Recorded sessions on demand will be available 24 hours after the session ends

Abstract Body

Background and Aim

Assessment of work of breathing (WOB) is essential to titrate respiratory support in preterm infants. Conventional WOB measurements using esophageal pressure are invasive and labor-intensive. In search of a less invasive and less complex method, we aimed to assess the relation between diaphragm activity measured with transcutaneous electromyography (dEMG) and WOB.

Methods

Clinically stable preterm infants (<37 weeks of gestation) on nCPAP were eligible. Esophageal pressure, tidal volume and diaphragm activity were recorded at different PEEP levels. Inspiratory WOB (WOB_i, normalized for tidal volume)) was calculated and compared to dEMG peak activity (dEMG_peak). Breath outcomes were sorted and averaged in ten incremental deciles, for correlation analysis between the two entities.

Results

Fifteen preterm infants (27.8 ± 1.8 weeks, birth weight 955 ± 259 grams) on nCPAP (median PEEP 4 (4-5) cmH₂O) were included. The level of WOB was low and showed little variation with different levels of PEEP (WOB_i 1.20 (0.90 – 1.76) vs. 1.25 (0.98 – 2.1) cmH₂O for PEEP 2 vs. PEEP 6 respectively). The same was true for dEMG_peak (1.5 (1.2 – 1.9) vs. 1.5 (1.2 – 2.1) µV). Correlation between dEMG activity and WOB was modest but variable with a median Pearson R of 0.65 (range -0.74 to 0.95) for dEMG_peak vs. WOB_i.

Conclusion

In stable preterm infants comparison of WOB and diaphragm activity showed a modest but variable correlation and these results warrant further investigation of dEMG as measure for WOB in infants with less stable lung conditions.

Abstract Body

BACKGROUND: Preterm infants are at risk of bradycardia and apnea. Current monitoring of respiratory rate (RR) by chest impedance (CI) is susceptible to motion artifacts. Detection of RR by esophageal electromyography (EMG) has been validated in ventilated patients. Kalman filters are signal filters designed to optimize estimation of signal parameters in the presence of interferences. We aimed to test feasibility of RR extraction by esophageal low frequency electrode motion (LFEM) in non-ventilated infants.

METHODS: Multichannel esophageal LFEM was recorded simultaneously to CI in neonates between 32 and 42 weeks postmenstrual age. Esophageal RR was extracted using a Kalman filter. We selected 100 esophageal segments of two minutes duration with significant bradycardia and/or apnea. A physician visually labeled RR in those segments as reference. A second blinded physician labeled a subset of 20 segments to test inter-rater reliability using Spearman’s coefficient. Esophageal RR and RR from CI were compared to the reference using Wilcoxon signed rank test with 5% significance level. Results are presented as median (interquartile range) breaths per minute (bpm).

RESULTS: Kalman filtering extracted RR in all labeled segments. Esophageal RR significantly differed from CI RR (7.31 (2.88-10.26) bpm, P<0.001). Inter-rater reliability was 0.73 bpm. Esophageal RR did not significantly differ from visually labeled RR (0.34 (-2.75-2.65) bpm, P=0.62), whereas CI RR was lower (-6.96 (-10.50-(-4.14)) bpm, P<0.001).

CONCLUSION: Extraction of RR by esophageal LFEM is feasible and closer to visually obtained RR than that from CI. These results indicate relative robustness to motion artifacts in non-ventilated preterm infants.

Abstract Body

Background and aims

Preterm infants present an increased risk of abnormal lung function and respiratory morbidity during infancy. Tidal breathing flow-volume (TBFV) measurements may be valuable in the follow up of these patients, as they are feasible at all post-natal ages and can identify airway obstruction. We aimed at assessing: 1) the trend of TBFV paraments over the first 2yrs of life; 2) the association between TBFV parameters and perinatal risk factors; 3) the predictive value of TBFV parameters for hospitalisation for respiratory infections (HRI), in preterm infants ± Bronchopulmonary Dysplasia (BPD).

Methods

We retrospectively analysed TBFV measurements performed at 0-6, 6-12 and 12-24months corrected age in 97 infants ≤32wks’ gestation. We assessed the association between TBFV parameters and perinatal factors using linear regressions, and the association between TBFV parameters and subsequent HRI using logistic regressions. To compare nested models, we used the area under the curve and the likelihood ratio test (LRT).

