Neurally adjusted ventilatory assist improves survival, and its early application accelerates weaning in preterm infants

Abstract

Background

Evidence to show that neurally adjusted ventilatory assist (NAVA) improves clinical outcomes is lacking. We aimed to analyze whether NAVA improves respiratory outcomes in preterm infants who require invasive mechanical ventilation.

Methods

A retrospective cohort study was conducted in 122 very low birthweight infants who required invasive mechanical ventilation for more than 24 h at one tertiary neonatal intensive care unit in Korea from January 2016 to June 2023. Subjects were divided into three groups: early NAVA for those supported with NAVA before the seventh day of life (n = 18), late NAVA for those supported with NAVA later than the seventh day (n = 18), and conventional for those supported with conventional ventilation modes other than NAVA (n = 86).

Results

There was no difference in the composite outcome of bronchopulmonary dysplasia or death among the three groups. Neonates who had been supported with NAVA at some point had lower odds of mortality than those who had not (adjusted odds ratio [aOR] 0.09, 95% CI 0.01–0.90, p = 0.040 for the early NAVA group; aOR 0.15, 95% CI 0.03–0.81, p = 0.027 for the late NAVA group). The adjusted hazard ratio for invasive mechanical ventilation weaning was higher in neonates supported with NAVA within the first week of life than in those supported with other ventilation modes (aHR 2.02, 95% CI 1.14–3.57, p = 0.015).

Conclusions

Neurally adjusted ventilatory assist application was associated with lower odds of mortality, and its early application from the first few days of life helped preterm infants wean from invasive mechanical ventilation sooner.

INTRODUCTION

Invasive mechanical respiratory support can be lifesaving for preterm infants but it can also cause lung injury, particularly if a high pressure and/or volume is used. Studies have therefore focused on lung‐protective ventilatory strategies or fast weaning from invasive respiratory support with switching to noninvasive respiratory support. ¹ However, no meta‐analysis has demonstrated a significant reduction in bronchopulmonary dysplasia (BPD) or mortality rate among preterm infants with the use of a single mode of ventilation in comparison with other modes of ventilation. ² , ³

Neurally adjusted ventilatory assist (NAVA) provides synchronized and proportional respiratory support based on a patient's electrical activity of the diaphragm (Edi). ⁴ Studies in preterm infants have shown that NAVA, when compared with conventional ventilation modes, lowers the peak inspiratory pressures, tidal volumes and work of breathing, ⁵ , ⁶ , ⁷ and it achieves better synchronization ⁸ , ⁹ and oxygenation. ⁷ , ¹⁰ These physiological benefits may reduce the risk of barotrauma or volutrauma, and may ultimately prevent the development of chronic lung diseases. However, evidence showing that NAVA improves clinical outcomes is lacking. ¹¹ We thus aimed to analyze whether NAVA improves respiratory outcomes in preterm infants who require invasive mechanical ventilation. This study sought to determine whether NAVA, when used as a primary or rescue mode of ventilation, results in the promotion of weaning from mechanical ventilation and a reduction in BPD or death among preterm infants in comparison with other forms of mechanical ventilation.

METHODS

This retrospective cohort study included 175 very low birthweight infants who required invasive mechanical ventilation for more than 24 h in the level III neonatal intensive care unit (NICU) of Inha University Hospital, South Korea, from January 2016 to June 2023. There were no subjects with congenital or chromosomal abnormalities. We excluded infants who were transferred from other hospitals after 24 h from birth (n = 6), those who were intubated after 24 h from birth (n = 12), and those who died before 7 days of life (n = 35) (Figure 1). From 122 included infants, the conventional group included 86 infants supported with any conventional ventilation modes except NAVA. Among infants supported with NAVA for at least 24 h regardless of the use of other modes, the early NAVA group included those supported with NAVA before the seventh day of life (n = 18), and the late NAVA group included those supported with NAVA after the seventh day (n = 18).

FIGURE 1.

Study participants. Values are number (%) or median (IQR). AC, assist control; BW, birthweight; DOL, day of life; GA, gestational age; HFOV, high‐frequency oscillatory ventilation; MV, mechanical ventilation; NAVA, neurally adjusted ventilatory assist; PMA, postmenstrual age; PS, pressure support; SIMV, synchronous intermittent mandatory ventilation; VG, volume guarantee; VLBW, very low birthweight.

