Nasal continuous positive airway pressure (NCPAP) is a simple, low cost and non-invasive method of ventilating a sick newborn [1]. Bubble CPAP is the most commonly used modality for delivery of NCPAP [2,3]. Traditionally, short bi-nasal prongs have remained the standard of care for delivery of NCPAP. The limitations of delivering NCPAP with prongs include mechanical difficulties in maintaining the nasal prongs, poor tolerance of the infant to the apparatus, difficulties in positioning the neonate, columella injury and septal deformities [4-6].

Nasal masks are increasing being used for delivering CPAP in recent times due to their ease of application [7]. A randomized trial in neonates <31 weeks gestation comparing nasal mask with binasal prongs showed less intubation rate within 72 hours for the treatment of respiratory distress syndrome (RDS) or in post-extubation setting with nasal mask [8]. A recent randomized controlled trial (RCT) from India reported a 6% reduction in the oxygen requirement at 2 hours of CPAP initiation with nasal mask as compared to nasal prongs [9]. Nasal trauma has been reported with the use of both nasal masks and prongs and occurs equally with each interface [10,11]. There is need for more evidence before nasal masks can replace short binasal prongs. Our aim was to compare the effectiveness of these two modes of CPAP delivery in an Indian scenario using Bubble CPAP.

Intervention: Enrolled infants were randomized to receive either Nasal mask (group 1) or Nasal prongs (group 2) as a mode of NCAP delivery interface. Randomization was done using a computer generated randomization chart with sealed opaque, sequentially numbered envelopes. The physician on call opened sequentially numbered sealed opaque envelopes and randomized infants to respective groups. Access to envelopes was restricted to designated physicians. Blinding of intervention and outcome measurement was not feasible because of the nature of intervention

Infants in the Mask group were delivered NCPAP using Fisher and Paykel Infant Nasal Mask in small (BC800), medium (BC801) and large (BC802) sizes based on best estimate using the nasal mask scale provided by the company. Masks were connected to Fisher and Paykel ‘Bubble CPAP system’(BC151) using Fisher and Paykel ‘Flexi Trunk Midline Interface’ (BC191 - 70 mm) and appropriate sized Fisher and Paykel ‘Infant Bonnet’ depending on Head circumference (BC300 - small, BC303 - medium, BC306 – large). Infants in the Prong group were delivered NCPAP using appropriate sized Hudson RCI　Infant Nasal Prong CPAP cannula system (size 0 and 1). The prongs were connected to Fisher and Paykel ‘Bubble CPAP system’(BC151) directly using pins and rubber bands over appropriate sized bonnets provided with the Hudson Nasal prong CPAP cannula system.

PDA was diagnosed clinically and confirmed by echocardiography wherever possible. IVH was defined by bed side sonography using the Papile classification [12]. NEC was classified according to Bell’s classification, as modified by Kliegman and Walsh as stage II or greater [13]. ROP was defined according to the international classification of retinopathy of prematurity [14]. BPD was defined according to the National Institutes of Health consensus definition [15]. Full feeds were defined as feeds that reached 150 mL/kg per day and sustained for 3 consecutive days. All secondary outcomes were determined before discharge home from hospital unless stated otherwise.

On the basis of NICU data from previous two years before this study, around 40% of infants with a completed gestational age <34 weeks who received NCPAP as primary treatment of respiratory distress required intubation and invasive ventilation in our NICU. The interface on which infants started was not recorded but was likely to be prongs in the majority. We hypothesized that using nasal mask would reduce the need for intubation and mechanicical ventilation to 20% (an absolute reduction of 20%). With a two sided alpha error of 0.05 and beta error of 0.2 (power 80%), the estimated sample size was 118 (59 in each group).

Statistical analyses: Baseline characteristics and outcome measures on nominal scales were analyzed by chi square test or Fisher exact test as appropriate. Baseline characteristics and outcome measures on continuous scales were analyzed by using two sample t test or Mann Whitney U test as appropriate. Statistical significance was considered if the P value was <0.05. Statistical analysis was performed by applying intention to treat principle. SPSS software for Windows version 18 (IBM SPSS, Inc, Chicago, IL) was used for statistical analysis.

A total of 340 infants born at <34 weeks gestation were admitted in NICU during the study period, out of which 181 infants were assessed for inclusion (Fig. 1). 118 infants were randomly assigned with 61 to Mask and 57 to Prongs group. The baseline demographic characteristics of enrolled infants were similar (Table I).