Results

Time to peak tidal expiratory flow/expiratory time ratio (tPTEF/tE) was lower than normal at all time points. Longer duration of oxygen-dependence, intubation and respiratory support were associated with reduced tPTEF/tE at all time points. For each z-score increase in tPTEF/tE, the OR for re-hospitalizations decreased by 0.70. tPTEF/tE added significantly to BPD classifications alone in predicting re-hospitalizations (AUC = 0.80 vs. 0.73, p-value for LRT < 0.001) (figure 1).

Conclusions

Preterm infants present persistent airway obstruction over the first 2yrs. TBFV measurements have the potential to identify infants at high risk of HRI regardless of BPD.

Abstract Body

Background and aims

Transcutaneous carbon dioxide (TcCO₂) monitoring is a non-invasive, continuous method of monitoring partial pressure of CO₂ (PCO₂). It provides an attractive alternative to potentially painful arterial or capillary blood sampling in a neonatal population. We aimed to determine the agreement of TcCO₂ with blood gas PCO₂ (BG PCO₂) in our tertiary neonatal intensive care unit (NICU).

Methods

This was a single centre prospective observational study performed in a tertiary NICU from August 2019 to November 2019. Routine care was continued and TcCO₂ simultaneously monitored. Agreement was assessed by Bland-Altman analysis and concordance correlation coefficient. End-user feedback on ease of use of TcCO₂ monitors was collected from staff.

Results

698 paired TcCO₂/BG PCO₂ samples from 19 invasively and non-invasively ventilated infants were analysed. Mean corrected gestation was 35⁺³ weeks (range 25⁺⁰ – 51⁺⁵) and weight 1985g (range 675g – 4900g). Overall bias was -0.30 (SD 1.21, p<0.0001) with good concordance (CCC= 0.80). 69% (95% CI 65%-72%, p=0.0003) of samples fell within our pre-defined clinically acceptable difference of 1 kPa. Agreement was more favourable for infants who were non-invasively ventilated or had a corrected gestational age >30 weeks. Staff feedback was uniformly positive and no infants suffered skin damage secondary to the heated sensors.

Conclusions

Our results suggest that TcCO₂ monitoring is a reliable assessment tool that can be used as an adjunct to blood gas analysis in a neonatal population. This can help reduce the number of invasive blood tests that are performed, particularly for non-invasively ventilated infants.

Abstract Body

Thoracic impedance apnoea monitoring is widely used in NICUs, but may fail to detect apnoea through misinterpretation of cardiac impedance change as breathing (poor sensitivity), and conversely may not recognise shallow breathing, falsely signalling apnoea (poor positive predictive value, PPV). We aimed to examine the performance of impedance-derived apnoea detection in routine clinical practice.

In preterm infants (n=35) impedance-derived apnoea alarm condition and capsule pneumography-derived apnoeic events (≥15 sec) were recorded. In pooled and per patient data, we examined the performance of impedance-derived apnoea monitoring, with the following categorization a) true positive (TP): impedance monitor detects apnoeic event; b) false negative (FN): impedance monitor fails to detect apnoea and c) false positive (FP): impedance monitor indicates apnoea despite continuation of respiratory effort on capsule pneumography. Sensitivity [TP/(TP+FN)] and PPV [TP/(TP+FP)] were determined.

Birth gestation was 27 (26-28) weeks [median (IQR)] and post-natal age 17 (11-23) days. A total of 1323 apnoeic events were identified by capsule pneumography in 3370 h of recordings (Table). The 773 impedance-derived apnoea alarms showed poor correspondence with detected apnoea (Table).

Impedance monitoring is both insensitive and poorly predictive for apnoea. An alternative apnoea monitoring system is needed to improve care for preterm infants.

Abstract Body

BACKGROUND AND AIMS:

Both lung ultrasound score (LUSscore) and N-terminal pro-B-type natriuretic peptide (NTproBNP) are early predictors of bronchopulmonary dysplasia (BPD) in patients born before 32 weeks of gestation (PTB32W). We studied if the predictive ability was increased using both biomarkers.

METHODS:
We performed lung ultrasound (LUS) in PTB32W at birth, at three days of age, at one and two weeks. We included patients with NT-proBNP at birth, at one or two weeks of age, and created two groups: group 1,non-BPD patients; group 2, BPD patients. We used logistic regression to calculate the best model to predict BPD using gestational age (GA) at birth, birth weight (BW), LUSscore and NTproBNP.

