Utility of saturation trends to predict successful weaning of nasal CPAP in very preterm neonates

Abstract

BACKGROUND:

There is no objective criteria to wean CPAP in preterm neonates. We aimed to assess the accuracy of ‘saturation trends’ to predict successful CPAP discontinuation.

METHODS:

We included very preterm neonates who required CPAP. Index tests were ‘saturation trends’. Outcome was successful CPAP discontinuation, defined as baby stable in room air for 72 h.

RESULTS:

We had 120 neonates with mean±SD gestation 28.6±1.8 weeks. 96 (80%) neonates had successful discontinuation and 24 (20%) failed. Neonates with successful discontinuation had significantly greater ‘saturation trends’ during 24 h before discontinuing CPAP compared to those who failed [64.3 (48.1–83.7) vs. 47.3 (23.0–65.0), p = 0.001]. Saturations > 95% while on CPAP with 21% FiO₂ for > 60% time had 63% sensitivity and 70% specificity to predict successful CPAP discontinuation.

CONCLUSION:

‘Saturation trends’ is a readily available objective parameter that can be used to guide weaning CPAP in preterm neonates.

Keywords

Neonate predictive accuracy pulse-oximetry saturation trends

1 Background

Continuous positive airway pressure (CPAP) is the most commonly used respiratory support in preterm neonates. There are well-defined criteria for initiation of CPAP that are based on severity of respiratory distress assessed by several scoring systems, oxygen requirement and other parameters [1, 2]. However, there are no objective criteria to decide on weaning and discontinuation of CPAP in preterm neonates. Though maintenance of oxygen saturations within the target range is considered a major prerequisite to wean CPAP, it is often evaluated based on intermittent vitals monitoring performed by nurses.

Pulse-oximeters give ‘saturation trends’, which refers to the proportion of time, when saturations were in a particular range, during a specified period of time. This data is readily available in pulse-oximeters. In our unit, we regularly use ‘saturation trends’ as a guide to wean and discontinue respiratory support in neonates. However, on literature search, we could find only a few studies on the utility of saturation trends in newborn infants. Hence, we conducted this study to assess the accuracy of saturation trends to predict successful discontinuation of CPAP in preterm neonates.

2 Methods

This prospective observational diagnostic test accuracy evaluation study was conducted in a tertiary care neonatal unit in India from May 2022 to April 2023. We included all neonates who were born at <32 weeks’ gestation and required CPAP support at any postnatal age. Exclusion Criteria were (1) baby had to be changed from CPAP to mechanical ventilation or high-flow nasal cannula (HFNC), and (2) death or discharge against medical advice before CPAP discontinuation. We obtained Institutional Ethics Committee approval before commencement (EC/AP/890/04/2022). The study was prospectively registered in Clinical Trial Registry of India (CTRI/2022/04/041883). Informed written consent was obtained from parents prior to recruitment.

Saturation monitoring was performed using Masimo Rad-7 pulse-oximeter, which gives oxygen saturation trends during the previous 1–24 hours. Saturation trends during the selected time frame refer to the proportion of time the baby’s saturations are in ranges of <80, 80–85, 86–90, 91–95 and 96–100%. Index tests were saturation trends during (1) 0–24 h prior to CPAP discontinuation, (2) 24–48 h prior to CPAP discontinuation, and (3) 0–24 h after CPAP discontinuation. Outcome was successful discontinuation of CPAP, defined as baby remaining stable in room air for 72 h after discontinuation. Failure of CPAP discontinuation was diagnosed when baby had to be reconnected to CPAP within 72 h of discontinuation. Indications to restart CPAP support were (1) respiratory distress defined as Downe’s score≥3, (2) any apnoea requiring positive pressure ventilation, or ≥2 episodes of apnoea requiring tactile stimulation or reverting on its own, and (3) repeated episodes of desaturations to <85%, defined as ≥3 episodes in 6 h. Blood gas analysis was not routinely performed for all neonates after CPAP discontinuation.

