Aerosol therapy in relation to retinopathy of prematurity in mechanically ventilated preterm infants

Aerosol inhalation has been widely used to manage pulmonary diseases of children and adults. However, the distribution of particle deposition in the airway is different between infants and adults [1]. Recently, inhalation therapy is increasingly administrated to preterm infants. Several inhalants are used for mechanically ventilated infants, such as bronchodilators, corticosteroids, xanthine derivatives, mucoactive agents, and antibiotics. Inhaled bronchodilators help relieve airway constriction and improve lung compliance in premature infants with respiratory distress [2‐4] and bronchopulmonary dysplasia (BPD) [5, 6]. The frequency and treatment duration of inhaled bronchodilators markedly vary because of no evidence-based guidelines [7]. In addition to bronchodilators, inhaled corticosteroids and xanthine derivatives have replaced systemic ones for managing BPD in preterm infants [8, 9]. Furthermore, the use of inhaled mucoactive medications makes secretions easier to transport and increases the efficiency of cough or mucus clearance, since secretion clearance is hampered by weakness and restrictive lungs [10]. However, the role of mucoactive agents in pulmonary critical care for mechanically ventilated preterm infants is not yet clear [11].

Retinopathy of prematurity (ROP) is a common disease of premature infants [12]. Exposing premature infants to supplemental oxygen is an important risk factor for developing ROP [13, 14]. Increased duration and number of hyperoxia events may increase the risk of ROP [15, 16]. Therefore, consensus guideline for resuscitation of preterm infants has been revised to deliver a low fraction of inspired oxygen [17].

Excess oxygen may be delivered to premature infants during inhalation therapy. However, any correlation of ROP with inhalation therapy has never been determined. Thus in the current study, we explored whether excess oxygen via inhalation therapy can cause ROP in extremely premature infants. The findings of this study provide clinical implications for therapeutic strategies to avoid the development of ROP.

In total, 205 preterm infants were eligible for enrollment, including 54 infants with a GA of 24~25 weeks, 69 infants with a GA of 26~27 weeks, and 82 infants with a GA of 28~29 weeks. The incidence of ROP was 55.6% in infants with a GA of 24~25 weeks, 44.9% in those with a GA of 26~27 weeks, and 14.6% in those with a GA of 28~29 weeks (Fig. 1).

There were 154 infants (75.1%) who received aerosol therapy via an endotracheal tube. Drugs used for aerosol therapy included bronchodilators (ipratropium bromide, fenoterol, and salbutamol) and mucoactive agents (acetylcysteine and sodium bicarbonate). The most commonly used inhalation drug was acetylcysteine (99/154 = 64.3%). The use for bronchodilators and mucoactive agents was to decrease airway resistance and to promote mucin breakdown, respectively [20].

Demographic characteristics of enrolled infants are shown in Table 1. There were significant differences in sex and GA between infants with and without aerosol therapy. Males were more common among infants with aerosol therapy. Furthermore, the incidence of aerosol therapy was inversely proportional to GA: 85.2% at GA of 24~25 weeks, 78.2% at GA of 26~27 weeks, and 65.9% at GA of 28~29 weeks. Overall, the mortality rate after 1 month of age was 5.4%. There was no significant difference in the mortality between the aerosol and control groups. All deaths in both groups were related with sepsis.

Table 2 illustrates the association between ROP and settings of mechanical ventilation. There was a significant correlation of ROP with the following factors – aerosol therapy, high frequency ventilation (HFV), inhaled nitric oxide (iNO), and the duration of intubation. In specific, ROP was more common in infants with aerosol therapy, iNO, use of HFV longer than 1 week, and intubation period more than 6 weeks. The relationship between ROP and aerosol therapy by three categories of GA is shown in Table 3. There was a trend of increasing ROP among infants with aerosol therapy in all categories of GA.

A risk assessment of ROP was conducted using a multivariate logistic regression model (Table 4). Adjusted confounders included sex, aerosol therapy, GA, iNO, HFV, and intubation period. The regression analysis showed that inhalation therapy carried a greater risk of ROP. In addition, infants with a GA of 24~27 weeks were at greater risk of ROP than those with a GA of 28~29 weeks. Other factors – including sex, HFV, iNO, and intubation period – did not carry significant risk.

This study depicts the association between aerosol therapy and ROP among mechanically ventilated preterm infants. To our best knowledge, this is the first study to investigate the potential impact of aerosol therapy on the risk of ROP. Our findings suggest that aerosol administration of therapeutic drugs via an endotracheal tube in combination with pure oxygen may result in harmful effects of ROP in infants with a GA of 24~29 weeks.

ROP is a multifactorial disease. One of the most significant risk factors for ROP is a low GA, as shown in our and other studies [12]. Immature vascularization induces an increased susceptibility of the retina to oxidative damage. Therefore, our study enrolled infants with a GA of 24~29 weeks. Furthermore, our study aimed to clarify the possible impact of aerosol therapy on ROP. Thus, we only investigated infants who received aerosol therapy before the development of ROP.

In this study, inhalants were delivered with 100% oxygen via a nebulizer. It is well documented that high oxygen saturation can increase the risk of ROP [21]. However, a lower target range of oxygenation may result in an increase in mortality [22]. Therefore, adequate oxygen during aerosol therapy is mandatory to maximize benefits and minimize harms. To reduce the potential risk of ROP, several factors should be considered while conducting aerosol therapy. Our data suggest that an oxygen blender or metered dose inhaler-spacer device may be needed for inhalation therapy to reduce excess oxygen exposure [23]. In addition, using the mechanical ventilator or a mesh nebulizer to produce the aerosol can decrease the fraction of oxygen concentration. Furthermore, the total time outside the target saturation range is associated with the development of ROP [16]. Thus, shortening the duration or frequency of high-oxygen inhalation may be helpful in decreasing the risk of ROP.

The role of oxygen in the development of ROP is complex. Many reports showed a reduction in ROP with supplemental oxygen in a later course of treatment after ROP had already developed [24, 25]. Accordingly, aerosol therapy with pure oxygen may be fine when an infant already has ROP.

In this study, we did not focus on the medications used in aerosol therapy. One would question whether medicines such as acetylcysteine may have an impact on the development of ROP. Our study found no significant difference in ROP between inhalants with and without acetylcysteine (data not shown). Furthermore, it has been shown that acetylcysteine may have a protective effect against ROP [26]. Thus we believe that ROP is not simply modulated by inhaled medicines, although the potential effects of inhalation drugs cannot be completely ruled out.

In our study, the univariate analyses showed ROP was more common in infants with HFV, iNO, and prolonged intubation. However, the multivariate model did not find a significant correlation of ROP with these ventilation settings, which are in accordance with previous reports of meta-analysis illustrating that HFV and iNO did not increase the risk of ROP in preterm infants [27, 28].

There are limitations to our study. First, this study was a retrospective design. Randomized controlled trials are needed to determine the causal effect of aerosol therapy on ROP. Second, the effectiveness of aerosol therapy for premature neonates was not evidence-based [21]. No clear long-term benefits have been demonstrated yet. Third, we did not investigate the potential impact of delivery devices. Nevertheless, the delivery devices were not changed throughout the study period.

Aerosol therapy is increasingly administered to premature infants during the first hospital month. Possible adverse effects of aerosol therapy need to be elucidated. Our study implies an increased risk of ROP with aerosol medicines administrated using pure oxygen. Appropriate oxygen delivery may play a role in reducing the progression of ROP. Clinicians should use great caution when conducting aerosol therapy with excess oxygen in preterm infants.