The effect of 3% and 6% hypertonic saline in viral bronchiolitis: a randomised controlled trial

Introduction

Acute viral bronchiolitis is a viral infection of the lower respiratory tract in infants [1, 2]. Common viruses that cause bronchiolitis include respiratory syncytial virus (RSV), human metapneumovirus, adenovirus, (para)influenza virus, rhinovirus and coronavirus [3]. Bronchiolitis is the leading cause of infant hospitalisation during the first year of life and impacts millions of infants worldwide [4]. The annual costs due to hospitalisation for bronchiolitis amount to more than $500 million in the USA alone [5]. So far, potentially therapeutic interventions, like nebulised corticosteroids, epinephrine, beta agonists, antibiotics and recombinant human deoxyribonuclease have been shown not to be effective [6–10]. Therefore, treatment of bronchiolitis remains supportive with supplemental oxygen, maintaining fluid balance and mechanical ventilation, if required [11].

Airway oedema and mucus plugging are the predominant pathological features in infants with acute viral bronchiolitis [12]. Nebulised hypertonic saline (HS), using sodium chloride, may have positive effects on bronchiolitis, as it may reduce submucosal swelling, decrease inflammatory mediators and the viscosity of mucus and improve mucociliairy clearance [13–16]. In vitro, HS improves mucus rheologic properties (elasticity and viscosity) and accelerates mucus transport rates [17]. Studies in the past decade have suggested that nebulised mildly HS reduced the duration of stay in hospital for infants with mild-to-moderately severe bronchiolitis [4, 18–25]. All but one of these studies included small numbers of patients from single centres and used poorly defined outcomes; study groups were heterogeneous and some were not clearly defined [4, 26–28]. Studies in cystic fibrosis patients have shown that nebulised HS in a concentration of 5–7% is more effective than isotonic saline, and this effect appeared to be dose dependent [29–34]. In infants aged 4 months and older with cystic fibrosis, HS was proven to be safe [35]. Therefore, we performed a randomised, controlled trial of different concentrations of nebulised HS in infants hospitalised for viral bronchiolitis. We hypothesised that nebulisation with HS would dose-dependently reduce the length of hospital stay and the duration of supportive therapy.

Results

Between November 2009 and May 2011, 292 previously healthy children with moderate-to-severe viral bronchiolitis were included in the study and randomised (fig 1). Prior to enrolment, most patients received nasal-decongestants: 21 received salbutamol (by spacer), 10 paracetamol, four amoxicillin, three omeprazol, three nystatine, and one received valproinic acid.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

Enrolment of patients to the study and the follow-up and analyses procedures.

Two patients were discharged due to rapid clinical amelioration and were, therefore, withdrawn from the study before the first dose of the trial medication was given. A total of 290 infants (median age 3.4 months, range 10 days to 23 months) were enrolled in the study and analysed in the intention-to-treat analysis. 43 (15%) patients were excluded for the per-protocol analysis and 18 (6.2%) patients did not complete the treatment because of adverse events. Of these, nine patients left due to coughing or agitation during and after nebulisation and nine others because of clinical deterioration, such as increased need for oxygen, effort of breathing or progressive airway obstruction. Withdrawal of informed consent led to the withdrawal of a further nine (3.1%) patients. Thirteen infants were excluded from the study due to protocol deviation. Three patients (1.2%) required mechanical ventilation at a PICU during the course of disease: one patient deteriorated 5 h after the first inhalation with the 6% HS; the clinical condition of one patient, randomised to the 6% group, worsened after 6 days of treatment and required intensive care admission; and one patient, receiving NS, had been respiratory insufficient on admission and this clinical condition did not change after receiving one dose of the study medication while awaiting transfer to the PICU (fig. 1). The number of withdrawals (per reason) did not significantly differ between intervention groups (p=0.47).

