Characterisation of the Oxygenation Response to Inspired Oxygen Adjustments in Preterm Infants

Transforming neonatal care with artificial intelligence: challenges, ethical consideration, and opportunities

Artificial intelligence (AI) offers tremendous potential to transform neonatology through improved diagnostics, personalized treatments, and earlier prevention of complications. However, there are many challenges to address before AI is ready for clinical practice. This review defines key AI concepts and discusses ethical considerations and implicit biases associated with AI. Next we will review literature examples of AI already being explored in neonatology research and we will suggest future potentials for AI work. Examples discussed in this article include predicting outcomes such as sepsis, optimizing oxygen therapy, and image analysis to detect brain injury and retinopathy of prematurity. Realizing AI's potential necessitates collaboration between diverse stakeholders across the entire process of incorporating AI tools in the NICU to address testability, usability, bias, and transparency. With multi-center and multi-disciplinary collaboration, AI holds tremendous potential to transform the future of neonatology.

Reducing the time delay of oxygen transport to the neonate on continuous positive airway pressure support: A bench study

Background

Premature newborns often require oxygen support as part of their therapy. Systems for oxygen administration are developed to assure adequate oxygenation of newborns. Several factors were identified in the systems that contribute to the time delay between the change in the set inspiratory oxygen fraction and its actual delivery to tissues. In this study, we aimed to reduce the physical delay in oxygen delivery to newborns.

Methods

We developed an O₂ Flush System (O₂-FS) that brings the source of oxygen as close to a patient as possible to make oxygen available for rapid delivery that compensates for the physical delay in the ventilator circuit. The O₂-FS system is built around an electromechanical on/off valve. We validated the O₂-FS concept in experiments with non-invasive Continuous Positive Airways Pressure (CPAP) ventilators.

Results

The O₂-FS accelerated oxygen delivery with all the tested systems and arrangements, typically by 5-15 s. We also observed that the application of supplemental oxygen increased the pressure in the ventilator circuit by 3-4 cmH₂O which may mitigate the apneic pauses that are common in premature newborns.

Conclusions

The O₂-FS system may work as a universal accessory of the CPAP lung ventilator and shorten the distribution of oxygen to the patient during oxygen desaturation events, possibly eliminating or interrupting apneic pauses in neonates, for whom oxygen therapy is an essential treatment. In clinical practice, the O₂-FS could help maintain normoxemic saturation values through adequate oxygen dosing in preterm neonates, thus reducing morbidity and mortality.

Comparison of two devices for automated oxygen control in preterm infants: a randomised crossover trial

OBJECTIVE

To compare the effect of two different automated oxygen control devices on target range (TR) time and occurrence of hypoxaemic and hyperoxaemic episodes.

DESIGN

Randomised cross-over study.

SETTING

Tertiary level neonatal unit in the Netherlands.

PATIENTS

Preterm infants (n=15) born between 24+0 and 29+6 days of gestation, receiving invasive or non-invasive respiratory support with oxygen saturation (SpO₂) TR of 91%-95%. Median gestational age 26 weeks and 4 days (IQR 25 weeks 3 days-27 weeks 6 days) and postnatal age 19 (IQR 17-24) days.

INTERVENTIONS

Inspired oxygen concentration was titrated by the OxyGenie controller (SLE6000 ventilator) and the CLiO₂ controller (AVEA ventilator) for 24 hours each, in a random sequence, with the respiratory support mode kept constant.

MAIN OUTCOME MEASURES

Time spent within set SpO₂ TR (91%-95% with supplemental oxygen and 91%-100% without supplemental oxygen).

RESULTS

Time spent within the SpO₂ TR was higher during OxyGenie control (80.2 (72.6-82.4)% vs 68.5 (56.7-79.3)%, p<0.005). Less time was spent above TR while in supplemental oxygen (6.3 (5.1-9.9)% vs 15.9 (11.5-30.7)%, p<0.005) but more time spent below TR during OxyGenie control (14.7 (11.8%-17.2%) vs 9.3 (8.2-12.6)%, p<0.05). There was no significant difference in time with SpO₂ <80% (0.5 (0.1-1.0)% vs 0.2 (0.1-0.4)%, p=0.061). Long-lasting SpO₂ deviations occurred less frequently during OxyGenie control.

CONCLUSIONS

The OxyGenie control algorithm was more effective in keeping the oxygen saturation within TR and preventing hyperoxaemia and equally effective in preventing hypoxaemia (SpO₂ <80%), although at the cost of a small increase in mild hypoxaemia.

TRIAL REGISTRY NUMBER

NCT03877198.

