Skip to main content
NCBI home page
Eye logoLink to Eye
. 2023 Oct 18;38(5):877–884. doi: 10.1038/s41433-023-02796-2

Efficacy comparison of 21 interventions to prevent retinopathy of prematurity: a Bayesian network meta-analysis of randomized controlled trials

Miao Zhou 1,2,#, Pei-Chen Duan 1,#, Dan-Lin Li 1, Jing-Hong Liang 3, Gang Liang 4, Hua Xu 5,✉, Chen-Wei Pan 1,✉
PMCID: PMC10965999  PMID: 37853107

Abstract

Background

The objective of this study was to evaluate the comparative efficacy of current interventions for the prevention of retinopathy of prematurity (ROP) in premature infants.

Methods

A network meta-analysis (NMA) was performed. We searched PubMed, Web of Science, Scopus, Embase, and the Cochrane Library for relevant studies from their inception to May 5, 2022. Publications were eligible for our study if they were randomized controlled trials (RCTs) involving preterm infants at <37 weeks of gestational age and reported the incidence of any-stage ROP treated with the interventions of interest. The overall effect was pooled using the random effects model.

Results

We identified 106 RCTs (involving 23894 participants). This NMA showed that vitamin A supplementation markedly reduced the incidence of ROP, in comparison with placebo (odds ratio [OR] = 0.59, 95% credible interval [95% CrI] 0.33, 0.85), fish oil-based lipid emulsion (OR = 0.57, 95% CrI 0.24, 0.90), early erythropoietin (OR = 0.51, 95% CrI 0.34, 0.98), probiotics (OR = 0.48, 95% CrI 0.32, 0.97), and human milk (OR = 0.50, 95% CrI 0.21, 0.78). Vitamin A supplementation has the highest probability of being the best option for reducing the ROP risk compared with the other 20 interventions based on its surface under the cumulative ranking curve (SUCRA) value (SUCRA = 92.50%, 95% CrI 0.71, 1.00).

Conclusions

Our findings suggest that among 21 interventions, vitamin A supplementation might be the best method of preventing ROP. This NMA offers an important resource for further efforts to develop preventive strategies for ROP.

Subject terms: Paediatrics, Public health, Retinal diseases

Introduction

Retinopathy of prematurity (ROP) is a leading cause of childhood blindness [1, 2]. Due to the tremendous improvement in supportive services, the higher survival rate of preterm infants has unexpectedly increased the incidence of ROP than before, equivalently [3–5]. Although there have been several therapies for severe ROP, they are either destructive (laser therapy and cryotherapy) [6] or expensive (anti-vascular endothelial growth factor treatment) [7]. Current therapeutic options may also have potentially serious complications [8]. Therefore, at this stage, valid prevention remains the best strategy for avoiding ROP-related blindness [9].

Although preterm delivery has been identified as a major risk factor for ROP, a variety of other clinical factors are also associated with an increased risk of ROP, including low gestational age (GA) [10], low birth weight (BW) [11], hyperoxia following delivery [12, 13], and several other modifiable clinical factors such as blood transfusion [14], perinatal infection [15], and poor postnatal weight gain [16]. While there have been many trials [17–19] and systematic reviews [1, 20, 21] of potential interventions for ROP, the findings remain conflicting. Moreover, little attention has been paid to comparing these categories of preventive strategies to rank the effectiveness of the available interventions for preventing ROP.

Network meta-analysis (NMA) is a useful approach to comparing the effects and efficacies of multiple interventions in medical research [22–25]. In this study, we conducted a NMA with the aim of recommending an optimal prevention strategy for ROP by summarizing and analyzing the available high-quality evidence. The findings should provide valuable information for the development of convincing evidence-based clinical guidelines for preventing ROP in premature infants.

Methods

This NMA of randomized controlled trials (RCTs) was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [26] and the PRISMA extension for NMA (PRISMA-NMA) [27]. The protocol for this study was registered on INPLASY (registration number: INPLASY2022110002).

Selection criteria

In accordance with the Population, Intervention, Comparison, Outcomes, and Study selection criteria, RCTs were included if they enrolled infants of <37 weeks GA, compared any of the interventions of interest, and demonstrated the incidence of any-stage ROP. The severity of ROP can be classified into five stages, and “any-stage ROP” refers to stages 1–5 [28].

The choice of intervention measures was based on recently published reviews [1, 29, 30] and the current understanding of the pathophysiology of ROP. Because ROP is a typically biphasic disease [28], the timing of initiation of the interventions was also taken into consideration. Finally, these interventions fell into seven main categories and 21 types: (1) postnatal nutrition: human milk (HM), fish-oil-based lipid emulsion (FO), and nutritional supplementation with vitamin A (VA), vitamin E (VE), inositol (IN), or lutein (LU); (2) blood transfusion: the use of hemoglobin transfusion guidelines (TG), early initiation of erythropoietin (early EPO, EE, initiated at <8 days of age), late initiation of erythropoietin (late EPO, LE, initiated at >8 days of age), and all EPO (AE, early and late EPO); (3) infection reduction: fluconazole prophylaxis (FP), probiotics (PS), or lactoferrin supplementation (LS); (4) nonsteroidal anti-inflammatory drugs: indomethacin (IND), ibuprofen (IB); (5) corticosteroids: early initiation of dexamethasone (ED, initiated within the first week of life), late initiation of dexamethasone (LD, initiated at >7 days of age); (6) caffeine: early initiation of caffeine (EC, initiated at <3 days of age), late initiation of caffeine (LC, initiated at ≥3 days of age); and (7) other interventions: d-penicillamine (DP), superoxide dismutase (SOD). Where two or more trials reported the same population, the primary study was preferentially included to avoid duplicate data.