In total, 247 infants completed the study and were analysed in the per-protocol analysis; 84 received 3% HS, 83 received 6% HS and 80 received NS (fig. 1). Patient baseline characteristics did not significantly differ between the three groups (table 2). Results of the intention-to-treat and per-protocol analyses did not differ for baseline characteristics or for any primary or secondary outcomes (data not shown). Besides the study medication, patients received: nasal-decongestants n=161, paracetamol n=30, antibiotics n=11, salbutamol n=6, ibuprophen n=1, nystatine n=1, and ranitidine n=1. The use of nasal decongestant (xylomethazoline and sodium chloride) was equally distributed among the study groups (data not shown).

The median duration of hospital stay was 69 h in the 3% HS group, 70 h in the 6% HS group and 53 h in the NS group (p=0.29) (table 3). Differences in time to discharge between the three groups were not statistically significant (p=0.26) (fig. 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2–

The percentage of infants remaining in hospital displayed against time until discharge, according to intervention.

Supplemental oxygen was needed in 50 (59.5%) infants in the 3% HS group, 53 (63.9%) infants in the 6% HS group, and 51 (63.8%) infants in the 0.9% NS group (p=0.70), with a median duration of 54, 54 and 40 h, respectively, (p=0.14).

Additional tube feeding was required by 29 (34.5%) infants in the 3% HS group, 31 (37.3%) infants in the 6% HS group and 22 (27.5%) infants in the 0.9% NS group (p=0.39), with a median duration of 62, 52 and 54 h, respectively, (p=0.87) (table 3).

The median Wang score at discharge was 2.0 in the NS group and in the 3% HS group and 1.0 in the 6% HS group (p=0.53). The Wang score at discharge improved regardless of the treatment regimens (tables 3 and 4), without significant differences between the groups (p=0.80).

A substantial number of adverse effects were noted in all treatment groups (table 5). Except for cough, which occurred significantly more often in the HS groups (p=0.03), no differences were found between groups. Withdrawals because of adverse events did not differ between groups (p=0.59). Most adverse events were likely to be related to the underlying viral bronchiolitis.

Discussion

In this double-blind, randomised, multicentre trial we compared 3% and 6% HS with NS in children hospitalised with viral bronchiolitis. Neither the 3% HS nor the 6% HS reduced the duration of stay in hospital, the Wang clinical severity score at discharge, the duration of supplemental oxygen or tube feeding compared to NS. These results are in contrast with results published in the literature [18–22]. Pooled results from three small studies [18–20] demonstrated a reduction in the duration of hospitalisation by 0.9 days in favour of the 3% HS group when compared with the NS group, whereas Luo and colleagues [21, 22] found a reduction by 1.4 and 1.6 days, respectively, in their studies in mild-to-moderate and moderate-to-severe bronchiolitis. However, Al-Ansari et al. [24] compared nebulisation of 3% HS and 5% HS to NS in relatively mild bronchiolitis in infants staying at a short-stay unit and did not show an effect on duration of stay at hospital, which he described as affected by medical and social factors.

Although criteria for diagnosis of viral bronchiolitis were less well defined in most of these studies except in the study of Kuzik et al. [20], differences in study populations between our study and previous studies are unlikely to explain the differences between the different results. Our study population was comparable with respect to age, clinical severity, duration of illness before admission and predominantly RSV-positive infants. Additionally to the exclusion criterion of previous wheezing, atopic manifestations, such as eczema or food allergy, we tried to avoid inclusion of infants who might benefit from salbutamol. Indeed, a recent study showed that in children with bronchiolitis, the Wang score improved significantly more in atopic children than in nonatopic children when salbutamol was added to the medication [23], which supports our approach. Taking into account the possible rapid responders to a single dose of salbutamol [12], it could be argued that our study protocol led to the exclusion of rapid responders, possibly influencing the results. In total, 14% of all children screened for inclusion had to be excluded because of a positive reaction to salbutamol. However, even if these infants had been included and randomly assigned to the study groups, the number would have been too small to influence the results significantly.

A meta-analysis suggested an inverse relationship between severity of disease and effect of nebulised HS [4]. Hence, differences in severity of bronchiolitis between the studied populations could explain the different findings. However, this does not seem to have been the case, as at baseline the Wang scores in our study were 6.2 compared to 5.8 in the study accounting for half the weight of the meta-analysis [21], while the other studies of the meta analysis included patients with higher Wang scores [18–20]. Our study was not designed to examine subgroups with different disease severity, and further studies should elucidate if HS might be specifically useful in subgroups of infants with more severe disease.

