Introduction

Respiratory syncytial virus (RSV) bronchiolitis is the most common cause of lower respiratory tract infection and the leading reason for hospitalization among infants in developed countries [1]. Since currently no therapeutic option is available for the efficient management of this multifactorial bronchiolar obstruction [2], treatment is mainly based on supportive care, with oxygen and hydration. Admission to a paediatric intensive care unit (PICU) is required in 2–6% of the cases, particularly in ex-premature and young infants with evolving respiratory distress [3, 4], to provide mechanical ventilation, if necessary. Indeed, these patients have evidence of severe obstructive airway disease with markedly increased respiratory system resistance, air trapping and decreased dynamic compliance [5]. The consequence is a higher risk of ventilatory failure secondary to an excessive load of the respiratory muscles [6] which are more susceptible to fatigue due to the immature pattern of ventilatory muscles fibres in early infancy [7].

Nasal continuous positive airway pressure (nCPAP) has been shown to effectively decrease respiratory muscle load in children with mechanical upper airway obstruction [8]. No data, however, has been reported for young infants with more distal obstructive airway diseases, such as severe acute viral bronchiolitis. Specific technical concerns have hindered nCPAP application in this population for a long time. Particularly, uncontrolled air leaks and the need to maintain a constant airway pressure throughout the entire respiratory cycle limited its potential benefits (e.g., decrease in respiratory work) [9]. The development of a device, which converts the kinetic energy of a jet of fresh gas [10], currently allows the delivery of a stable CPAP in neonates and young infants [11].

We, thus, wanted to know whether nCPAP delivered by a system specifically designed for young infants and neonates could efficiently decrease respiratory distress symptoms in this group of patients by unloading respiratory muscles during RSV acute obstruction [12]. To this aim we evaluated the short-term effects of nCPAP application during severe RSV bronchiolitis on (i) the clinical respiratory distress score, as assessed by the modified Wood’s Clinical Asthma Score (m-WCAS) [13], and on (ii) the breathing pattern and muscular respiratory effort, estimated from oesophageal (Pes) and gastric (Pgas) pressures.

Material and methods

Patients

We considered eligible for this study only infants less than 3 months of age, who were hospitalised at the PICU of the Arnaud de Villeneuve University Hospital for RSV bronchiolitis. We did not set any weight limit and used the following inclusion criteria:

  1. 1.

    Clinical diagnosis of bronchiolitis, as defined by the presence of upper respiratory tract infection followed by tachypnea, chest retraction, wheezing and hyperinflation on chest radiograph

  2. 2.

    Confirmation of RSV-positive bronchiolitis by enzyme immunoassay of a nasopharyngeal swab specimen

  3. 3.

    Severe respiratory distress, as defined by capillary PCO2 > 50 mmHg and by m-WCAS > 5; infants with RSV-related apnea alone were not included in this study

  4. 4.

    Absence of pneumothorax on chest radiograph, or corticosteroid or bronchodilator treatment during the first 2 h of the enrolment

  5. 5.

    Signed authorisation by the parents

Clinical score and monitoring

A single observer, who was not involved in patient’s care and pressure recordings, quantified the respiratory distress with the m-WCAS (Table 1), and used a previously described visual analogue scale [13] to standardise the scoring of accessory muscles’ recruitment and expiratory wheezing. We carried out continuous monitoring of heart rate, pulse oximetry (SpO2) and intermittent blood pressure measurement by oscillometry with an Intellivue MP70 cardioscope (Philips Medical Systems, The Netherlands), and measured transcutaneous PO2 (PtcO2) and PCO2 (PtcCO2) using a Tina TCM4 (Radiometer, Denmark).

Table 1 Modified Wood’s clinical asthma score (m-WCAS)
Full size table

Pressure measurements

We calibrated the pressure transducer with a 20-cm water column before each measurement. We assessed the inspiratory effort by oesophageal pressure measurement (Pes) using a neonatal oesophageal balloon catheter (Marquat, France) mounted on a differential pressure transducer system (T2000, Gould, USA). We inserted the catheter orally and advanced it gently until the balloon was in the middle portion of the oesophagus. We then checked that the oesophageal balloon was in a satisfactory position by continuous on-line Pes monitoring and adjusted it until we obtained a clear oscillating signal in the presence of a negative deflection during inspiration. In the five oldest (i.e., 39- to 90-day-old) infants, we could evaluate also the expiratory muscles activity in parallel with Pes by inserting both an oesophageal and a gastric balloon catheter. We evaluated the placement of the gastric balloon by gentle manual pressure on the infant’s abdomen to observe the fluctuations in gastric pressure (Pgas) in the absence of Pes fluctuations. The same investigator, who was not involved in the care of the patients and was unaware of their scorings, performed all these measurements.