The decision about which ventilator or which ventilation mode to use was made by the clinician in charge. The invasive mechanical ventilation for preterm infants was traditionally started using Babylog VN500 (Dräger) in assist control with volume guarantee or Servo‐n/Servo‐i (Getinge) in synchronized intermittent mandatory ventilation modes in this unit. After NAVA was introduced in 2014, as it became the primary ventilation mode in this unit, it was applied for intubated infants whenever they had active spontaneous breathing. Some clinicians, however, preferred to use conventional modes of ventilation. Sometimes, when clinical deterioration occurred during NAVA, conventional modes or high‐frequency oscillatory ventilation (HFOV) were used briefly and then the patient was returned to NAVA after respiratory stabilization. The HFOV mode was used only for the rescue strategy when previous ventilation modes were unable to provide safe ventilation or gas exchange in the setting of severe hypoxemic respiratory failure. During the NAVA, the trigger Edi level was set as 0.5 μV above the minimum Edi. Attending physicians decided and adjusted the NAVA level in the range of 1.0–2.5 cmH₂O/μV to target the peak Edi of 5–15 μV. Positive end‐expiratory pressure, fraction of inspired oxygen and backup ventilation (pressure control mode) settings were adjusted according to the infant's condition. The apnea time was set as 2–5 s, according to each patient's pulmonary status, respiratory drive, and the severity of desaturation and bradycardia during apnea. After the set apnea time without the Edi signal, the backup ventilation started with a preset peak pressure and frequency until the Edi trigger resumed. The usual backup rate was 20–50 breaths/min and the inspiratory time was 0.3–0.5 s. The safety limit for peak inspiratory pressure was usually set as 25–40 cmH₂O.

The primary outcome was BPD at 36 weeks postmenstrual age (PMA) or death. The secondary outcomes were the development of BPD, high‐grade (≥2) BPD, death, and weaning from mechanical ventilation. We used the Jensen classification to define BPD and graded its severity at 36 weeks PMA:¹² grade 1 if 2 L/min or less was required using a nasal cannula; grade 2 if more than 2 L/min was required using nasal cannula or other forms of non‐invasive ventilation support; and grade 3 if invasive mechanical ventilation was required.

The following clinical data were collected from electrical medical records: gestational age at birth, birthweight, sex, body temperature at NICU admission, histological chorioamnionitis, respiratory distress syndrome, patent ductus arteriosus requiring medical or surgical treatment, pathogen‐proven sepsis, necrotizing enterocolitis of stage 2 or higher, and death. Regarding mechanical ventilation data, every mode of mechanical ventilation and the start/end date of each mode, intubation and extubation dates and second and third in‐/ex‐tubation dates if the infant was reintubated, total duration of invasive mechanical ventilation, and respiratory support details at 36 weeks PMA and discharge was collected. If the interval between the prior extubation and the next intubation was less than 7 days, the last extubation date was recorded as the next extubation date.

The study was approved by the institutional review board of Inha University Hospital, Incheon, South Korea (IRB no. 2023–08‐030). The requirement for written informed consent was waived by the board due to the retrospective nature of this study.

Statistical analyses

Data are presented as numbers (percentages) for categorical variables and as medians with interquartile ranges (IQRs) for continuous variables. Disparities in baseline characteristics among the three groups were assessed using the Kruskal–Wallis test with the Bonferroni correction for continuous variables and Fisher's exact test for categorical variables in univariate analyses. To examine the impact of the three groups on BPD, death, and composite outcomes, a multivariate logistic regression analysis was conducted, adjusting for clinical factors including sex, bodyweight, gestational age, body temperature at NICU admission, histological chorioamnionitis, patent ductus arteriosus, sepsis, and necrotizing enterocolitis. Kaplan–Meier survival curves were employed to assess the time until weaning from invasive mechanical ventilation and mortality. The weaning point was set as the last extubation date. To investigate differences among the groups regarding weaning from mechanical ventilation and mortality, multivariate Cox proportional hazard models for survival analysis was used, adjusting for clinical variables such as sex, bodyweight, gestational age, body temperature at NICU admission, chorioamnionitis, patent ductus arteriosus, sepsis, and necrotizing enterocolitis. All statistical analyses were performed using R software version 4.1.2. If p <0.05, this was regarded as statistically significant for all statistical tests.