We cannot rule out that differences in type of nebulisers and frequency of nebulisation played a role. We used the same nebuliser with a MMAD of 4 μm in all study patients, leading to a deposition mainly in the lower airways [38]. Potentially, a nebuliser with smaller MMAD and another lung deposition might be of influence [39]. The type of nebuliser and MMAD were not described in the other studies, therefore, subsequent comparison is not possible. Only two previous studies contained participants who had been nebulised more frequently than in our study [20, 22]. However, there was no kinetic data on the half-life of the effect of HS and frequent nebulisation is not advised [20]. Hence, it is unlikely that differences in type of nebulisation and frequency of nebulisation could explain our findings.

Most previously published studies were single centre studies. It is possible that differences like local habits of diagnosis and treatment may have influenced the results. The current study was a multicentre study, which could possibly reduce these factors. Randomisation was undertaken for each centre and, therefore, allowed an analysis per centre; but we found no significant differences between the centres.

The strengths of our study include the use of strict inclusion and discharge criteria. We used a strict, clinically relevant definition for time of discharge, which was not influenced by social or administrative factors [24]. Hence, we confined the length of stay in hospital to the time in hospital due to medical need. This might explain our relatively short durations for hospital stay (53, 69 and 70 h for NS, 3% HS and 6% HS, respectively), compared with 2.6–6.0 days for 3% HS and 3.0–7.4 days for NS in previous studies [18–22]. None of these studies took care of potentially influencing social and administrative factors; most studies only used physician-based clinical decisions.

A possible explanation why this study did not find a difference between the intervention groups and the control group could be that NS may not be a true placebo. However, the design of our study cannot rule out a positive effect of NS nebulisation in infants with viral bronchiolitis. This question could be addressed by a noninferiority trial comparing NS and HS. Physiologically, an adequate quantity of airway lining fluid is crucial for normal functioning of airway mucociliary clearance. Bronchiolitis is known to deplete water content of the airway lining fluid, causing a reduction of mucociliary clearance [12]. In vivo, the change in airway surface liquid and the improvement in mucus clearance after NS or HS inhalation is probably a direct effect of the total mass of sodium chloride added to the airway surface [13]. In ambulatory bronchiolitis patients, nebulisation with a high volume of NS was as effective as that of a lower volume of 3% HS [40]. So far, this has not been shown for more severe illnesses. Earlier studies that showed a positive effect of 3% HS also performed nebulisation three times a day, which resulted in the same total dose [18, 19, 21]. Consequently, although a higher dose of total sodium chloride may have influence, it is unlikely to explain the different study results. Our findings were supported by a recent study comparing NS to 7% HS and using an inhalation frequency of four times a day that did not show a positive effect of HS, even with a higher total dose of sodium chloride [41]. Ipek et al. [23] compared the effect of nebulised NS and HS with and without salbutamol in outpatients with mild bronchiolitis. All study groups showed a significant decrease in clinical severity score, without differences between the groups, suggesting that NS might have some therapeutic effect as well in non-atopic patients. This supports our findings that no significant differences could be found between the use of HS and NS in bronchiolitis patients.

Although no serious adverse effects were observed in the current study, a considerable number of children showed adverse effects (table 5), which largely may have been due to bronchiolitis itself. In 6.2% of all included patients, adverse effects were the reason for withdrawal. Moreover, coughing was more prevalent in both HS groups. This requires further attention, since previous studies did not include exact descriptions of adverse effects.

In conclusion, we showed that nebulised 3% HS or 6% HS did not reduce the duration of hospital stay, the Wang clinical severity score at discharge, the need for supplemental oxygen or the need for tube feeding in children aged up to 2 years who are hospitalised with viral bronchiolitis, when compared with the use of NS. Our results do not support the routine use of HS in infants with bronchiolitis.