Study protocol

We conducted this prospective study during two RSV epidemic periods, from November 2004 to March 2006. We placed the infants who met the inclusion criteria under humidified oxygen/air blend on a heating base and adjusted the FiO2 so as to reach a SpO2 of 94–98% during the entire period of study. After 10 min of observation in a quite environment, the observer filled in the m-WCAS. Then the designed investigator measured Pes and Pgas, in the five oldest children. Fifteen minutes later, the observer evaluated again the m-WCAS and we validated the pressure values only if we did not find any differences between the two m-WCAS. After these baseline measurements, we applied nCPAP using an Infant Flow Ventilator (Electro Medical Equipment, UK) equipped with an adequate humidifier (Fisher and Paykel 850, CA) while leaving in place the oesophageal and gastric catheters. This device delivered nCPAP (+6 cm H2O) via a mask, connected to a twin injector nozzle fixed to the patient by a specific bonnet. We quantified again m-WCAS, Pes, Paw and Pgas 1, and again 6 h after nCPAP application.

Assessment of respiratory load and data analysis

Each result, in basal conditions and during nCPAP delivery, corresponded to the mean value of 30 consecutive and regular respiratory cycles. We calculated the respiratory rate, inspiratory (Ti) and expiratory (Te) time and Pes swings from the Pes traces (Fig. 1).

Fig. 1
figure 1

Tracings of gastric (Pgas) and oesophageal (Pes) pressures in one of the infants with acute bronchiolitis. Left panel In spontaneous ventilation, Pes swings (thin arrows) reach 50 cm H2O during inspiration and Pgas swings (thick arrows) reach 3–4 cm H2O during expiration. Right panel After application of +6 cm H2O nCPAP for 1 h, we observed a significant decrease in Pes swings during inspiration and disappearance of Pgas swings during expiration. Paw airway pressure

Full size image

We used the pressure–time product (PTP) per breath and per minute, defined as the area under the pressure–time curve, proposed to estimate the metabolic cost of breathing [14]. We quantified this area by scanning the surface, using the NIH image programme. For this study, we did not refer PTP to the chest wall static recoil pressure–time curve relationship [15], because we were unable to obtain accurate tidal volume measurements in infants of this age. We calculated the inspiratory muscles’ pressure–time product per breath (PTPesinsp/breath) by measuring the mean area under the Pes signal between the onset of the inspiratory effort and the end of inspiration (∫Pes dt). We expressed PTPesinsp per minute by multiplying the pressure–time products per breath by the breathing frequency (PTPesinsp/min = ∫Pes dt × breathing frequency). Swings of gastric pressure and the expiratory muscles pressure–time product (PTPgasexp) were assessed from the Pgas traces. We obtained the PTPgasexp/breath (∫Pgas dt) by measuring the mean area under the positive inflection of the Pgas signals between the onset of the expiratory effort and the end of expiration, defined as the return to baseline of the Pgas signal. We calculated PTPgasexp/minute by multiplying the pressure–time products per breath by the breathing frequency (PTPgasexp/min = ∫Pgas dt × breathing frequency). We did not include the positive peak of Pgas, induced by diaphragmatic contraction during inspiration, for assessing Pgas swings and PTPgasexp.

We used a one-way repeated analysis of variance to compare changes in parameters over the studied period with post-test comparison of means using the Tukey–Kramer’s test, and the Pearson’s correlation coefficient to assess the correlations between continuous variables. We expressed data as mean (SE) and considered significant only differences with a P value < 0.05. We performed all statistical analyses with SAS (SAS Institute, NC) in collaboration with the Department of Statistics at the Montpellier University Hospital (EB).

Ethical consideration

We obtained informed consent from the parents of all infants, and the University Hospital Human Subjects Committee of Montpellier approved this protocol.

Results

From November 2004 to March 2006, 34 infants with respiratory distress attributable to RSV bronchiolitis were admitted to the PICU. Among them, 20 were excluded because their clinical score was <5, and two for being older than 3 months. All parents gave their consent to enter this protocol. The selected 12 infants had a mean postnatal age of 43 (±6) days, weight of 3.6 (±0.3) kg and PRISM score of 8.6 (±1.2). Seven were born before 36 completed weeks of gestation, at a median gestational age of 33 (32.5–36.0) weeks, but none of them had a chronic lung disease. Two other infants had a ventricular septal defect with a moderate shunt that did not induce pulmonary arterial hypertension and/or cardiac failure. Patients 5 and 12 received nebulised β-adrenergic agents 4–6 h before the enrolment.