We routinely use bubble CPAP delivered through short binasal prongs. CPAP is usually started and continued at 6 cm/H₂O. Fraction of inspired oxygen (FiO₂) is titrated real time to maintain the saturations within the target range of 91–95%. Weaning from CPAP is initiated when the baby is stable on 21% FiO₂ for ∼24 h. Weaning is performed by cycling on and off CPAP. The duration of off-periods is gradually increased as 2, 4, 8 and 24 h depending on the tolerability of the neonate.

Saturation trends during the on- and off-CPAP periods were entered separately in pre-designed study proforma. Saturation trends prior to CPAP discontinuation refer to the saturations during the on-CPAP periods.

The consultant and trainee doctors and the nurses who were involved in decision-making and care of the neonates were masked to the saturation trends. Other baseline and demographic characteristics and clinical details of the study infants were collected prospectively.

2.1 Sample size and Statistical analysis

For 90% success rate of CPAP discontinuation based on our data, and an assumed 90% sensitivity, 7% precision, and 10% attrition, the sample size required was 120.

Descriptive data are presented as mean and standard deviation (SD), median and interquartile range (IQR) or number and percentage, as appropriate. Mann-Whitney’s U test was used to compare the ‘saturation trends’ between the groups. Receiver operating characteristic (ROC) curves were used to establish the best cut-off values for the saturation trends data. The optimal cut-offs obtained from the ROC were used to calculate the sensitivity, specificity, positive predictive value, negative predictive value and likelihood ratios (LR+ and LR–). All statistical analyses were performed using SPSS 23.0. A p value less than 0.05 was considered statistically significant.

3 Results

We recruited 120 neonates (eFigure 1). The mean±SD gestational age was 28.6±1.8 weeks and birth weight was 1123±323 g (Table 1). 105 (87.5%) neonates required CPAP at birth for respiratory distress syndrome or transient tachypnea of newborn, while 15 (12.5%) required CPAP later for other indications. 96 (80%) neonates had successful discontinuation and 24 (20%) had failure of CPAP discontinuation.

Table 1
Baseline characteristics

n =

Gestational age; weeks; mean (SD) 28.6 (1.8)

Birth weight; g; mean (SD) 1123 (323)

Gender, boys; n (%) 64 (53.3)

APGAR 1 min; median (IQR) 6 (5, 7)

APGAR 5 min; median (IQR) 8 (7, 8)

Initiation of CPAP, n (%)

On day 1 105 (87.5)

Later 15 (12.5)

Duration of CPAP; days; median (IQR) 9 (4, 31.5)

	n =
Gestational age; weeks; mean (SD)	28.6 (1.8)
Birth weight; g; mean (SD)	1123 (323)
Gender, boys; n (%)	64 (53.3)
APGAR 1 min; median (IQR)	6 (5, 7)
APGAR 5 min; median (IQR)	8 (7, 8)
Initiation of CPAP, n (%)
On day 1	105 (87.5)
Later	15 (12.5)
Duration of CPAP; days; median (IQR)	9 (4, 31.5)

Neonates with successful CPAP discontinuation had significantly greater ‘saturation trends’ compared to those who failed (Table 2). Based on area under ROC curve (AUC), the best predictors for successful CPAP discontinuation were (1) proportion of time when saturations were >95% during 0–24 h before CPAP discontinuation, while on CPAP with 21% FiO₂ [64.3 (48.1–83.7) vs. 47.3 (23.0–65.0); AUC 0.723 (0.625–0.821); p = 0.001]. The best cut-off of 60% had 63% sensitivity and 70% specificity (eFigure 2, Table 3), and (2) proportion of time with saturations > 95% during 0–24 h after CPAP discontinuation [56.4 (41.3–69.8) vs. 33.8 (27.4–49.9); AUC 0.763 (0.622–0.904); p = 0.006]. The best cut-off of 50% had 61% sensitivity and 80% specificity.