Although oxygen supplementation in the delivery room and in the neonatal intensive care unit are commonly used for the respiratory support of premature neonates [30], we did not consider management with supplemental oxygen as one of the interventions of interest in this analysis, mainly because studies of oxygen saturation always compare higher and lower oxygen saturation targets rather than a placebo group or another intervention. Moreover, the potential adverse outcomes of oxygen management on neonatal death before hospital discharge dampened our interest in this intervention [31, 32].

Search strategy and study selection

PubMed, Web of Science, Scopus, Embase, and the Cochrane Library were searched for pertinent literature, from their inception to May 5, 2022. We applied PubMed medical subject headings and keywords in combination with Boolean logical operators as comprehensively as possible, using “retinopathy of prematurity”, “preterm infants”, “nutrition”, “anti-infective”, “transfusion”, “randomized controlled trials”, and additional relevant conceptual keywords. The detailed strategy is shown in Supplementary search strategy. We also searched the reference lists of identified studies and references to articles to avoid omitting potentially relevant studies.

The literature search was conducted independently by two authors (MZ and PCD) to assess the eligibility of studies for inclusion. We first removed duplicate publications and screened the studies for their eligibility for initial inclusion based simultaneously on their titles and abstracts. With full-text screening, the two authors independently undertook the retrieval and perusal of articles for further assessment. Any disagreement was settled by discussion with and arbitration by a third author (CWP).

Data collection and quality assessment

Two authors (MZ and PCD) independently extracted the following information from each study: intervention details, demographics of the participants, and outcomes of interest. For the outcomes of interest, the available dichotomous data were collected, including the number of premature infants with any-stage ROP and the total sample size in each arm. A standardized data extraction form was adapted based on the guidelines of the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS) checklist [33], covering study information, database information, population characteristics, interventions, outcomes, and analysis. Disagreements were resolved by a third author (CWP). While when the data were not provided in the eligible literature, we would use the effect measures reported by the individual trials (e.g., odds ratios) along with the corresponding 95% confidence intervals, or other statistical indices for the NMA.

We appraised the risk of bias (ROB) in all of the included studies with the Cochrane Risk of Bias Tool (RoB 2.0) [34]. Two investigators (MZ and PCD) assessed each trial independently according to the relevant quality criteria, and any disagreement was resolved by discussion and consensus. ROB was evaluated in Review Manager 5.4.1.

Statistical analysis

We conducted a Bayesian random-effects NMA to calculate the odds ratio (OR) and 95% credible interval (CrI), with the noninformative priors integrating direct and indirect evidence, to compare 21 preventive interventions for ROP [35, 36]. The extent of heterogeneity between different studies was quantified using the I2 statistical test, and values >50% indicated a significant degree of heterogeneity [37]. To adequately ensure the similarity of various strategy comparisons which could provide valid indirect inferences, we compared the clinical and methodological characteristics such as participants and experimental designs across all the included studies to apprise the transitivity assumption [38]. We evaluated global inconsistency in the overall network by fitting the design-by-treatment interaction approach [39, 40]. Local inconsistency between direct and indirect evidence was assessed by the node-splitting method [39]. Evidence of publication bias was evaluated using a funnel plot and the Egger’s test using Stata 12.0. We used a radar plot to provide a visual representation of the relationships between each intervention and the study characteristics [41].

The posterior densities of the unknown parameters were estimated using Markov-chain Monte Carlo simulations [42]. Three parallel relevant Markov chains were set up to run with arbitrarily chosen initial values [43]. Each chain used 500,000 iterations. The first 20,000 iterations represented the burn-in period and were discarded. The surface under the cumulative ranking curve (SUCRA) value for each intervention was taken as the estimated probability used to cumulatively rank the preventive interventions. SUCRA takes a value between 0 and 1, where 0 certainly denotes the worst intervention and 1 certainly indicates the best intervention [44]. The NMA was performed using WinBUGS 1.4.3. We also performed sensitivity analyses by removing trials with a total sample size of <100 infants to ensure the robustness of our findings [45]. We conducted further subgroup analyses stratified by the total sample size (≥100 vs <100), gestational age (≥28 weeks vs <28 weeks) and birth weight (≥1000 g vs <1000 g) of each trial.

Results

Figure 1 shows the flowchart of publication selection. After searching five databases and scanning other sources, we selected a total of 1102 records from the initial literature search, after duplicates were removed. Finally, 106 publications were considered eligible for our NMA.

Fig. 1.

Fig. 1

Open in a new tab

Selection process for articles in the review (n = 106).

Baseline characteristics and quality assessment

These studies involved 23894 participants across 21 different preventive interventions and provided data published between 1981 and 2021. Notably, 12,063 infants were assigned to the intervention group and 11831 were assigned to the control group. Table 1 summarizes the key characteristics of the eligible trials and participants across the 106 studies. Detailed descriptions of the study and participant characteristics of each trial are provided in Supplementary Table 1. Forty-eight studies were published in the last decade (45.3%). Twenty-two trials examined the effect of all EPO on ROP prevention. Thirty-nine studies had sample sizes of 21–50 participants in the intervention group. Eighty-nine (84%) of the eligible trials were conducted in very-low-birth-weight (VLBW, birth weight <1500 g) infants. In 48 studies, the participants were born at <28 weeks GA. The overall and individual-study-level quality of the studies are shown in Supplementary Figs. 1 and 2, respectively. The overall ROB was low.

Table 1.

Demographic characteristics of included studies and their participants (n = 106).

