Respiratory severity score as a predictive factor for severe bronchopulmonary dysplasia or death in extremely preterm infants

Bronchopulmonary dysplasia (BPD) is a serious pulmonary condition and one of the most common causes of mortality and morbidity in preterm infants. Although numerous studies have been performed to identify risk factors of BPD to develop preventative therapies, the exact pathogenesis and predictors for BPD remain unclear [1]. In addition, the incidence of BPD has increased over the past 2 to 3 decades despite many advances in neonatal care for preventing BPD, such as gentle ventilation strategies, antenatal steroids, surfactants, and postnatal steroids [2‐5]. Because infants with severe BPD require prolonged mechanical ventilation and are associated with the long-lasting burdens of pulmonary and neurodevelopmental sequelae [6‐11], research for identifying predictive factors for the earlier diagnosis of BPD severity is valuable; the accurate detection of severe BPD at an early stage may enable the initiation of therapies when they may be more effective and may avoid unnecessary therapies, such as corticosteroids, thereby reducing their potential hazards [12].

The respiratory severity score (RSS) is a simplified severity score consisting of the mean airway pressure (MAP) multiplied by the fraction of inspired oxygen (FiO₂). The RSS has been used in place of the oxygen index (OI) to assess the need for respiratory support in infants requiring assisted ventilation because the lack of arterial access in infants precludes monitoring the partial pressure of arterial oxygen (PaO₂) [13]. A previous study revealed a strong association between the RSS and OI at saturation levels of arterial oxygen (SaO₂) between 88 and 94% in intubated infants [14]. Recently, Malkar et al. demonstrated that an RSS ≥6 on day 30 of life was associated with increased mortality and mechanical ventilation duration in premature infants supported by mechanical ventilation for more than 30 days [15]. The study also suggested the possible usefulness of the RSS as a predictive factor for outcomes in neonatal respiratory conditions.

The objective of the current study was to evaluate the correlation between the RSS and subsequent severe BPD or death.

This was a retrospective study conducted in the neonatal intensive care unit at Seoul National University Children’s Hospital between January 2010 and December 2014. Approval for this study was obtained from the Seoul National University Hospital Institutional Review Board.

Recently, several trials have focused on the prevention of BPD in very preterm infants, using methods such as steroid administration, inhaled nitric oxide, and stem cell therapy [24‐30]. Although those therapies might be beneficial for reducing the incidence of BPD, a lack of evidence exists regarding their safety and effectiveness. Hence, research on preventing BPD must focus on high-risk populations for severe BPD by assessing the balance between potential benefits and harms.

Many clinical models have been developed to predict the risk of BPD. The Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network has developed a predictive model for BPD that is available as a web-based tool to provide estimates of BPD based on demographic variables and respiratory support by postnatal day [31]. This tool might be helpful to identify patients who are at a higher risk for BPD and are most likely to benefit from postnatal preventive treatment. However, Asian race was not considered in this model. Later, Onland et al. reviewed 26 clinical prediction models for BPD in 2013 and concluded that most existing clinical prediction models are poor to moderate predictors of BPD [32]. None of the previous BPD prediction models have been widely adopted.

In the present study, we focused on using the RSS to predict severe BPD or death. RSSs at postnatal days 14, 21, and 28 were significantly associated with death or severe BPD after adjusting for other BPD-related factors. In addition, the RSS could provide an estimate of severe BPD or death at a relatively early stage. At the aspect for prediction, RSS at postnatal day 14 might provide early identification of infants at high risk for BPD or death whilereas RSS at postnatal day 28 might have less meaningful in practice. RSS at postnatal day 21 was the most highly associated with severe BPD or death.

The RSS has been used in several major randomized control trials as a surrogate for the OI in assessing the severity of respiratory illness [25, 33]. The OI is one of well-known respiratory indices for quantifying the severity of oxygenation failure [34, 35]. However, the OI requires an indwelling arterial catheter, which in turn is associated with multiple complications including thromboembolism and ischemic injury. In addition, the OI cannot reflect real-time pulmonary status and needs frequent blood samplings which lead to severe anemia in preterm infants [36]. In current clinical practice, arterial blood sampling is much less common than in the past. As alternative non-invasive methods of assessing the severity of respiratory failure, oxygen saturation index calculated by substituting SpO₂ for PaO₂ showed a good correlation with OI [37]. As a strong association between the RSS and the OI in preterm infants under mechanical ventilation at oxygen saturation between 88 and 94% was revealed [14], the RSS can be used instead of the OI when preterm infants maintain at targeted saturation. The RSS can easily determine the need for and assess the response to therapeutic interventions in infants requiring mechanical ventilation.

In this study, we included only patients who required intubated ventilatory support. The intubation rate during the first day of life was 92.6%, which was similar to the value of 89% reported by the Neonatal Research Network, and the moderate to severe BPD rate was 47%, which was similar to or slightly higher than the rates of other studies. We did not use any minimally invasive surfactant therapy techniques during the study period. Although that might have increased the risk of intubation for surfactant instillation, we initially considered the application of PEEP in the delivery room according to the guidelines of the neonatal resuscitation program, as well as the selective use of surfactants. The association between a longer duration of mechanical ventilation and chronic lung disease is well known. Avoiding intubation and using non-invasive ventilator support are necessary to minimize exposure to mechanical ventilation. However, in the Surfactant, Positive Pressure, and Oxygenation Randomized Trial (SUPPORT), 83% of extremely low birth weight infants initially assigned to non-invasive support required endotracheal intubation and mechanical ventilator care [38]. Morley CJ et al. also showed that among preterm infants 25 to 28 weeks in gestational age who had adequate respiratory effort and were assigned to non-invasive support, almost half required endotracheal intubation and mechanical ventilation [39]. In practice, preterm infants with younger gestational ages still require invasive mechanical ventilatory support. The present study showed that the RSS can be used for one of predictors of severity of BPD in these infants who require intubation and mechanical ventilator support. However, there are some limitations that infants who require intubation because of other causes instead of lung conditions, such as airway problems or apnea of prematurity, were included in the study. In addition, RSS can be highly affected on ventilatory support strategies and need to be standardized for validation. Recently, several new modalities for non-invasive respiratory support, such as non-invasive intermittent positive pressure ventilation, and non-invasive neurally adjusted ventilatory assist, have been developed. Those new techniques might provide more sophisticated respiratory support without intubation to extremely preterm infants in near future. Therefore, further investigation would be important to determine the effect of various non-invasive techniques on the outcomes of the present study.

Additionally, because this analysis was retrospective and based on experiences within a single unit, further research with a larger prospective cohort study is necessary to validate the usefulness of the RSS for predicting the outcome at a postmenstrual age of 40 weeks. RSS can be used one of covariates for developing a better clinical model to predict the risk of severe BPD or death.

In conclusion, the present study demonstrated that there were strong correlations between RSSs at postnatal days 14, 21, and 28 and subsequent severe BPD or death. Furthermore, an RSS ≥ 3.0 at postnatal day 14 and an RSS ≥ 3.6 at postnatal day 21 were reliable values for predicting severe BPD or death. The RSS could provide a simple estimate of severe BPD or death at various postnatal ages and might be helpful for determining the proper time for early interventions.