RESULTS:

We included: 71 patients in group 1, with median GA of 29 weeks (interquartile rank (IQR) 29-31); and 10 in group 2, with median GA of 27.5 weeks (IQR 24-29). There was a significant difference between groups in NTproBNP at two weeks, and in LUSscores from three days of birth onwards. The best model to predict BPD included BW, NTproBNP at two weeks and LUSscore at one week, with an adjusted R² (aR²) of 0.57, p=0.004, and an area under the receiver operating characteristic curve (AUCROC) of 0.96 (95%CI 0.87-1). This model improves the one using BW and NTproBNP at two weeks (aR²= 0.51, p=0.002, AUCROC=0.93) and the one using BW and LUSscore at one week (aR²= 0.48, p<0.0001, AUCROC=0.93).

CONCLUSIONS:

LUSscore at one week can improve the BPD prediction obtained by NTproBNP at two weeks of age in PTB32W.

Abstract Body

Background and aims:

Extremely preterm infants are exposed to an increased risk of postnatal essential fatty acids shortage. Supplementation with docosahexaenoic acid (DHA) and arachidonic acid (ARA) have been shown to be beneficial in prevention of retinopathy of prematurity. In contrast, enteral supplementation with DHA increased pulmonary morbidity in some studies. We aimed to compare differences in pulmonary outcome at various time points.

Methods:

Participants were randomized to receive either an enteral emulsion containing ARA and DHA (100 and 50 mg/kg/day, respectively) (ARA:DHA) or standard care (SC) from birth until 40 weeks of postmenstrual age (PMA).

Results:

Out of 206 randomized infants with gestational age <28 weeks, 174 infants surviving to PMA 40 weeks were analysed. Bronchopulmonary dysplasia (BPD) were distributed in ARA:DHA and SC groups as follows: no BPD: 4.9% vs 5.4%, mild: 36.6% vs 42.4%, moderate: 24.4% vs 26.1% and severe: 34.1% vs 26.1% (p=0.71). 40.2% in ARA:DHA vs 45.7% in SC needed no breathing aid at 36 weeks PMA. 26.8% vs 26.1% needed low flow nasal cannula (LFNC), and 25.6% vs 26.1% high flow nasal cannula (HFNC) (p=0.42). At 40 weeks PMA, 59.8% vs 73.9% required no breathing aid, 26.8% vs 19.6% required LFNC, and 11.0% vs 5.4% HFNC (p=0.22). Duration of supplemental oxygen requirement were similar in both groups (p=0.18) (figure 1).

Conclusions:

Enteral DHA:ARA supplementation did not significantly alter pulmonary morbidity in extremely preterm infants. A tendency towards more severe BPD and more respiratory support could be noted and will be further evaluated.

Abstract Body

Continuous positive airway pressure (CPAP) with nasal prongs or mask is the recommended first-line treatment for newborn infants in respiratory distress and relies on a tight seal between patient and device. Measuring leakage without affecting the child is challenging and quantifying studies are scarce.

Our aim was to measure absolute leakage with nasal mask and prongs during neonatal CPAP treatment and evaluate leakage reducing actions.

Using the flow-through technique, we developed a clinically applicable method to measure absolute leakage without affecting the infant.

Our multi-center, randomized, cross-over study included 50 clinically stable infants with nasal CPAP treatment, aged between 28 and 44 weeks of gestational age. The interface order was randomized and each interface was applied by a nurse blinded to the current leakage. Leakage was measured for 30 seconds during quiet breathing. Leakage reducing actions included changing the interface size, manually adjusting the interface, applying a light pressure at the feeding line entry, adding or removing a pacifier and changing body position.

Analyses showed a significantly lower leakage (mean difference 0.86 LPM, 95% CI 0.07 to 1.65, p= 0.034) with prongs (median leakage 2.01 LPM, IQR 1.80) compared to nasal mask (median leakage 2.45 LPM, IQR 4.13). Leakage reducing actions lead to significant reductions in leakage (p<0.01) for both prongs (1.28 LPM, IQR 1.36) and nasal mask (2.40 LPM, IQR 3.99). The absolute leakage measuring method is plausible and clinically applicable.

Further studies on the clinical effects of leakage, and how interfaces can be optimized, are mandated.