Publication Year Birth Weight (g)
 1981–1990 13 (12.3%) 652–1000 44 (41.5%)
 1991–2000 14 (13.2%) 1001–1250 32 (30.2%)
 2001–2010 31 (29.2%) 1251–1500 13 (12.3%)
 2011– 48 (45.3%) NA 17 (16.0%)
Interventions Gestational Age at Birtha (wk)
 Human Milk 6 (5.7%) 24.0–28.0 48 (45.3%)
 Fish Oil-Based Lipid Emulsion 6 (5.7%) 28.1–30.0 31 (29.2%)
 Vitamin A 5 (4.7%) 30.1– 10 (9.4%)
 Vitamin E 7 (6.6%) NA 17 (16.0%)
 Lutein 3 (2.8%)
 Inositol 5 (4.7%) Proportion of Male Sexa (%)
 PRBC Transfusion Guidelines 4 (3.8%) ≥50 55 (51.9%)
 Early Erythropoietin 20 (18.9%) <50 26 (24.5%)
 Late Erythropoietin 7 (6.6%) NA 25 (23.6%)
 All Erythropoietin 22 (20.8%)
 Fluconazole Prophylaxis 4 (3.8%) SGAa (n)
 Lactoferrin Supplementation 3 (2.8%) 1–10 8 (7.5%)
 Probiotics Supplementation 12 (11.3%) 11–20 6 (5.7%)
 Ibuprofen 5 (4.7%) 21– 7 (6.6%)
 Indomethacin 3 (2.8%) NA 85 (80.2%)
 Early Caffeine 3 (2.8%)
 Late Caffeine 1 (0.9%) Proportion of Received Antenatal Steriodsa (%)
 Early Dexamethasone 5 (4.7%) 0–30.0 2 (1.9%)
 Late Dexamethasone 4 (3.8%) 30.1–60.0 6 (5.7%)
 D-penicillamine 4 (3.8%) 60.1–90.0 31 (29.2%)
 Superoxide Dismutase 2 (1.9%) 90.1– 13 (12.3%)
NA 54 (50.9%)
Sample Sizea (n)
 7–20 10 (9.4%) Apgar Scorea
 21–50 39 (36.8%) 6.0–6.9 14 (13.2%)
 51–100 24 (22.6%) 7.0–7.9 19 (17.9%)
 101–500 28 (26.4%) 8.0– 9 (8.5%)
 501– 5 (4.7%) NA 64 (60.4%)
Open in a new tab

NA not applicable, SGA small for gestational age.

aOnly intervention group presented.

Results of the NMA between interventions and the incidence of ROP

A radar plot is shown in Fig. 2. The central node of the plot represented the placebo group, and each category of intervention was connected to the node in the center by lines, which displayed a direct relationship between interventions and the placebo. The total sample size of each intervention (n = 406–4651), SUCRA values (2.68–92.51%), and the number of the arms (n = 1–29) fell into 10 levels, reflected in the distance from the central node. The included trials contained 21 intervention arms: all EPO was the most commonly recorded intervention in 29 arms (total sample size, n = 4651), followed by early EPO in 27 arms (n = 4194); PS in 14 arms (n = 3894); VE in 13 arms (n = 1094); FO in 10 arms (n = 574); late EPO in 9 arms (n = 602); DP in 7 arms (n = 836); HM (n = 1495), VA (n = 624), IN (n = 931), TG (n = 549), and ED (n = 920) in 7 arms; LU in 6 arms (n = 406); IB (n = 369) and LD (n = 661) in 5 arms; FP (n = 967), EC (n = 2067), LS (n = 2597), and SOD (n = 1208) in 4 arms; IND in 3 arms (n = 1144); and LC in 1 arm (n = 108).

Fig. 2. Radar plot.

Fig. 2

Open in a new tab

Legend: SUCRA values (3.03%–92.10%); sample size (n = 406–4651); arms (n = 1–29). Abbreviations: SUCRA: surface under the cumulative ranking curve.

A league table of all pairwise comparisons is presented in Fig. 3. We evaluated the efficacy of the 21 interventions, from baseline to the end of the intervention, in the prevention of ROP in the 106 trials (n = 23,894) and found that only VA supplementation was statistically significantly superior to the placebo (OR = 0.59, 95% CrI 0.33, 0.85). VA was also significantly more beneficial than FO (OR = 0.57, 95% CrI 0.24, 0.90), EE (OR = 0.51, 95% CrI 0.34, 0.98), PS (OR = 0.48, 95% CrI 0.32, 0.97), and HM (OR = 0.50, 95% CrI 0.21, 0.78). EC was significantly more effective than PS (OR = 0.66, 95% CrI 0.36, 0.97), HM (OR = 0.65, 95% CrI 0.31, 0.99), and FP (OR = 0.59, 95% CrI 0.25, 0.94). Only SOD was significantly superior to DP (OR = 0.41, 95% CrI 0.10, 0.72).

Fig. 3. Relative effects of interventions according to the network meta-analysis.

Fig. 3

Open in a new tab

Legend: Non-significant pairwise comparisons; significant pairwise comparisons; SUCRA values; interventions. Abbreviations: PLA Placebo, HM human milk, FO fish-oil-based lipid emulsion, VA vitamin A, VE vitamin E, LU lutein, IN inositol, TG PRBC Transfusion Guidelines, EE early erythropoietin, LE late erythropoietin, AE all erythropoietin, FP fluconazole prophylaxis, PS probiotic supplementation, LS lactoferrin supplementation, IB ibuprofen, IND indomethacin, EC early caffeine, LC late caffeine, ED early dexamethasone, LD late dexamethasone, DP D-penicillamine, SOD superoxide dismutase, SUCRA surface under the cumulative ranking curve, OR odds ratio, 95% CrI 95% credible interval.