RESULTS

A total of 122 preterm infants born at a median (IQR) 27⁺¹ (25⁺³–28⁺⁶) weeks of gestation with a median (IQR) birthweight of 912 (753–1133) g were included in the study. In the conventional group, the most frequently used mode of ventilation was assist control with volume guarantee (n = 74), followed by HFOV (n = 50) and synchronous intermittent mandatory ventilation with pressure support (n = 18). For six infants (7%), all of three modes of ventilation were used, for 44 infants (51%), two modes were used, and for 36 infants (42%), one mode was used. In the early NAVA group, NAVA was started at a median (IQR) 2 (1–4) day of life and a PMA 26⁺⁶ (25⁺⁴–28⁺³) weeks. Among infants in the early NAVA group, NAVA alone was used for only six infants (33.3%), and one and two other modes were used for four and eight infants before switching to NAVA, respectively. There was no adverse event related to application of NAVA early in their life. In the late NAVA group, NAVA was started at the median (IQR) 27 (22–37) day of life and a PMA 29⁺⁵ (27⁺⁴–31⁺¹) weeks. All infants in the late NAVA group were exposed to conventional ventilation modes before switching to NAVA (Figure 1).

In the comparison of baseline characteristics (Table 1), gestational age was significantly lower in the late NAVA group than in the conventional group (median [IQR], 25⁺⁵ [24⁺⁴–26⁺²] weeks vs. 28⁺² [26⁺⁶–29⁺²] weeks, p = 0.008). Dexamethasone was used more frequently in the late NAVA group for the purpose of facilitating ventilator weaning than in the conventional group (61.1% vs. 17.4%, p < 0.001).

TABLE 1.

Baseline characteristics.

	Conventional (n = 86)	Early NAVA (n = 18)	Late NAVA (n = 18)	p value a
Gestational age, week	28+2 (26+6–29+2)	26+4 (25+4–28+4)	25+5 (24+6–26+2) b	0.016
Birthweight, g	1030 (870–1240)	880 (820–1090)	810 (720–910)	0.079
Sex, male: female	50:36	8:10	10:8	0.586
Initial body temperature,°C	36.5 (36.3–36.8)	36.7 (36.5–36.8)	36.4 (36.2–36.7)	0.085
Chorioamnionitis	33 (38.4)	8 (44.4)	5 (27.8)	0.619
Respiratory distress syndrome	85 (98.8)	18 (100)	17 (94.4)	0.505
Patent ductus arteriosus	43 (50.0)	13 (72.2)	13 (72.2)	0.073
Sepsis	16 (18.6)	2 (11.1)	6 (33.3)	0.274
Necrotizing enterocolitis, stage ≥2	18 (20.9)	3 (16.7)	4 (22.2)	0.999
Use of dexamethasone	15 (17.4)	5 (27.8)	11 (61.1) c	<0.001

Note: Values are presented as numbers (%) or median (interquartile range).

Abbreviation: NAVA, neurally adjusted ventilatory assist.

^a Fisher exact test or Kruskal–Wallis test with Bonferroni test.

^b Significant difference between conventional and late NAVA groups (p = 0.008).

^c Significant difference between conventional and late NAVA groups (p < 0.001).

There was no difference in the composite outcome of BPD or death among the three groups. The late NAVA group had higher odds of BPD (adjusted odds ratio [aOR] 18.09, 95% CI 1.57–208.40, p = 0.020) than the conventional group. However, there was no difference in BPD of grade 2 or higher (aOR 0.32, 95% CI 0.06–1.60, p = 0.166). Both the early and late NAVA groups had lower odds of mortality than the conventional group (aOR 0.09, 95% CI 0.01–0.90, p = 0.040 for the early NAVA group; aOR 0.15, 95% CI 0.03–0.81, p = 0.027 for the late NAVA group) (Table 2).

TABLE 2.

Outcomes with multivariate logistic regression analysis.

	Conventional (other modes except NAVA)	Early NAVA		Late NAVA
		aOR a (95% CI)	p Value	aOR a (95% CI)	p Value
BPD b or death	Reference	0.36 (0.08–1.53)	0.165	4.33 (0.13–145.24)	0.414
BPD b	Reference	0.56 (0.12–2.55)	0.455	18.09 (1.57–208.40)	0.020
BPD, grade ≥ 2 b	Reference	1.57 (0.37–6.74)	0.545	0.32 (0.06–1.60)	0.166
Death	Reference	0.09 (0.01–0.90)	0.040	0.15 (0.03–0.81)	0.027

Abbreviations: aOR, adjusted odds ratio; BPD, bronchopulmonary dysplasia; NAVA, neurally adjusted ventilatory assist.

