the 9 reviews. One review included no studies (ie, was an ”empty” review),¹⁴  all the other reviews included only RCTs, and there was 1 network meta-analysis.¹⁹  The number of studies in the reviews ranged from 4 (CPAP) to 45 (probiotics) (median 8, interquartile range [IQR] 5–18, mean 13, standard deviations [SD] 14). The number of participants ranged from 119 (CPAP) to 12 320 (probiotics) (median 1456, IQR 322–3201, mean 2829, SD 3987). Three reviews included under 1000 infants.^17,20,21  All reviews only included hospitalized infants. Only 1 review included studies from low income countries,¹⁹  and 2 included only high income countries.^15,20 

All 9 systematic reviews “fully addressed” (ie, examined all components of the PICOs) of 9 research questions (mother’s own milk, donor human milk, fortification, infant formula, emollients, probiotics, or any CPAP, early and prophylactic CPAP). Fifteen research questions (KMC, early initiation, responsive feeding, advancement, exclusive breastfeeding (EBF) duration, iron, zinc, vitamin D, vitamin A, calcium and phosphorous, multiple micronutrients, CPAP pressure source, methyl xanthines, family involvement, and family support) had no systematic review.

The AMSTAR quality assessment is summarized in Table 3. We were not able to assess the quality of 2 reviews; 1 was a network meta-analysis,¹⁹  and 1 was an empty review¹⁴  which found no trials. All the other 7 reviews were rated as high. All reviews used the Cochrane risk of bias (ROB) tool,²³  and the GRADE system for assessing the certainty of the body of evidence.^24,25  All used a comprehensive search strategy, all assessed the potential impact of bias on the results of the meta-analysis, and accounted for bias in discussing the results, and all assessed heterogeneity and small study bias. Two reviews did not report if they performed study selection in duplicate and 5 reviews did not report on the funding of studies.

In our overview of systematic reviews, we found 9 systematic reviews of 121 studies and 25 465 preterm or LBW infants published in the last 36 months that addressed 9 of our 24 research questions. The remaining 15 research questions had no systematic review that addressed the research question. The included systematic reviews assessed mother’s own milk, donor human milk, fortification, infant formula, emollients, probiotics, any CPAP, and early and prophylactic CPAP. However, important interventions such as KMC, early initiation, responsive feeding, advancement, duration of EBF, iron, zinc, vitamin A, vitamin D, calcium and phosphorous, multiple micronutrients, methyl xanthines, family involvement, and family support did not have a recent systematic review that covered all PICO elements.

Thirty nine reviews met the inclusion criteria partially (ie, covered some element of each PICO) but were excluded from the analysis. Four reviews were excluded as they did not include meta-analyses, 2 reviews assessed the wrong interventions, and 15 did not assess all critical outcomes (mortality, morbidity, growth, and neurodevelopment) (eg, they reviewed only single morbidities such as necrotizing enterocolitis, bronchopulmonary dysplasia, apnea, sepsis, or biomarkers). We had decided apriori that each of the outcomes, mortality, morbidity, growth, and neurodevelopment, were separately “crucial” to the understanding of impact of interventions in preterm and LBW infants. Thus, we did not consider a review to be “complete” unless each of these outcomes were included. Sixteen reviews were excluded because they restricted their target populations to very LBW or very preterm infants. Infants between 32 and 36 weeks’ gestation have higher risks of mortality and morbidity than term infants,^3,26  and we felt that that complete reviews should also assess effectiveness in these older infants. Encouragingly, most reviews included term LBW infants, highlighting the now well-known vulnerabilities of these infants.^2–4  However, all the reviews only included studies from hospitalized infants and only 1 study was from a low income country.

All the 9 reviews were rated as high quality and the authors used standard tools for assessing risk of bias and the certainty of the body of evidence, used a comprehensive search strategy, and assessed the potential impact of bias on the results of the meta-analysis and heterogeneity. The authors assessed small study bias if data were available, otherwise, they reported reasons for not completing this (ie, not enough studies available). The only reasons for rating the systematic reviews down were that 2 did not complete study selection in duplicate, and 5 reviews did not state the funding sources of included studies.

There are many current publications of the “estimates” of the prevalence of LBW and prematurity and the “risks” and burden of ill health in preterm and LBW infants.^1–5,26  However, to our knowledge, this is the first overview of systematic reviews of trials of health care interventions for the care of preterm and LBW infants.¹⁰  This lack of “overview” of trials is concerning as we located 48 systematic reviews and 700 trials evaluating health care for preterm and LBW infants just in the last 36 months. In 2010, a global research priority setting exercise identified many research gaps for the care of the preterm and LBW infant.⁶  Important questions about the efficacy and effectiveness of interventions for preterm and LBW infants have been answered over the last 10 years, yet our overview has also shown that important gaps in the global evidence base remain.

Our overview had some limitations. We restricted our search only to those interventions identified as a priority at a World Health Organization scoping meeting. Other interventions that may be of importance to preterm or LBW infants were not considered as they were out of scope for this review. We are planning an additional overview of systematic reviews using similar methods to assess these other interventions.

Overall, it is encouraging that almost a third of our research questions were addressed by high quality systematic reviews of over 120 trials and 26 000 infants. However, we found important gaps. Interventions, such as feeding of preterm and LBW infants, especially “what to feed” (micronutrient deficiencies), “how to feed” (ie, the mechanics of feeding young infants), and reviews of family involvement and support had no recent systematic reviews that fully addressed important research questions. All studies were from hospitalized infants and only 1 was from a low income country. These deficits are concerning and need to be urgently addressed.

Systematic reviews of intervention trials are a crucial component of the knowledge base for the health care of all populations. International agencies and the research community must maintain their focus on support of rigorous systematic assessment and meta-analysies of intervention studies. This is especially important for infants with the highest burden of ill health such as those who are born preterm and LBW.