A SUCRA line was drawn to rank the hierarchy of each preventive intervention (Fig. 4 and Supplementary Fig. 3). According to SUCRA, VA (SUCRA = 92.50%, 95% CrI 0.71, 1.00) had the highest probability of minimizing the risk of ROP compared with the other 20 interventions. LC (SUCRA = 92.50%, 95% CrI 0.05, 1.00) and EC (SUCRA = 80.82%, 95% CrI 0.38, 0.95) also had relatively high rankings across the 21 interventions. Followed by SOD (SUCRA = 70.22%, 95% CrI 0.19, 0.95), ED (SUCRA = 70.17%, 95% CrI 0.19, 0.95), VE (SUCRA = 61.31%, 95% CrI 0.19, 0.90), LS (SUCRA = 59.24%, 95% CrI 0.10, 0.95), IN (SUCRA = 57.23%, 95% CrI 0.10, 0.90), LU (SUCRA = 53.24%, 95% CrI 0.05, 0.95), AE (SUCRA = 52.29%, 95% CrI 0.19, 0.81), FO (SUCRA = 45.06%, 95% CrI 0.05, 0.86), IB (SUCRA = 44.49%, 95% CrI 0.05, 0.90), IND (SUCRA = 39.58%, 95% CrI 0.05, 0.90), TG (SUCRA = 38.36%, 95% CrI 0.05, 0.90), EE (SUCRA = 38.19%, 95% CrI 0.10, 0.71), LD (SUCRA = 32.68%, 95% CrI 0.05, 0.86), LE (SUCRA = 32.22%, 95% CrI 0.05, 0.81), PS (SUCRA = 31.27%, 95% CrI 0.05, 0.71), HM (SUCRA = 30.03%, 95% CrI 0.05, 0.76), FP (SUCRA = 22.01%, 95% CrI 0.00, 0.76), DP (SUCRA = 2.68%, 95% CrI 0.00, 0.29) got an inferior ranking.

Fig. 4. Results of network meta-analysis using the placebo as the referent intervention.

Fig. 4

Open in a new tab

SUCRA surface under the cumulative ranking curve, 95% CrI 95% credible interval.

Heterogeneity, inconsistency, publication bias, and sensitivity analysis

There was no statistical heterogeneity in our NMA of RCTs (I2  =  0%). Neither global inconsistency nor local inconsistency were found between direct and indirect evidence (på 0.05). A funnel plot revealed the slight asymmetrical distribution of the comparisons (Supplementary Fig. 4). There was no significant publication bias as indicated by Egger’s test (p = 0.083). After trials with a small sample size (<100 infants) were removed, sensitivity analyses revealed that only VA supplementation was still significantly superior to the placebo (OR = 0.57, 95% CrI 0.33, 0.99; Supplementary Fig. 5), consistent with the main analysis.

Subgroup analysis

In studies with a total sample size of ≥100 infants, VA supplementation was significantly more beneficial than the placebo (OR = 0.57, 95% CrI 0.33, 0.99; Supplementary Fig. 6). In studies with enrolled infants of <28 weeks GA, D-penicillamine was significantly more inferior than the placebo (OR = 0.18, 95% CrI 0.04, 0.74; Supplementary Fig. 7). In trials conducted in infants with BW ≥ 1000 g, inositol was significantly superior to the placebo (OR = 0.44, 95% CrI 0.21, 0.94; Supplementary Fig. 8). No significant difference was found in the comparisons between other interventions and the placebo in the six subgroup analyses.

Discussion

To our knowledge, this is the first comprehensive Bayesian NMA of interventions aimed at different clinical factors related to the increased risk of ROP. In this study, we summarized the available evidence and found that vitamin A supplementation had the highest probability of being the most effective intervention in preventing ROP (SUCRA = 0.92). The large body of high-quality RCTs that was included, together with the NMA methodological approach applied in this study, should ensure useful clinical guidance for the clinical management of ROP.

Most of the risk factors associated with ROP are avoidable. Therefore, various preventive measures have been used at different levels in health institutions [46]. Among these interventions, vitamin A is well established as an essential micronutrient for the maintenance of normal visual function [47]. Interestingly, our results suggest that vitamin A supplementation may be the preferred option for the prevention of ROP in premature infants. Despite the absence of consensus about the effect of vitamin A on improving the outcomes of ROP, a range of relative studies have been done [19, 48–50], and most of them are consistent with this study when comes to affirmation of the efficacy of vitamin A decreasing the chance of ROP occurrence [19, 48, 49]. Possible mechanisms may be conducted by allowing sufficient rhodopsin for phototransduction, lowering the metabolic requirements of the developing photoreceptor, or protecting the rod outer segments from the toxicity of hyperoxia or hypoxia related to the onset and progression of ROP [50].

Our finding that vitamin A may be the optimal intervention for the prevention of ROP has both clinical and research significance. From a clinical perspective, although in the short term, the first and foremost key approach to preventing ROP is still the adoption of a screening strategy [46, 51], this strategy potentially entails some programmatic challenges in developing countries, including: (1) the low adherence to screening guidelines due to the tremendous differences in resource accessibility among different facilities; (2) the large discrepancy between the number of infants who require ROP screening and the number of trained ophthalmologists; and (3) the lack of inexpensive and easy-to-handle imaging facilities that can be used by non-ophthalmologists, such as neonatologists [46]. Therefore, from the perspective of intervention, our results may provide a more cost-effective and easy-to-access strategy than screening in developing countries. From a research standpoint, further trials are required to confirm the efficacy of this strategy and to clarify the optimal delivery method and dose of vitamin A required to reduce the ROP incidence and avoid overdosing, which have not been defined for ROP prevention, despite multiple trials [52, 53]. Unfortunately, many factors can lead to overdosing with vitamin A [54]. For example, parenteral nutrition, which is generally administered to preterm infants and typically contains vitamin A, may also influence the plasma levels of vitamin A [19]. Therefore, further trials are required to establish protocols for the detection and maintenance of serum vitamin A levels to ensure the effectiveness of vitamin A in preventing ROP.