^a Adjusted for gestational age, bodyweight, sex, initial body temperature, chorioamnionitis, patent ductus arteriosus, sepsis, and necrotizing enterocolitis.

^b Jensen classification was used to define BPD and its severity at 36 weeks postmenstrual age; grade 1 if required 2 L/min nasal cannula or less, grade 2 if required more than 2 L/min nasal cannula or other forms of non‐invasive ventilation, and grade 3 if required invasive mechanical ventilation.

Bronchopulmonary dysplasia was developed in 29.1%, 44.4%, and 83.3% of infants in the conventional, early NAVA, and late NAVA group, respectively (p = 0.582 for the conventional group vs. the early NAVA group; p < 0.001 for the conventional group vs. the late NAVA group; p = 0.026 for the early NAVA group vs. the late NAVA group) (Supporting Information, Table S1). However, the incidence of high‐grade BPD showed no difference among the three groups (16.3%, 38.9%, and 27.8%, respectively; p = 0.214). The mechanical ventilation duration was significantly longer in the late NAVA group (p = 0.002 for the conventional group vs. the late NAVA group; p = 0.004 for the early NAVA group vs. the late NAVA group). However, because death occurred more frequently in the conventional group than in the early or late NAVA groups, the mechanical ventilation duration alone could not be used to evaluate the weaning effect of NAVA. The adjusted hazard ratio (aHR) of weaning from invasive mechanical ventilation was higher in neonates supported with NAVA before the seventh day of life (aHR 2.02, 95% CI 1.14–3.57, p = 0.015) than in those supported with other conventional ventilation modes (Figure 2 and Supporting Information, Table S2). Most infants in the early NAVA group succeeded in weaning from mechanical ventilation before 4 weeks of age. In this model, a higher gestational age and initial body temperature also had higher HRs for earlier weaning from mechanical ventilation (aHR 1.39, 95% CI 1.14–1.69, p = 0.001 and aHR 1.76, 95% CI 1.07–2.88, p = 0.025, respectively).

FIGURE 2.

Adjusted curve for weaning from mechanical ventilation. A multivariate Cox proportional hazard model was used. This curve shows that the early NAVA group weaned successfully from invasive ventilation sooner than the other groups. NAVA, neurally adjusted ventilatory assist.

The multivariate Cox proportional hazard survival analysis demonstrated that the late NAVA group had a longer survival time (aHR 0.15, 95% CI 0.03–0.76, p = 0.022) than the conventional group (Figure 3 and Supporting Information, Table S3). Furthermore, female sex and a higher gestational age showed lower HRs for mortality.

FIGURE 3.

Adjusted survival curve. A multivariate Cox proportional hazard model was used. This curve shows that the late NAVA group survived longer than the conventional group. NAVA, neurally adjusted ventilatory assist.

DISCUSSION

Despite advances in neonatology and technology, the incidence of BPD has remained unchanged over the past three decades. ¹³ , ¹⁴ Tremendous efforts have been made to reduce ventilator‐induced lung injuries using new technology and device but no single study has demonstrated differences in mortality and respiratory morbidity rates related to the use of new ventilator technologies. ¹⁵

This is the first study to demonstrate that NAVA resulted in better clinical outcomes than other ventilation modes. Both the early and late NAVA groups had lower odds of death compared to the conventional group, even though two thirds of infants in the early NAVA group and all infants in the late NAVA group were exposed to other conventional modes. The incidence of BPD was significantly higher in the late NAVA group; however, because NAVA application was started at the median (IQR) 27 (22–37) days of life for them, we could assume that the infants in the late NAVA group already had evolving or established BPD. The infants in the late NAVA group had already received invasive ventilation with several conventional modes for approximately 4 weeks before switching to NAVA, so it cannot be asserted that the higher OR of BPD in the late NAVA group was an effect of NAVA application but rather that NAVA might be used as a rescue therapy or to wean from invasive mechanical ventilation in BPD patients. Perhaps as a result of rescue uses, there was no difference in the incidence of grade 2 or higher BPD between the conventional group and the late NAVA group, even though there was a significantly higher incidence of BPD in the late NAVA group (Supporting Information, Table S1). A longer duration of invasive mechanical ventilation is known to be related to higher mortality. ¹⁶ The important point here is that the longer duration of mechanical ventilation in the late NAVA group (median [IQR], 35 [22–37] vs. 7 [4–30] days, p = 0.002) did not increase the mortality rate but, rather, the survival of infants in the late NAVA group was significantly higher than that in the conventional group (aHR 0.15, 95% CI 0.03–0.76, p = 0.022). In this analysis, infants who were female and had a higher gestational age also exhibited lower HRs for mortality, which is consistent with neonatal outcome data demonstrated in previous studies. ¹⁷ , ¹⁸