Table 2

Comparison of saturation trends between neonates with successful and failure of weaning CPAP

Saturation trends	Successful weaning (n = 96)	Failure of weaning (n = 24)	P value^*	AUC of ROC curve
24 h prior to CPAP discontinuation, while on 21% FiO₂
>95	64.3 (48.1, 83.7)	47.3 (23.0, 65.0)	0.001	0.723 (0.625, 0.821)
>90	94.1 (89, 97.3)	92.1 (84.5, 94.3)	0.007	0.681 (0.557, 0.805)
<80	0.1 (0, 0.6)	0.3 (0, 1.25)	0.004	0.691 (0.550, 0.832)
24-48 h prior to CPAP discontinuation, while on 21% FiO₂
>95	63.2 (46.2, 76.9)	45.4 (30.6, 57.0)	0.049	0.636 (0.514, 0.758)
>90	94.95 (89, 98.1)	90.7 (85.1, 98.3)	0.012	0.674 (0.549, 0.799)
<80	0.2 (0, 0.6)	0.85 (0, 3.18)	0.004	0.697 (0.567, 0.826)
24 h after CPAP discontinuation
>95	56.4 (41.3, 69.8)	33.8 (27.4, 49.9)	0.006	0.763 (0.622, 0.904)
>90	94.95 (90.3, 98.2)	87.5 (82, 93.7)	0.043	0.689 (0.489, 0.902)
<80	0.2 (0, 0.6)	2 (0.08, 5.8)	0.040	0.693 (0.484, 0.902)

^*Mann-Whitney U test.

Table 3

Markers of diagnostic accuracy of saturation trends

Proportion of time when oxygen saturations > 95%	Sensiti-vity	Specifi-city	Positive predictive value	Negative predictive value	Likelihood ratio+	Likelihood ratio –
During 24 h prior to CPAP discontinuation, while on 21% FiO₂	62.1	69.6	89.4	30.8	2.0	0.5
During 24 h after CPAP discontinuation	60.0	80.0	96.6	17.4	3.0	0.5

4 Discussion

Our study shows that saturation trends from electronic patient monitoring has good accuracy to predict successful discontinuation of CPAP in very preterm neonates.

The readiness to wean and discontinue CPAP in preterm neonates is often evaluated based on subjective assessments by clinicians and feedback from nurses. Though maintenance of oxygen saturations within the target range is an important aspect of this assessment, it is evaluated based on intermittent monitoring, which may under or overestimate the amount of time spent outside the target range [3]. ‘Saturation trends’ from electronic patient monitoring is an objective and continuous measurement, and is readily available from the pulse-oximeter. Infants who have better saturation trends are likely to have better lung compliance and gas exchange, and hence are more likely to achieve successful CPAP discontinuation.

In a similar study, Gentle et al showed that preterm neonates who had successful CPAP discontinuation had significantly better saturation trends compared to those who failed [4]. Saturations > 95% for 31.6% time had 75% sensitivity and 75% specificity to predict successful CPAP discontinuation. Mascoll-Robertson and colleagues evaluated saturation trends to predict transition from CPAP or HFNC/room air/low flow oxygen in very preterm infants [5]. Infants who failed had saturations≤85% for a greater proportion of time compared to those who had successful transition (10.7±11.9 vs. 3.3±4.7%; p = 0.02). The differences in sensitivity and specificity for ‘saturation trends’ between our study and previous studies could be due to differences in target saturations, criteria used for weaning CPAP and definition used for success/failure of CPAP discontinuation.

Data from electronic patient monitoring is being explored to predict morbidities and mortality in neonates [6]. Heart rate characteristics index was found to predict late-onset sepsis and necrotising enterocolitis [7, 8]. Similarly, heart rate and oxygen saturations data downloaded from pulse-oximeters was shown to predict need for NICU admission, intraventricular hemorrhage, bronchopulmonary dysplasia and mortality [9, 10].

