Impact of early caffeine administration on respiratory outcomes in very preterm infants initially receiving invasive mechanical ventilation

Abstract

Objective

The guidelines recommend early caffeine administration for preterm infants requiring non-invasive mechanical ventilation since earlier treatment is associated with better outcomes. The objective was to evaluate the impact of early caffeine therapy (within 24 hours after birth) on respiratory outcomes in very preterm infants who were initially receiving invasive mechanical ventilation.

Methods

This was an observation cohort study from 1 January 2018 to 31 December 2022 based on a database that was prospectively collected and maintained. Infants who initially received invasive mechanical ventilation were divided into two groups based on the timing of caffeine within the first 24 hours after birth (early) and within 48 hours of birth or later (late). Generalised linear mixed models with a random effect model for the centre were used to assess the impact of different caffeine initiation times on neonatal outcomes.

Results

Among the cohort of 9880 infants born at <32 weeks gestation, 2381 were eligible for this study (early 1758 (73.8%) and late 623 (26.2%)). For infants born at more than 28 weeks of gestation, the adjusted generalised linear mixed model showed that the duration of invasive mechanical ventilation was 1.34 (95% CI −2.40 to –0.27) days shorter and the incidence of moderate-to-severe bronchopulmonary dysplasia (BPD) was lower (adjusted OR 0.63; 95% CI 0.41 to 0.96) in the early caffeine group compared with the late caffeine group.

Conclusion

In very preterm infants who initially receive invasive mechanical ventilation, early administration of caffeine within 24 hours after birth can shorten the duration of invasive mechanical ventilation, reduce the incidence of moderate-to-severe BPD and improve respiratory outcomes. The very early initiation of caffeine treatment does not appear to be associated with any adverse outcomes.

Trial registration number

ChiCTR1900025234.

Keywords: Pulmonary Disease, Chronic Obstructive; Drug induced Lung Disease

Introduction

With the continuous improvement of neonatal care, the survival rate of very premature infants has significantly increased. However, due to the presence of underdeveloped lungs, bronchopulmonary dysplasia (BPD) remains a poor outcome in these very premature infants.^1 2^ Our previous study showed that the median duration of invasive mechanical ventilation was 26 (17–41) days in very preterm infants, and a longer duration of invasive mechanical ventilation was an independent risk factor for the development of moderate-to-severe BPD.^3^ The Caffeine for Apnoea of Prematurity (CAP) study confirmed that caffeine promoted successful extubation, reduced the incidence of BPD and improved neurodevelopmental outcomes in preterm infants.^4 5^

With the publication of the CAP study results, clinical practice has changed in many centres, and the use of caffeine for prophylactic purposes has increased significantly (ie, administration is started very soon after birth or before a diagnosis of apnoea of prematurity is made).^6 7^ Subgroup analysis of the CAP trial showed that early use of caffeine within 72 hours after birth decreased the incidence of BPD.^6^ The European Consensus Guidelines recommend considering early administration of caffeine for infants at high risk of mechanical ventilation, such as premature infants receiving non-invasive mechanical ventilation.^8^ While international guidelines favour non-invasive respiratory support, the majority of preterm infants still inevitably require invasive mechanical ventilation after admission.^9^ Therefore, can the early use of caffeine improve respiratory outcomes in very preterm infants who initially receive invasive mechanical ventilation?

We hypothesised that the early use of caffeine would provide similar benefits without evidence of short-term harm in very preterm infants who initially received invasive mechanical ventilation. The aim of this study, which was based on a prospective observational cohort study, was to determine whether early caffeine administration (within 24 hours after birth) contributes to early withdrawal and improved respiratory outcomes in very preterm infants who initially received initial invasive mechanical ventilation.

Methods

Study design and population

This was a multicentre observation cohort study based on a database that was prospectively collected and maintained. The study included all preterm neonates who were born at less than 32 weeks of gestation and admitted to 47 participating NICUs in the Sino-Northern Neonatal Network (SNN) (http://www.snn-med.com/) from 1 January 2018 to 31 December 2022. Neonates who were born outside a tertiary-level NICU, who initially received non-invasive positive pressure ventilation, who had congenital anomalies, who did not receive caffeine therapy, who were missing data and who died before 24 hours after birth were excluded from the study.

Data sources

The data were collected as part of the SNN routine data collection.^10^ Prospectively, maternal pregnancy and delivery data were collected soon after birth, and infant data were collected until death and hospital discharge/transfer. Patients were divided into the following two groups according to the time of caffeine early initiation (neonates who received caffeine within the first 24 hours after birth) and late initiation (neonates who received caffeine starting ≥24 hours after birth). The usual practice in each centre is to administer a loading dose of 20 mg/kg of caffeine base, with a daily maintenance dose of 5–10 mg/kg initiated 24 hours after the loading dose.