The role of caffeine in ROP prevention has recently attracted increasing attention [55, 56]. However, the efficacy and timing of caffeine administration in preventing ROP do not have unanimous consensus [57]. In our study, although the efficacy of the late and early use of caffeine ranked second and third among the 21 interventions, the NMA comparison showed no significant advantage compared with the placebo group, consistent with the findings of a previous study [58]. The potential additive or even synergistic antioxidant effects of caffeine and SOD, the efficacy of which ranked fourth on our NMA, may be a possible explanation. The mechanism underlying the action of caffeine in preventing ROP is unclear but may involve its regulation of angiogenesis and oxidative stress [29]. Petrucci et al. suggested that caffeine can be hardly regarded as an antioxidant [59], while its coadministration with recombinant human SOD could change this result [60]. This raises the question of whether the antioxidant properties of caffeine act synergistically with those of recombinant human Cu, Zn-SOD [60]. However, we advise caution when interpreting these findings because the poor efficacy of caffeine may not be attributable to the intervention itself but to the small number of RCTs included and subjects enrolled (only one RCT addressed the late initiation of caffeine). Although The Caffeine for Apnea of Prematurity Trial reported that caffeine may also reduce the incidence of ROP [56], evidence for the efficacy of caffeine beyond the indications of interest including ROP is of low quality [61]. Therefore, larger and multicentre RCTs are required to confirm the efficacy of caffeine in the prevention of ROP in premature infants.

Our study is the first comprehensive Bayesian network meta-analysis comparing the effectiveness of interventions to prevent ROP in preterm infants, which updates the comprehensive conventional meta-analysis, undertaken in 2016 [1]. There are several differences between the 2016 study and our study. First, in terms of the number and characteristics of the included studies, the previous publication included 67 studies consisting of both RCTs and observational studies, whereas our study included only RCTs and identified 106 RCTs. Moreover, given the methodological limitations, the 2016 study only evaluated the effect of one intervention against that of a placebo and therefore provided less-robust information. Our study used the Bayesian NMA technique to extend that previous study and to provide an up-to-date overview of the efficacy of various intervention strategies in preventing ROP. We also included more intervention strategies than the 2016 paper, based on recent reviews [29, 30], so our results should be more comprehensive than the previous meta-analysis.

Several limitations of our study must be noted. First, the findings of this NMA were potentially influenced by inevitable confounding factors (e.g., the total sample size of each RCT, the GA and BW of infants, or the quality of trials). The comparability of the studies included in the NMA may be limited by heterogeneity in terms of population characteristics, intervention protocols, and outcome measurements. The different results of different subgroup analyses in the study support that. Heterogeneity among included studies can affect the validity of the pooled effect estimates and limit the generalizability of the results. Besides, no dosage or toxicity information for these interventions in infants can be inferred from this study. Finally, indirect comparisons with NMA may lead to misleading results [62]. Considering that the precision of NMA can be increased by the use of individual participant data (IPD), we suggest future NMAs on this research topic use IPD to improve the precision of the results. And this would also allow precise subgroup and dose-response analyses. Further IPD-based NMAs and multi-center trials should be conducted to provide a more precise estimation of intervention effects. Trials are also warranted to gather and synthesize information on the dosage and toxicity of the interventions in infants.

Our study indicated that vitamin A supplementation may be the best intervention for reducing the likelihood of ROP in preterm infants. However, it must be noted that this study only provides avenues for further research and no dosage or toxicity information can be inferred from this meta-analysis. Additional research is required to better understand the relative benefits of each interventional strategy used to prevent ROP.

Summary

What was known before

  • While there have been many trials and systematic reviews of potential interventions for retinopathy of prematurity (ROP), the findings remain conflicting. Moreover, little attention has been paid to comparing these categories of preventive strategies to rank the effectiveness of the available interventions for preventing ROP.

What this study adds

  • To our knowledge, this is the first comprehensive Bayesian network meta-analysis (NMA) of interventions aimed at different clinical factors related to the increased risk of ROP. Our findings suggest that among 21 interventions, vitamin A supplementation might be the best method of preventing ROP. This NMA offers an important resource for further efforts to develop preventive strategies for ROP.

Supplementary information

Supplementary Materials (8.4MB, pdf)

Author contributions

Conception and design of the research: HX, GL, and CWP; Acquisition and interpretation of the data: MZ, PCD, and DLL; Statistical analysis and writing of the manuscript: MZ, PCD, and JHL; Critical revision of the manuscript: HX, GL, and CWP.

Funding

This study was supported by the National Natural Science Foundation of China (82122059) and Suzhou Science and Technology Development Program (SKY2022172).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

These authors contributed equally: Miao Zhou, Pei-Chen Duan.

Contributor Information

Hua Xu, Email: xuhua01431@suda.edu.cn.

Chen-Wei Pan, Email: pcwonly@gmail.com.

Supplementary information

The online version contains supplementary material available at 10.1038/s41433-023-02796-2.