Effects of nCPAP on respiratory distress

Severe respiratory distress at baseline was attested by a high respiratory distress score, oxygen requirement and respiratory acidosis. Application of nCPAP was associated with a decrease in m-WCAS within 1 h in 11 of the 12 patients (Fig. 2, left panel). Among the different components of the m-WCAS, only the scores for accessory muscles’ use [from 1.7 (±0.08) to 0.8 (±0.13), P = 0.002] and expiratory wheezing [from 1.3 (±0.18) to 0.3 (±0.13), P = 0.002] significantly decreased after nCPAP. nCPAP was also linked to lower FiO2 needed to reach the targeted SpO2, and a decrease in PtcCO2, heart rate and mean arterial blood pressure (Table 2).

Fig. 2
figure 2

Individual data of the 12 infants with acute viral bronchiolitis, during spontaneous breathing (SB) and nasal continuous positive airway pressure (nCPAP) delivery. Left panel Evolution of the modified Wood’s clinical asthma score (m-WCAS), which evaluates respiratory distress. Right panel Evolution of the inspiratory muscle pressure–time product (PTPesinsp), an index of inspiratory effort

Full size image
Table 2 Evolution of respiratory distress after nasal continuous positive airway pressure (nCPAP) application
Full size table

Effects of nCPAP on breathing pattern and respiratory effort

The baseline breathing pattern was characterized by tachypnea with a Ti/Te ratio close to 1. After 1 h of nCPAP, we observed a significant decrease in the Ti/Ttot ratio without any difference in the respiratory rate. These modifications were already visible after 1 h of nCPAP and did not change further after 6 h (Table 3).

Table 3 Evolution of the breathing pattern and the inspiratory effort after nasal continuous positive airway pressure (nCPAP) delivery
Full size table

In the hour immediately following nCPAP we detected a homogeneous reduction in the values of three parameters of inspiratory effort in comparison to those at baseline: specifically, a 53 (±5)% decrease of Pes swings, and a reduction of 58 (±4)% of PTPesinsp/breath and 59 (±5)% of PTPesinsp/min. Individual variations of PTPesinsp/min after nCPAP are presented in Fig. 2 (right panel).

In the five oldest infants with simultaneous double recordings, we quantified the expiratory muscles activity using Pgas swings, PTPgasexp/breath, and PTPgasexp/min, which, at baseline, were 3.7 (±0.4) cm H2O, 1 (±0.3) cm H2O s, and 55 (±10) cm H2O s min−1, respectively. After 1 h of nCPAP, the expiratory muscle activity completely disappeared (Fig. 1).

Correlations between respiratory distress score and inspiratory effort

At baseline and after nCPAP, we found significant correlations between m-WCAS and, respectively, Pes swings (r = 0.56, P = 0.004), PTPesinsp/breath (r = 0.60, P = 0.001) and PTPesinsp/min (r = 0.61, P = 0.001). After nCPAP, we also detected a significant correlation between the decrease in Pes swings and accessory muscles’ use score (r = 0.8, P = 0.001), and the reduction in Pes swings and wheezing score (r = 0.70, P = 0.02), but did not observe similar correlations for PTPes/breath and PTPes/min. We did not look for correlations between respiratory distress score and expiratory effort due to the paucity of available data.

Patients’ follow-up

No complications, such as barotraumas, occurred during nCPAP treatment, and none of the infants worsened or required endo-tracheal intubation. We maintained nCPAP for 135 (±36) h. All infants were discharged from the unit, after a mean length of stay in the PICU of 7 (±0.8) days. No patient required PICU readmission after discharge.

Discussion

This study demonstrates that nCPAP can adequately decrease respiratory effort in young infants with acute RSV bronchiolitis and that this physiological effect is correlated with a significant reduction in the respiratory distress score as quantified by the m-WCAS.

RSV bronchiolitis and respiratory distress score

Among the different respiratory scoring systems developed to evaluate acute bronchiolitis in hospitalised infants, we selected m-WCAS because it includes only clinical items that can be quantified in children of any age and is potentially accessible to different care providers. Although we previously noted good inter-observer agreement among physicians in our PICU [13], for this study, the same person, who was not involved in the care and pressure recordings, performed all the scorings to assess the m-WCAS ability to relate disease severity and responsiveness to therapy. The baseline scores suggest that our population was sicker and more homogeneous than those of previous studies, where the same scoring system was used [16, 17], and clearly indicate that all our patients had severe bronchiolitis with a potential risk of tracheal intubation.

Effects of nCPAP on the respiratory distress score

After 1 h of nCPAP, we observed a sharp reduction in m-WCAS that was only due to a decrease in accessory muscles’ use and expiratory wheezing. Since the evaluation of these parameters is rather difficult [18, 19], we used also a visual analogue scale, similar to that for pain measurement except that this scale was represented by a 20 cm instead of 10 cm line to allow more precise scorings [13]. Thanks to this aid, these two m-WCAS components could provide an accurate estimation of the breathing workload, as suggested by the correlation between their decrease and the Pes swing drop after nCPAP.