Outcomes

The primary outcomes of this study were the duration of invasive mechanical ventilation and the incidence of moderate-to-severe BPD. The secondary outcomes were the duration of non-invasive respiratory support, the duration of oxygen supplementation, pulmonary haemorrhage, failure to wean, necrotising enterocolitis (NEC) and death before discharge.

Definitions

Estimates of gestational age were made, in descending order of preference, according to ultrasonography, the last menstrual period or the New Ballard score.^11^ The outcomes included early-onset sepsis (≤72 hours) (defined by cultures positive for bacteria or fungi and antibiotic therapy duration ≥5 days or the intention to treat but death <5 days),^12^ NEC (stages 2–3),^13^ severe intraventricular haemorrhage^13^ and moderate-to-severe BPD (the definition of moderate to severe BPD was based on previously published criteria by Jobe and Bancalari).^14^ The diagnosis of PDA was made clinically (a wide pulse pressure, prominent precordial pulsations, bounding pulses and a continuous systolic murmur at the second left parasternal area and/or signs of heart failure), with or without echocardiography.

The duration of invasive mechanical ventilation was defined as the total number of days the infant required mechanical ventilation support. The duration of non-invasive respiratory support was defined as the total number of days during which the infant received nasal intermittent positive pressure ventilation, bilevel positive airway pressure, continuous positive airway pressure or high-flow nasal cannula therapy. The duration of oxygen supplementation was defined as the total number of days the infant received supplemental oxygen. Reintubation within 72 hours was defined as extubation failure.^15^

Statistical analysis

All the data were analysed by using SPSS V.26.0 software. Continuous variables are presented as medians (IQRs) while categorical variables are expressed as frequencies (percentages). The Mann-Whitney U test or χ^2^ test was used to compare the differences between the two groups according to the distribution of the data. To control for these potential confounding factors, participants in the two groups were matched 1 using propensity score matching based on statistically significant baseline characteristics (gestational age, birth weight and surfactant use). Any baseline characteristics that did not exhibit appropriate balance between the two groups were subsequently included in generalised linear mixed models. Generalised linear mixed models with a random effect model for the centre were used to assess the impact of different caffeine initiation times on neonatal outcomes. In this study, a p<0.05 was considered to indicate statistical significance.

Results

Baseline characteristics

A total of 9880 infants born at <32 weeks gestation were admitted to participating NICUs during the study. Based on the inclusion and exclusion criteria, 2381 very preterm infants with initial invasive mechanical ventilation were ultimately included in this study; 1758 (73.8%) were included in the early initiation group and 623 (26.2%) were included in the late initiation group (figure 1).

Figure 1 Flow diagram of the study participants.

The baseline characteristics of the cohort are described in table 1. After propensity score matching, the majority of covariates were balanced between the two groups in both subcohorts. Compared with the late caffeine initiation group, the early initiation group (infants born at <28 weeks gestation) had a greater percentage of males and a somewhat lower Apgar score at 5 min (table 1). For infants born at 28^+0^–31^+6^ weeks gestation in the early initiation group, there were statistically significant differences in male sex and hsPDA (table 1).

Primary outcomes

Univariate analysis of infants born at 28^+0^–31^+6^ weeks of gestation revealed that the duration of invasive mechanical ventilation (early median 4 days (IQR 2.5–7.9) vs late 6 (3.7–10.8) days, p<0.05) was shorter, and the risk of moderate-to-severe BPD (10.7% vs 16.2%; p=0.013) was lower in the infants for whom caffeine was given early than in those for whom it was given late (table 2). The adjusted generalised linear mixed model showed that the duration of invasive mechanical ventilation was 1.34 (95% CI −2.40 to –0.27; p=0.014) days shorter and the incidence of moderate-to-severe BPD was lower (adjusted OR, AOR 0.63; 95% CI 0.41 to 0.96) in the early caffeine group compared with the late caffeine group (table 2). For infants born at <28 weeks, univariate analysis revealed that the early caffeine group had a shorter duration of invasive mechanical ventilation (early median 8 days (IQR 4.2–21.3) vs late 14 (6.7–25.8) days; p=0.108) and a lower incidence of moderate-to-severe BPD (32.1% vs 34.9%; p=0.662) than the late caffeine group did, although the differences were not statistically significant (table 2).

Secondary outcomes

For infants born at 28^+0^–31^+6^ weeks gestation, the adjusted generalised linear mixed model showed that the duration of oxygen supplementation was 5.38 lower (95% CI −8.10 to –2.66; p<0.001) in the early caffeine group compared with the late caffeine group (table 2). For infants born at <28 weeks, the adjusted generalised linear mixed model showed that the early caffeine group had a lower incidence of pulmonary haemorrhage (AOR 0.39; 95% CI 0.17 to 0.92) than the late caffeine group (table 2). Within both gestational age ranges, there were no statistically significant differences in the duration of non-invasive respiratory support, failure to wean or death before discharge between the two groups (p>0.05).