References

  • 1.Fang JL, Sorita A, Carey WA, Colby CE, Murad MH, Alahdab F. Interventions to prevent retinopathy of prematurity: a meta-analysis. Pediatrics. 2016;137:e20153387. doi: 10.1542/peds.2015-3387. [DOI] [PubMed] [Google Scholar]
  • 2.Cavallaro G, Filippi L, Bagnoli P, La Marca G, Cristofori G, Raffaeli G, et al. The pathophysiology of retinopathy of prematurity: an update of previous and recent knowledge. Acta Ophthalmol. 2014;92:2–20. doi: 10.1111/aos.12049. [DOI] [PubMed] [Google Scholar]
  • 3.Hack M, Wright LL, Shankaran S, Tyson JE, Horbar JD, Bauer CR, et al. Very-low-birth-weight outcomes of the National Institute of Child Health and Human Development Neonatal Network, November 1989 to October 1990. Am J Obstet Gynecol. 1995;172:457–64. doi: 10.1016/0002-9378(95)90557-x. [DOI] [PubMed] [Google Scholar]
  • 4.Stoll BJ, Hansen NI, Bell EF, Shankaran S, Laptook AR, Walsh MC, et al. Neonatal outcomes of extremely preterm infants from the NICHD Neonatal Research Network. Pediatrics. 2010;126:443–56. doi: 10.1542/peds.2009-2959. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Lemons JA, Bauer CR, Oh W, Korones SB, Papile LA, Stoll BJ, et al. Very low birth weight outcomes of the National Institute of Child health and human development neonatal research network, January 1995 through December 1996. NICHD Neonatal Research Network. Pediatrics. 2001;107:E1. doi: 10.1542/peds.107.1.e1. [DOI] [PubMed] [Google Scholar]
  • 6.Li Z, Zhang Y, Liao Y, Zeng R, Zeng P, Lan Y. Comparison of efficacy between anti-vascular endothelial growth factor (VEGF) and laser treatment in Type-1 and threshold retinopathy of prematurity (ROP) BMC Ophthalmol. 2018;18:19. doi: 10.1186/s12886-018-0685-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sankar MJ, Sankar J, Chandra P. Anti-vascular endothelial growth factor (VEGF) drugs for treatment of retinopathy of prematurity. Cochrane Database Syst Rev. 2018;1:CD009734. doi: 10.1002/14651858.CD009734.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.VanderVeen DK, Melia M, Yang MB, Hutchinson AK, Wilson LB, Lambert SR. Anti-vascular endothelial growth factor therapy for primary treatment of type 1 retinopathy of prematurity: a report by the American Academy of Ophthalmology. Ophthalmology. 2017;124:619–33. doi: 10.1016/j.ophtha.2016.12.025. [DOI] [PubMed] [Google Scholar]
  • 9.Wheatley CM, Dickinson JL, Mackey DA, Craig JE, Sale MM. Retinopathy of prematurity: recent advances in our understanding. Br J Ophthalmol. 2002;86:696–700. doi: 10.1136/bjo.86.6.696. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Chen M, Citil A, McCabe F, Leicht KM, Fiascone J, Dammann CE, et al. Infection, oxygen, and immaturity: interacting risk factors for retinopathy of prematurity. Neonatology. 2011;99:125–32. doi: 10.1159/000312821. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Fierson WM, Ophthalmology AAOPSo, American Academy Of O, American Association For Pediatric O, Strabismus, American Association Of Certified O. Screening examination of premature infants for retinopathy of prematurity. Pediatrics. 2018;142:e20183061. doi: 10.1542/peds.2018-3061. [DOI] [PubMed] [Google Scholar]
  • 12.Hartnett ME, Lane RH. Effects of oxygen on the development and severity of retinopathy of prematurity. J AAPOS. 2013;17:229–34. doi: 10.1016/j.jaapos.2012.12.155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hartnett ME. Pathophysiology and mechanisms of severe retinopathy of prematurity. Ophthalmology. 2015;122:200–10. doi: 10.1016/j.ophtha.2014.07.050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hauspurg AK, Allred EN, Vanderveen DK, Chen M, Bednarek FJ, Cole C, et al. Blood gases and retinopathy of prematurity: the ELGAN study. Neonatology. 2011;99:104–11. doi: 10.1159/000308454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Lee J, Dammann O. Perinatal infection, inflammation, and retinopathy of prematurity. Semin Fetal Neonatal Med. 2012;17:26–9. doi: 10.1016/j.siny.2011.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hellstrom A, Hard AL, Engstrom E, Niklasson A, Andersson E, Smith L, et al. Early weight gain predicts retinopathy in preterm infants: new, simple, efficient approach to screening. Pediatrics. 2009;123:e638–45. doi: 10.1542/peds.2008-2697. [DOI] [PubMed] [Google Scholar]
  • 17.ELFIN trial investigators group. Enteral lactoferrin supplementation for very preterm infants: a randomised placebo-controlled trial. Lancet. 2019;393:423–33. doi: 10.1016/S0140-6736(18)32221-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Fauchère JC, Koller BM, Tschopp A, Dame C, Ruegger C, Bucher HU. Safety of early high-dose recombinant erythropoietin for neuroprotection in very preterm infants. J Pediatr. 2015;167:52–7. doi: 10.1016/j.jpeds.2015.02.052. [DOI] [PubMed] [Google Scholar]
  • 19.Sun H, Cheng R, Wang Z. Early vitamin a supplementation improves the outcome of retinopathy of prematurity in extremely preterm infants. Retina. 2020;40:1176–84. doi: 10.1097/IAE.0000000000002543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Cota F, Costa S, Giannantonio C, Purcaro V, Catenazzi P, Vento G. Lutein supplementation and retinopathy of prematurity: a meta-analysis. J Matern Fetal Neonatal Med. 2022;35:175–80. [DOI] [PubMed]