Automation of oxygen titration in preterm infants: Current evidence and future challenges

For the preterm infant with respiratory insufficiency requiring supplemental oxygen, tight control of oxygen saturation (SpO₂) is advocated, but difficult to achieve in practice. Automated control of oxygen delivery has emerged as a potential solution, with six control algorithms currently embedded in commercially-available respiratory support devices. To date, most clinical evaluations of these algorithms have been short-lived crossover studies, in which a benefit of automated over manual control of oxygen titration has been uniformly noted, along with a reduction in severe SpO₂ deviations and need for manual FiO₂ adjustments. A single non-randomised study has examined the effect of implementation of automated oxygen control with the CLiO₂ algorithm as standard care for preterm infants; no clear benefits in relation to clinical outcomes were noted, although duration of mechanical ventilation was lessened. The results of randomised controlled trials are awaited. Beyond the gathering of evidence regarding a treatment effect, we contend that there is a need for a better understanding of the function of contemporary control algorithms under a range of clinical conditions, further exploration of techniques of adaptation to individualise algorithm performance, and a concerted effort to apply this technology in low resource settings in which the majority of preterm infants receive care. Attainment of these goals will be paramount in optimisation of oxygen therapy for preterm infants globally.

Automated oxygen control in preterm infants, how does it work and what to expect: a narrative review

BACKGROUND

Automated oxygen control systems are finding their way into contemporary ventilators for preterm infants, each with its own algorithm, strategy and effect.

OBJECTIVE

To provide guidance to clinicians seeking to comprehend automated oxygen control and possibly introduce this technology in their practice.

METHOD

A narrative review of the commercially available devices using different algorithms incorporating rule-based, proportional-integral-derivative and adaptive concepts are described and explained. An overview of how they work and, if available, the clinical effect is given.

RESULTS

All algorithms have shown a beneficial effect on the proportion of time that oxygen saturation is within target range, and a decrease in hyperoxia and severe hypoxia. Automated oxygen control may also reduce the workload for bedside staff. There is concern that such devices could mask clinical deterioration, however this has not been reported to date.

CONCLUSIONS

So far, trials involving different algorithms are heterogenous in design and no head-to-head comparisons have been made, making it difficult to differentiate which algorithm is most effective and what clinicians can expect from algorithms under certain conditions.

In vitro evaluation of delays in the adjustment of the fraction of inspired oxygen during CPAP: effect of flow and volume

BACKGROUND

Adjusting the fraction of inspired oxygen (FiO₂) delivered to preterm infants to keep their oxygen saturation within target range remains challenging. Closed-loop automated FiO₂ control increases the time infants spend within the assigned target range. The delay with which FiO₂ adjustments at the ventilator result in a change in the inspired gas limits the performance of both manual and automated controls.

OBJECTIVE

To evaluate the equilibration time (T_eq) between FiO₂ adjustments and changes in FiO₂ reaching the patient.

METHODS

In vitro determination of the delay in FiO₂ adjustments at the ventilator at 5 and 8 L/min of gas flow and two different humidifier/ventilator circuit volumes (840 and 432 mL).

RESULTS

T_eq values were 31, 23, 20 and 17 s for the volume-flow combinations 840 mL+5 L/min, 840 mL+8 L/min, 432 mL+5 L/min and 432 mL+8 L/min, respectively.

CONCLUSION

The identified delay seems clinically relevant and should be taken into account during manual and automatic control of FiO₂.

Is faster better? A randomised crossover study comparing algorithms for closed-loop automatic oxygen control

OBJECTIVE

Closed-loop automatic control (CLAC) of the fractional inspired oxygen (FiO₂) improved oxygen administration to preterm infants on respiratory support. We investigated whether a revised CLAC algorithm (CLAC_fast, ≤2 FiO₂ adjustments/min), compared with routine manual control (RMC_only), increased the proportion of time with arterial haemoglobin oxygen saturation measured by pulse oximetry within prespecified target ranges (Target%) while not being inferior to the original algorithm (CLAC_slow: ≤0.3 FiO₂ adjustments/min).

DESIGN

Unblinded randomised controlled crossover study comparing three modes of FiO₂ control in random order for 8 hours each: RMC supported by CLAC_fast was compared with RMC_only and RMC supported by CLAC_slow. A computer-generated list of random numbers using a block size of six was used for the allocation sequence.

SETTING

Two German tertiary university neonatal intensive care units.

PATIENTS

Of 23 randomised patients, 19 were analysed (mean±SD gestational age 27±2 weeks; age at randomisation 24±10 days) on non-invasive (n=18) or invasive (n=1) respiratory support at FiO₂ >0.21.

MAIN OUTCOME MEASURE

Target%.