Discussion

Early caffeine administration is recommended for infants at high risk of needing mechanical ventilation, such as premature infants receiving non-invasive mechanical ventilation. For infants who initially receive invasive mechanical ventilation, whether caffeine should be applied early is still controversial. This was a large, multicentre, population-based cohort study conducted to investigate the association between early caffeine administration and respiratory outcomes among very preterm infants who initially received invasive mechanical ventilation in China. Despite the lack of evidence from randomised clinical trials evaluating the very early use of caffeine in preterm neonates, most very preterm neonates in Chinese neonatal intensive care units (NICUs) included in this study received caffeine within the first 24 hours after birth. In our study, the rate of early caffeine administration was 73.8% among very preterm infants. Borszewska-Kornacka et al^16^ reported that the proportion of very preterm infants treated early with caffeine citrate (initial dose on the first day of life) was 80.1%. In this study, we found no serious adverse effects associated with early (prophylactic) caffeine initiation in preterm infants (born at <32 weeks gestation) compared with those who received caffeine later.

Multivariate analysis revealed that among infants born at 28^+0^–31^+6^ weeks gestation, early administration of caffeine was associated with a shorter duration of invasive mechanical ventilation. A randomised double-blind clinical trial showed that early caffeine initiation reduces reintubation.^17^ Infants who received caffeine early appeared to be ready for extubation earlier, and therefore, had a reduced duration of mechanical ventilation. A retrospective cohort study showed that the median duration of mechanical ventilation was reduced by 3 days in the early caffeine treatment group.^4^ Like in our study, Lodha et al^18^ reported that early caffeine application reduced the median duration of mechanical ventilation by 2 days. Early application of caffeine could improve the contraction and function of the diaphragm in very preterm infants, stimulate regular breathing and increase the sensitivity of the infant to carbon dioxide.^19 20^ Henderson-Smart and Davis^21^ reported that caffeine increases the chances of successful extubation of preterm infants within 1 week of birth. The results of our real-world study suggested that early caffeine administration within 24 hours after birth maximises the potential benefits among very preterm infants who initially receive invasive mechanical ventilation.

Our study also revealed a significant reduction in the incidence of BPD in infants with a gestational age of 28^+0^–31^+6^ weeks in the early caffeine group compared with the late caffeine group. Previous studies have shown that prolonged mechanical ventilation is an independent risk factor for BPD, and early caffeine initiation can reduce the duration of mechanical ventilation.^9 22^ A retrospective cohort study conducted by the CNN showed that early caffeine administration reduced the incidence of BPD in very preterm infants.^18^ Another retrospective cohort study also showed that early caffeine administration was associated with a reduced incidence of BPD.^23^ These findings suggested that early administration of caffeine within 24 hours after birth has a greater benefit in improving the incidence of BPD than other early administration times. The decrease in the incidence of BPD in the early caffeine treatment group may be related to the mechanism of caffeine, in addition to the shorter duration of mechanical ventilation and oxygenation supplementation in the early caffeine treatment group. Early application of caffeine could reduce the expression levels of chemokines and leucocytes in oxidised lung tissue, weaken the apoptosis mediated by inflammatory factors and reduce the release of inflammatory mediators to improve lung function, thereby preventing the development of BPD in preterm infants.^24^ This study provides real evidence for the early use of caffeine in premature infants who initially receive initial mechanical ventilation.

In extremely premature infants with a gestational age of less than 28 weeks, early caffeine administration also reduced the duration of invasive mechanical ventilation and the incidence of moderate-to-severe BPD. Although the difference was not statistically significant, there was still a trend towards improved respiratory outcomes. Respiratory maturity is lower in very preterm infants at a younger gestational age, and there may be many other factors that influence respiratory outcomes, such as genetic susceptibility,^25^ sepsis,^26^ patent ductus arteriosus,^27^ antibiotic application,^28^ admission hypothermia^29^ and lower socioeconomic status.^30^ Early caffeine administration did not dominate these factors, suggesting that early caffeine application alone did not result in significant improvement in respiratory outcomes in very preterm infants who were small for gestational age. In addition, the much smaller sample size and lower gestational age potentially impeded the ability to detect a significant difference.

This study has several limitations. First, these findings should be interpreted with caution because of the observational nature of our study. Second, due to time constraints, this study focused only on the short-term effects of caffeine and did not explore the long-term effects. Third, for infants with gestational ages less than 28 weeks, it is necessary to increase the sample size for better evidence. Future studies are planned to investigate the effects of early caffeine administration on long-term outcomes in preterm infants.

In conclusion, early caffeine therapy reduced the duration of invasive mechanical ventilation, decreased the incidence of moderate-to-severe BPD and improved respiratory outcomes in very preterm infants who initially received invasive mechanical ventilation. In clinical practice, a rigorous observational cohort study can also provide a reliable source of evidence for clinical practice when randomised controlled trials cannot be performed.