This is the largest study to evaluate the accuracy of saturation trends to predict successful CPAP discontinuation in preterm neonates. The drawbacks of using ‘saturation trends’ are (1) only some pulse-oximeters give ‘saturation trends’, and (2) babies may have transient false low saturations due to movement artefacts, which was not evaluated in this study. Though currently available pulse-oximeters reduce movement artifacts and we used disposable probes and changed it to a newer probe whenever necessary thereby limiting missed data points, this problem could not have been fully eliminated. It is essential to maintain good probe attachment and reduce movement artefacts to improve the predictive accuracy of saturation trends.

To conclude, ‘saturation trends’ is a readily available objective parameter that can be used as a guide to wean and discontinue CPAP in preterm neonates. Neonates with successful CPAP discontinuation had significantly greater ‘saturation trends’ compared to those who failed. The best predictor for successful CPAP discontinuation was the proportion of time with > 95% saturations while on CPAP with 21% FiO₂. The predictive accuracy of ‘saturation trends’ needs further evaluation in future studies with larger sample size.

Funding

None.

Conflict of interest

None.

Footnotes

Acknowledgments

None.

The supplementary material is available in the electronic version of this article: .

References

Downes

, Vidyasagar

, Morrow

, Boggs

. Respiratory distress syndrome of the newborn infant. Clin Pediatr. 1970;9(6):325–32.

Silverman

, Andersen

. A controlled clinical trial of effects of water mist on obstructive respiratory signs, death rate and necropsy findings among premature infants. Pediatrics. 1956;17:1–10.

Ruiz

, Trzaski

, Sink

, Hagadorn

. Transcribed oxygen saturation vs oximeter recordings in very low birth weight infants. J Perinatol. 2014;34(2):130–5.

Gentle

, Ambalavanan

, Carlo

. Oxygen saturation histograms predict nasal continuous positive airway pressure-weaning success in preterm infants. Pediatr Res. 2020;88(4):637–641.

Mascoll-Robertson

, Viscardi

, Woo

. The objective use of pulse oximetry to predict respiratory support transition in preterm infants: An observational pilot study. Respir Care. 2016;61(4):416–22.

Kumar

, Akangire

, Sullivan

, Fairchild

, Sampath

. Continuous vital sign analysis for predicting and preventing neonatal diseases in the twenty-first century: Big data to the forefront. Pediatr Res. 2020;87(2):210–20.

Coggins

, Weitkamp

, Grunwald

, Stark

, Reese

, Walsh

, et al. Heart rate characteristic index monitoring for bloodstream infection in an NICU: A 3-year experience. Arch Dis Child Fetal Neonatal Ed. 2016;101(4):F329–32.

Zeigler

, Ainsworth

, Fairchild

, Wynn

, Sullivan

. Sepsis and mortality prediction in very low birth weight infants: Analysis of HeRO and nSOFA. Am J Perinatol. 2023;40(4):407–14.

Solís-García

, Maderuelo-Rodríguez

, Perez-Pérez

, Torres-Soblechero

, Gutiérrez-Vélez

, Ramos-Navarro

, et al. Longitudinal analysis of continuous pulse oximetry as prognostic factor in neonatal respiratory distress. Am J Perinatol. 2022;39(6):677–82.

10.

Sullivan

, Wallman-Stokes

, Isler

, Sahni

, Moorman

, Fairchild

, et al. Early pulse oximetry data improves prediction of death and adverse outcomes in a two-center cohort of very low birth weight infants. Am J Perinatol. 2018;35(13):1331–8.

Utility of saturation trends to predict successful weaning of nasal CPAP in very preterm neonates – A prospective study

Abstract

BACKGROUND:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1 Background

2 Methods

2.1 Sample size and Statistical analysis

3 Results

Funding

Conflict of interest

Footnotes

Acknowledgments

References