RESULTS

Mean±SD [95% CI] Target% was 68%±11% [65% to 71%] for CLAC_fast versus 65%±11% [61% to 68%] for CLAC_slow versus 58%±11% [55% to 62%] for RMC_only. Prespecified hypothesis tests of: (A) superiority of CLAC_fast versus RMC_only and (B) non-inferiority of CLAC_fast versus CLAC_slow with margin of 5% yielded one-sided p values of <0.001 for both comparisons.

CONCLUSIONS

This revised and faster CLAC algorithm was still superior to routine care in infants on respiratory support and not inferior to a previously tested slower algorithm.

TRIAL REGISTRATION NUMBER

NCT03163108.

Physiological instability after respiratory pauses in preterm infants

BACKGROUND

The factors influencing the severity of apnea-related hypoxemia and bradycardia are incompletely characterized, especially in infants receiving noninvasive respiratory support.

OBJECTIVES

To identify the frequency and predictors of physiological instability (hypoxemia-oxygen saturation (SpO₂ ) <80%, or bradycardia-heart rate (HR) < 100 bpm) following respiratory pauses in infants receiving noninvasive respiratory support.

METHODS

Respiratory pause duration, derived from capsule pneumography, was measured in 30 preterm infants of gestation 30 (24-32) weeks [median (interquartile range)] receiving noninvasive respiratory support and supplemental oxygen. For identified pauses of 5 to 29 seconds duration, we measured the magnitude and duration of SpO₂ and HR reductions over a period starting at the pause onset and ending 60 seconds after resumption of breathing. Temporally clustered pauses (<60 seconds separation) were analyzed separately. The relative contribution of respiratory pauses to overall physiological instability was determined, and predictors of instability were sought in regression analysis, including demographic, clinical and situational variables as inputs.

RESULTS

In total, 17 105 isolated and 9180 clustered pauses were identified. Hypoxemia and bradycardia were more likely after longer duration and temporally-clustered pauses. However, the majority of such episodes occurred after 5 to 9 second pauses given their numerical preponderance, and short-lived pauses made a substantial contribution to physiological instability overall. Birth gestation, hemoglobin concentration, form of respiratory support, caffeine treatment, respiratory pause duration and temporal clustering were identified as predictors of instability.

CONCLUSIONS

Brief respiratory pauses, especially when clustered, contribute substantially to hypoxemia and bradycardia in preterm infants.

Engineering Evaluation of the Performance of an Automatic Peripheral Oxygen Controller Using a Neonatal Respiratory Model

Premature infants often require respiratory support with a varying concentration of the fraction of inspired oxygen FiO2 to keep the arterial oxygen saturation typically measured using a peripheral sensor (SpO2) within the desired range to avoid both hypoxia and hyperoxia. The widespread practice for controlling the fraction of inspired oxygen is by manual adjustment. Automatic control of the oxygen to assist care providers is desired. A novel closed-loop respiratory support device with dynamic adaptability is evaluated nonclinically by using a neonatal respiratory response model. The device demonstrated the ability to improve oxygen saturation control over manual control by increasing the proportion of time where SpO2 is within the desired range while minimizing the episodes and periods where SpO2 of the neonatal respiratory model is out of the target range.

Development and preclinical testing of an adaptive algorithm for automated control of inspired oxygen in the preterm infant

OBJECTIVE

To assess the performance of a novel algorithm for automated oxygen control using a simulation of oxygenation founded on in vivo data from preterm infants.

METHODS

A proportional-integral-derivative (PID) control algorithm was enhanced by (i) compensation for the non-linear SpO₂-PaO₂ relationship, (ii) adaptation to the severity of lung dysfunction and (iii) error attenuation within the target range. Algorithm function with and without enhancements was evaluated by iterative linking with a computerised simulation of oxygenation. Data for this simulation (FiO₂ and SpO₂ at 1 Hz) were sourced from extant recordings from preterm infants (n=16), and converted to a datastream of values for ventilation:perfusion ratio and shunt. Combination of this datastream second by second with the FiO₂ values from the algorithm under test produced a sequence of novel SpO₂ values, allowing time in the SpO₂ target range (91%-95%) and in various degrees of hypoxaemia and hyperoxaemia to be determined. A PID algorithm with 30 s lockout after each FiO₂ adjustment, and a proportional-derivative (PD) algorithm were also evaluated.

RESULTS

Separate addition of each enhancing feature to the PID algorithm showed a benefit, but not with uniformly positive effects. The fully enhanced algorithm was optimal for the combination of targeting the desired SpO₂ range and avoiding time in, and episodes of, hypoxaemia and hyperoxaemia. This algorithm performed better than one with a 30 s lockout, and considerably better than PD control.

CONCLUSIONS

An enhanced PID algorithm was very effective for automated oxygen control in a simulation of oxygenation, and deserves clinical evaluation.