Fig. 1 The flow of participants in the study.

TABLE I	Baseline Characteristics of the Study Population (N=118)

CPAP failure was seen in 13% infants on nasal mask and in 25% infants on nasal prongs, but failed to reach statistical significance (P = 0.15) (Table II). Incidence of pulmonary interstitial emphysema was significantly lesser in infants on nasal mask as compared to nasal prongs (P=0.03), although higher flows were required in mask group which was statistically significant [6.2 (0.7) vs 5.9 (0.4) L/min, P= 0.008]. There was a significant lower incidence of overall nasal trauma in mask group as compared to prongs (36% vs 58%; RR 0.62, 95% CI 0.41-0.93; P=0.02). In terms of severity of nasal trauma, the infants in mask group had lower incidence of moderate trauma as compared to those in prongs group (7% vs. 21%; RR 0.31, 95% CI 0.10-0.91; P=0.02) which was statistically significant. The two groups were similar in terms of mild (P=0.55) and severe (P=0.48) nasal trauma. All the infants in the study with mild and moderate nasal trauma recovered fully before hospital discharge except one with columella necrosis who had prolonged NCPAP support with prongs for 3 weeks.

TABLE II Comparison of Outcomes between Nasal Mask and Nasal Prongs Groups

In our study, we found a reduced NCPAP failure within 72 hours of initiation of respiratory support in preterms in the mask group when compared with the prongs group; however, this difference was statistically not significant. Incidence of PIE was significantly lower in the Mask group than the Prongs. There was a significantly lower inci-dence of moderate and overall nasal trauma in Mask group.

The major limitation of our study was its non-blinded design with potential for bias in particular with assessment of nasal trauma due to the very nature of the intervention. When we planned the study, we assumed NCPAP failure rate of 40% in Prongs group and 20% in Mask group. On completion of our study we found these rates were 25% and 13%, respectively. Our study was therefore underpowered to demonstrate difference, if any, between the two intervention interfaces. Another limitation of our study was it being a single center study, as NCPAP failure rates may vary in other units.

A study by Kerian, et al. [8], compared mask vs prongs using IFD, and found that in terms of NCPAP failure, nasal mask was superior than prongs which was statistically significant (28% vs 52%). Our overall CPAP failure rate was comparable to the previously done studies [16,17]. Air leaks are known complications of CPAP therapy [18]. We found low incidence of PIE in the mask group which was an unexpected finding in our study. Similar findings in the mask group have been previously reported [8], though not reaching statistical significance. Paradoxically, we found that delivering NCPAP with nasal masks required higher flows than prongs which was statistically significant. We assume that need for higher flows could be due to some leak at the interface level, which could have played a protective role for air leaks in the mask group. Further studies are required to elucidate, if any, the causal relation between the interface and air leaks.

Nasal trauma has been found to be a major drawback associated with NCPAP use. Injury pattern in the nasal mask group was primarily seen at the base of nasal bridge with occasional injuries at the junction between the nasal septum and the philtrum sparing the columella and septum. This may be because the mask rests on the nasal bridge and philtrum with constant pressure hampering local tissue perfusion which triggers breach of skin barrier leading to inflammation and nasal trauma. Injury pattern in prong group was mainly seen at columella and anterior part of nasal septum which may be due to constant pressure between the two prongs. These findings of different sites of injury are consistent with the pattern described in a previous study of nasal trauma [11].

We conclude that NCPAP with mask as interface is equally effective as providing NCPAP with short bi-nasal prongs, but causes less PIE and nasal trauma. We have used Bubble CPAP as a primary mode of respiratory support thus making our results more generalizable and applicable in resource-limited settings. As masks and prongs cause nasal trauma in differing distribution, we suggest that the interface to be alternated after every 72 hours if NCPAP to be used for prolonged duration. We recommend more multicentric RCTs with appropriate sample size to evaluate impact of various delivery interfaces on NCPAP success along with associated side effects.

Acknowledgments: Dean, LTMMC and Hospital, Mumbai for permission. Dr Nandkishor S Kabra (Associate professor, Department of Neonatology, Seth G.S. Medical college and KEM Hospital, Mumbai) for analyzing the data.

Contributors: SG, JM, HP: conceived and designed the study; JM, AU, SM: were involved in patient care; SG, DGH, HP: collected the data; SG, HP: analysis and interpretation of data; SG, DGH, JM: Drafting the manuscript. All authors approved the final manuscript.

Funding: None; Competing interests: None stated.