We are the first to demonstrate that the early NAVA application from the first week of life, exhibited earlier weaning from invasive mechanical ventilation in comparison with the conventional modes of ventilation (aHR 2.02, 95% CI 1.14–3.57, p = 0.015). For this early NAVA group, NAVA was started at a median (IQR) 2 (1–4) days of life and a PMA 26⁺⁶ (25⁺⁴–28⁺³) weeks for infants born at 26⁺⁴ (25⁺²–28⁺³) weeks of gestation. This means that NAVA was used as a primary mode of ventilation as early as possible in the very first few days of life. Based on this result, we can assume that early application of NAVA before the lungs are damaged may be beneficial to accelerate extubation for preterm infants. In addition to previous studies showing physiological benefits, ⁵ , ⁶ , ⁷ , ¹⁰ a recent study reported that even the most immature infants triggered most of their breaths by their own respiratory efforts, and this control matured with increasing PMA. ¹⁹ In this regard, we can assume that the physiological advantages of NAVA can result in the evident clinical benefit when it is applied from the first few days after birth before the lungs are damaged. Our analysis also demonstrated that a higher gestational age and initial body temperature were associated with early weaning from mechanical ventilation, which aligns with previous studies showing that a lower gestational age and admission hypothermia are associated with higher pulmonary morbidity in preterm infants. ²⁰ , ²¹

This study has weaknesses inherent to retrospective studies. First, the selection of ventilation mode was made at the discretion of the physician in charge and there were frequent changes in the settings, modes, and ventilator machine during the respiratory support. It was therefore difficult to investigate the pure effect of NAVA with this retrospective study. Second, each group was heterogeneous in terms of the number of subjects, baseline characteristics and number of ventilation modes used. It is very difficult to perform a prospective study using only one mode of ventilation for sick preterm infants in real clinical situations, and it may be unethical to adhere to only one mode in this vulnerable population. As there are many variables regarding the ventilation mode, e.g., pressure, tidal volume, inspiratory time, and respiratory rate, various effects of these settings may occur even for one mode of ventilation. Third, this was a single center study with a small number of patients in each group. However, perhaps because this was a single center study, the policy of selecting a ventilation mode or adjusting ventilator settings would be consistent, which may reduce bias in generating the study results. Clinical outcomes, including BPD and death, are affected not only by mechanical ventilation modes but also by many other factors, such as inflammation, pulmonary congestion by intracardiac shunt, and nutrition. A single intensive care policy to manage these conditions in this center would have had the effect of reducing bias.

CONCLUSION

We showed that NAVA application was associated with lower odds of mortality, and its early application from the first few days of life helped preterm infants wean from invasive mechanical ventilation sooner compared to conventional modes of ventilation. To date, there is very little clinical evidence and few strategies to help determine optimal the ventilation mode for infants with respiratory distress in the first few days after birth and for those with evolving BPD later. In this sense, our results add meaningful evidence that NAVA may be a promising optimal ventilation mode to prevent ventilator‐induced lung injury during the early phase, decrease the severity of BPD and reduce mortality during the late phase of ventilatory care for preterm infants. More well designed large‐scale studies are needed to further evaluate the clinical benefit of NAVA and guide its application in preterm infants.

AUTHOR CONTRIBUTIONS

J.L. conceptualized and designed the study; Y.L and J.L. collected data and carried out analyses; Y.L. drafted the manuscript; J.L revised the manuscript. All authors read and approved the final version of the manuscript.

CONFLICT OF INTEREST STATEMENT

J.L. has given academic lectures in scientific conferences supported by the Getinge company. Y.L. declares no conflict of interest.

Supporting information

Data S1.

Lee Y, Lee J. Neurally adjusted ventilatory assist improves survival, and its early application accelerates weaning in preterm infants. Pediatr Int. 2024;66:e15831. 10.1111/ped.15831