  • 21.Du Y, He Y, Wang YL, Zhou JG, Chen C. The efficacy and safety of inositol supplementation in preterm infants to prevent retinopathy of prematurity: a systematic review and meta-analysis. BMC Ophthalmol. 2019;19:135. doi: 10.1186/s12886-019-1140-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Cipriani A, Furukawa TA, Salanti G, Chaimani A, Atkinson LZ, Ogawa Y, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. Lancet. 2018;391:1357–66. doi: 10.1016/S0140-6736(17)32802-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.da Costa BR, Pereira TV, Saadat P, Rudnicki M, Iskander SM, Bodmer NS, et al. Effectiveness and safety of non-steroidal anti-inflammatory drugs and opioid treatment for knee and hip osteoarthritis: network meta-analysis. BMJ. 2021;375:n2321. doi: 10.1136/bmj.n2321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Anothaisintawee T, Attia J, Nickel JC, Thammakraisorn S, Numthavaj P, McEvoy M, et al. Management of chronic prostatitis/chronic pelvic pain syndrome: a systematic review and network meta-analysis. JAMA. 2011;305:78–86. doi: 10.1001/jama.2010.1913. [DOI] [PubMed] [Google Scholar]
  • 25.Huang J, Wen D, Wang Q, McAlinden C, Flitcroft I, Chen H, et al. Efficacy comparison of 16 interventions for myopia control in children: a network meta-analysis. Ophthalmology. 2016;123:697–708. doi: 10.1016/j.ophtha.2015.11.010. [DOI] [PubMed] [Google Scholar]
  • 26.Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8:336–41. doi: 10.1016/j.ijsu.2010.02.007. [DOI] [PubMed] [Google Scholar]
  • 27.Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, Cameron C, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med. 2015;162:777–84. doi: 10.7326/M14-2385. [DOI] [PubMed] [Google Scholar]
  • 28.Hellström A, Smith LE, Dammann O. Retinopathy of prematurity. Lancet. 2013;382:1445–57. doi: 10.1016/S0140-6736(13)60178-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Aranda JV, Qu J, Valencia GB, Beharry KD. Pharmacologic interventions for the prevention and treatment of retinopathy of prematurity. Semin Perinatol. 2019;43:360–6. doi: 10.1053/j.semperi.2019.05.009. [DOI] [PubMed] [Google Scholar]
  • 30.Filippi L, Dal Monte M. A safety review of drugs used for the treatment of retinopathy of prematurity. Expert Opin Drug Saf. 2020;19:1409–18. doi: 10.1080/14740338.2020.1826927. [DOI] [PubMed] [Google Scholar]
  • 31.Carlo WA, Finer NN, Walsh MC, Rich W, Gantz MG, Laptook AR, et al. Target ranges of oxygen saturation in extremely preterm infants. N Engl J Med. 2010;362:1959–69. doi: 10.1056/NEJMoa0911781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Australia B-I, United Kingdom Collaborative G, Tarnow-Mordi W, Stenson B, Kirby A, Juszczak E, et al. Outcomes of two trials of oxygen-saturation targets in preterm infants. N Engl J Med. 2016;374:749–60. doi: 10.1056/NEJMoa1514212. [DOI] [PubMed] [Google Scholar]
  • 33.Riley RD, Moons KGM, Snell KIE, Ensor J, Hooft L, Altman DG, et al. A guide to systematic review and meta-analysis of prognostic factor studies. BMJ (Clin Res ed) 2019;364:k4597. doi: 10.1136/bmj.k4597. [DOI] [PubMed] [Google Scholar]
  • 34.Cumpston M, Li T, Page MJ, Chandler J, Welch VA, Higgins JP, et al. Updated guidance for trusted systematic reviews: a new edition of the Cochrane Handbook for Systematic Reviews of Interventions. Cochrane Database Syst Rev. 2019;10:ED000142. doi: 10.1002/14651858.ED000142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ades AE, Sculpher M, Sutton A, Abrams K, Cooper N, Welton N, et al. Bayesian methods for evidence synthesis in cost-effectiveness analysis. PharmacoEconomics. 2006;24:1–19. doi: 10.2165/00019053-200624010-00001. [DOI] [PubMed] [Google Scholar]
  • 36.Sutton A, Ades AE, Cooper N, Abrams K. Use of indirect and mixed treatment comparisons for technology assessment. PharmacoEconomics. 2008;26:753–67. doi: 10.2165/00019053-200826090-00006. [DOI] [PubMed] [Google Scholar]
  • 37.Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ (Clin Res ed) 2003;327:557–60. doi: 10.1136/bmj.327.7414.557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Liang JH, Li J, Wu RK, Li JY, Qian S, Jia RX, et al. Effectiveness comparisons of various psychosocial therapies for children and adolescents with depression: a Bayesian network meta-analysis. Eur Child Adolesc Psychiatry. 2021;30:685–97. doi: 10.1007/s00787-020-01492-w. [DOI] [PubMed] [Google Scholar]
  • 39.Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. Stat Med. 2010;29:932–44. doi: 10.1002/sim.3767. [DOI] [PubMed] [Google Scholar]
  • 40.Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Consistency and inconsistency in network meta-analysis: concepts and models for multi-arm studies. Res Synth Methods. 2012;3:98–110. doi: 10.1002/jrsm.1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Liang JH, Lu L, Li JY, Qu XY, Li J, Qian S, et al. Contributions of modifiable risk factors to dementia incidence: a Bayesian network analysis. J Am Med Dir Assoc. 2020;21:1592–9.e1513. doi: 10.1016/j.jamda.2020.04.006. [DOI] [PubMed] [Google Scholar]
  • 42.Caldwell DM, Ades AE, Higgins JP. Simultaneous comparison of multiple treatments: combining direct and indirect evidence. BMJ (Clin Res ed) 2005;331:897–900. doi: 10.1136/bmj.331.7521.897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Mavridis D, Salanti G. A practical introduction to multivariate meta-analysis. Stat Methods Med Res. 2013;22:133–58. doi: 10.1177/0962280211432219. [DOI] [PubMed] [Google Scholar]