In this study, m-WCAS appeared to be an efficient tool to link respiratory effort intensity and responsiveness to therapy; nevertheless, we have specific comments about three patients. The two patients (i.e. 7 and 8) with the lowest baseline values for PTPesinsp had the lowest decrease in m-WCAS and PTPesinsp values following nCPAP application. However, their respiratory distress score, breathing pattern, blood gases and oxygen requirements at baseline were comparable to those of the other patients. This observation suggests that the identification of the minority of patients with severe bronchiolitis, who may not require nCPAP, may be difficult in the absence of Pes measurements. The patient with the slowest improvement of m-WCAS (i.e. 9), was admitted to the PICU at a late stage after respiratory distress onset, suggesting that rapidity of intervention could be the determinant for nCPAP efficacy.

RSV bronchiolitis and muscular respiratory effort

By monitoring Pes, we measured the consequences of RSV-induced airway obstruction on the inspiratory effort required to ensure alveolar ventilation. At baseline, the mean Pes swing developed at each inspiration was above 25 cm H2O, a value six times higher than the peak inspiratory pressure (about 4 cm H2O) we normally measure in healthy infants in our laboratory [20]. The inspiratory effort’s level of our patients was comparable to that reported previously in babies with acute viral bronchiolitis [6] and in older infants with upper airway obstruction [8, 21]. Their respiratory pattern in response to this higher inspiratory load was also modified, with increased breathing frequency and Ti/Ttot ratio. Such ventilatory adaptation has been observed in premature infants [22]. In our patients, the main consequence was a decrease in the expiratory length that might have induced dynamic hyperinflation and generated an intrinsic positive end expiratory pressure (PEEPi) [23]. Precise assessment of PEEPi would have required airflow monitoring [24], but we decided not to measure the tidal volume, because, in our experience, face-mask pneumotachography is poorly tolerated by young infants with acute bronchiolitis. Indeed, it creates agitation and respiratory pattern alterations that artificially modify Pes and m-WCAS.

We could perform simultaneous recordings of Pes and Pgas to evaluate expiratory muscles activity only in the oldest infants, and we limited this measurement to 2 h, because an orogastric tube was then required for regular aspiration of swallowed air from the stomach. In these patients, Pgasexp swings reached nearly 4 cm H2O, suggesting the recruitment of expiratory muscles, which has been previously observed in adult with severe chronic airflow obstruction [25].

Effects of nCPAP on the muscular respiratory effort

The variable-flow nCPAP ventilator used in this study was developed for premature infants, and could be used for our patients because of their young age and low weight. Theoretically, the high speed jet gas flow of this device allows the maintenance of a nearly constant pressure support during inspiration [10]. Consistently, we observed that nCPAP significantly decreased the Ti/Ttot ratio, Pes swings and PTPesinsp in all our patients. These findings indicate that this pressure support improves the breathing pattern and reduces the respiratory effort in very young infant with RSV infection, possibly through an alleviation of the bronchiolar obstruction and/or an offset of the patient’s inspiratory effort to overcome PEEPi [26, 27].

We also found that nCPAP completely abolished the expiratory muscles activity. Although expiratory muscles activity does not contribute to the inspiratory threshold load [24], increased expiratory effort could promote small airway collapse. The reduction of FiO2 and PCO2 observed in our patients may originate from nCPAP’s widening effect on the terminal airways, which prevents the airway collapse and ultimately improves alveolar ventilation [28]. The decrease in Pgas swings could be related to the change in the breathing pattern after nCPAP application, particularly the increase in Te that favours passive expiration [22]. However, Pgas swings’ reduction could be also due to the nCPAP driver used, as it shunted gas away from the infant during expiration, contrary to the constant nasal flow of supplemental oxygen delivery at baseline, which could lead to increased expiratory work [29, 30].

Tolerance of nCPAP

For this pilot study, we empirically set nCPAP to 6 cm H2O, because this level has been shown to be effective and safe in infants with mild respiratory distress [31]. We did not test higher nCPAP level that could be relevant in infants with particularly severe obstruction. Indeed, effect of optimized pressure level, adjusted to the individual Pes swings, should be assessed in further studies. In this study, nCPAP appeared to be relatively complication-free, particularly barotrauma. Although lung hyperinflation can significantly impair cardiac performance especially in preterm and young infants [32, 33], we think that the decrease in mean arterial blood pressure and heart rate we observed in our patients was more likely due to the clinical improvement and CO2 reduction than to an excessive distending pressure. As a matter of fact, these infants appeared more settled during nCPAP and their cardiovascular parameters returned to more physiological values for their age.

In conclusion, our study shows that nCPAP can rapidly and markedly unload respiratory muscles and decrease respiratory distress in infants with severe bronchiolitis. It also improves gasometric and hemodynamic parameters, and has no adverse effects in spite of the prolonged use.