  • 44.Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple-treatment meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011;64:163–71. doi: 10.1016/j.jclinepi.2010.03.016. [DOI] [PubMed] [Google Scholar]
  • 45.Chaimani A, Salanti G. Using network meta-analysis to evaluate the existence of small-study effects in a network of interventions. Res Synth Methods. 2012;3:161–76. doi: 10.1002/jrsm.57. [DOI] [PubMed] [Google Scholar]
  • 46.Senjam SS, Chandra P. Retinopathy of prematurity: addressing the emerging burden in developing countries. J Fam Med Prim Care. 2020;9:2600–5. doi: 10.4103/jfmpc.jfmpc_110_20. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Mactier H, Weaver LT. Vitamin A and preterm infants: what we know, what we don’t know, and what we need to know. Arch Dis Child Fetal Neonatal Ed. 2005;90:F103–8. doi: 10.1136/adc.2004.057547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.MacTier H, McCulloch DL, Hamilton R, Galloway P, Bradnam MS, Young D, et al. Vitamin A supplementation improves retinal function in infants at risk of retinopathy of prematurity. J Pediatr. 2012;160:954–9.e951. doi: 10.1016/j.jpeds.2011.12.013. [DOI] [PubMed] [Google Scholar]
  • 49.Garofoli F, Barilla D, Angelini M, Mazzucchelli I, De Silvestri A, Guagliano R, et al. Oral vitamin A supplementation for ROP prevention in VLBW preterm infants. Ital J Pediatr. 2020;46:77. doi: 10.1186/s13052-020-00837-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Ozkan H, Duman N, Kumral A, Kasap B, Ozer EA, Lebe B, et al. Inhibition of vascular endothelial growth factor-induced retinal neovascularization by retinoic acid in experimental retinopathy of prematurity. Physiol Res. 2006;55:267–75. doi: 10.33549/physiolres.930754. [DOI] [PubMed] [Google Scholar]
  • 51.Tsai AS, Chou HD, Ling XC, Al-Khaled T, Valikodath N, Cole E, et al. Assessment and management of retinopathy of prematurity in the era of anti-vascular endothelial growth factor (VEGF) Prog Retin Eye Res. 2021;88:101018. doi: 10.1016/j.preteyeres.2021.101018. [DOI] [PubMed] [Google Scholar]
  • 52.Agostoni C, Buonocore G, Carnielli VP, De Curtis M, Darmaun D, Decsi T, et al. Enteral nutrient supply for preterm infants: commentary from the European Society of Paediatric Gastroenterology, Hepatology and Nutrition Committee on Nutrition. J Pediatr Gastroenterol Nutr. 2010;50:85–91. doi: 10.1097/MPG.0b013e3181adaee0. [DOI] [PubMed] [Google Scholar]
  • 53.Landman J, Sive A, Heese HD, Van der Elst C, Sacks R. Comparison of enteral and intramuscular vitamin A supplementation in preterm infants. Early Hum Dev. 1992;30:163–70. doi: 10.1016/0378-3782(92)90143-5. [DOI] [PubMed] [Google Scholar]
  • 54.Darlow BA, Graham PJ. Vitamin A supplementation for preventing morbidity and mortality in very low birthweight infants. Cochrane Database Syst Rev. 2000; CD000501. [DOI] [PubMed]
  • 55.Zhang S, Zhou R, Li B, Li H, Wang Y, Gu X, et al. Caffeine preferentially protects against oxygen-induced retinopathy. FASEB J. 2017;31:3334–48. doi: 10.1096/fj.201601285R. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Schmidt B, Roberts RS, Davis P, Doyle LW, Barrington KJ, Ohlsson A, et al. Long-term effects of caffeine therapy for apnea of prematurity. N Engl J Med. 2007;357:1893–902. doi: 10.1056/NEJMoa073679. [DOI] [PubMed] [Google Scholar]
  • 57.Hussein MA, Coats DK, Khan H, Paysse EA, Steinkuller PG, Kong L, et al. Evaluating the association of autonomic drug use to the development and severity of retinopathy of prematurity. J AAPOS. 2014;18:332–7. doi: 10.1016/j.jaapos.2014.02.012. [DOI] [PubMed] [Google Scholar]
  • 58.Davis PG, Schmidt B, Roberts RS, Doyle LW, Asztalos E, Haslam R, et al. Caffeine for Apnea of Prematurity trial: benefits may vary in subgroups. J Pediatr. 2010;156:382–7. doi: 10.1016/j.jpeds.2009.09.069. [DOI] [PubMed] [Google Scholar]
  • 59.Petrucci R, Zollo G, Curulli A, Marrosu G. A new insight into the oxidative mechanism of caffeine and related methylxanthines in aprotic medium: may caffeine be really considered as an antioxidant? Biochim Biophys Acta Gen Subj. 2018;1862:1781–9. doi: 10.1016/j.bbagen.2018.05.011. [DOI] [PubMed] [Google Scholar]
  • 60.Liu R, Gang L, Shen X, Xu H, Wu F, Sheng L. Binding characteristics and superimposed antioxidant properties of caffeine combined with superoxide dismutase. ACS Omega. 2019;4:17417–24. doi: 10.1021/acsomega.9b02205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Davis PG. When to start and stop caffeine and why respiratory status matters. Semin Fetal Neonatal Med. 2020;25:101175. doi: 10.1016/j.siny.2020.101175. [DOI] [PubMed] [Google Scholar]
  • 62.Bahji A, Pierce M, Wong J, Roberge JN, Ortega I, Patten S. Comparative efficacy and acceptability of psychotherapies for self-harm and suicidal behavior among children and adolescents a systematic review and network meta-analysis. JAMA Netw Open. 2021;4:e216614. doi: 10.1001/jamanetworkopen.2021.6614. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Materials (8.4MB, pdf)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Eye are provided here courtesy of Nature Publishing Group

RESOURCES