Ventilatory Support, Extubation, and Cerebral Perfusion Changes in Pre-Term Neonates: A Near Infrared Spectroscopy Study

Early extubation is considered to be beneficial for pre-term neonates. On the other hand, premature extubation can cause lung derecruitment, compromised gas exchange, and need for reintubation, which may be associated with severe brain injury caused by sudden cerebral blood flow changes. We used near infrared spectroscopy (NIRS) to investigate changes in cerebral oxygenation (rScO2) and fractional tissue oxygen extraction (+) after extubation in pre-term infants. This is a single-center retrospective study of NIRS data at extubation time of all consecutive pre-term neonates born at our institution over a 1-year period. Comparison between subgroups was performed. Nineteen patients were included; average gestational age (GA) was 29.4 weeks. No significant change was noted in rScO2 and cFTOE after extubation in the whole population. GA and germinal matrix hemorrhage (GMH)-intraventricular hemorrhage (IVH) showed a significant change in rScO2 and cFTOE after extubation. A significant increase in cFTOE was noted in patients with previous GMH-IVH (+0.040; p = 0.05). To conclude, extubation per se was not associated with significant change in cerebral oxygenation and perfusion. Patients with a diagnosed GMH-IVH showed an increase in cFTOE, suggesting perturbation in cerebral perfusion suggesting further understanding during this challenging phenomenon. Larger studies are required to corroborate our findings.


Introduction
Extubating a pre-term infant is a challenging decision, especially during the first few days of life when there is a high risk of germinal matrix hemorrhage (GMH)intraventricular hemorrhage (IVH).GMH-IVH is the most frequent lesion affecting infants born before the gestational age (GA) of 34 weeks, with incidence rates as high as 20-30%, 1 and with variable outcomes.3][4] Pathogenesis of IVH is multifactorial, and is closely linked to vascular immaturity within the germinal matrix and fluctuation of cerebral blood flow (CBF). 5,6Pre-term neonates with low GA typically exhibit impaired capacity for autoregulation of CBF. 7,8These neonates are particularly sensitive to CBF fluctuations, especially in cases of respiratory distress, 9 pneumothorax, 10 and mechanical ventilation, which is a known risk factor for GMH-IVH. 11arly extubation is considered to improve neurological outcomes of pre-term neonates. 12Prolonged mechanical ventilation exposes the pre-term infant to several adverse outcomes, including a higher risk of developing bronchopulmonary dysplasia 13 and neurodevelopmental impairment. 14On the other hand, premature inappropriate extubation can cause lung derecruitment, compromise gas exchange, and increase respiratory fatigue, necessitating reintubation. 15Reintubation may potentially increase hemodynamic perturbations, especially those affecting CBF. 16This phenomenon may be aggravated by multiple attempts at reintubation, which is known to increase the risk of developing severe GMH-IVH. 17ear infrared spectroscopy (NIRS) is a potential tool to monitor the hemodynamic status in the neonatal brain.This technology can allow for continuous and non-invasive monitoring of cerebral tissue oxygenation (rScO 2 ), even in most vulnerable infants. 18Starting from cerebral rSO 2 and peripheral saturation (SpO 2 ), it is possible to calculate the cerebral fractional tissue oxygen extraction (cFTOE), which is a better index of CBF than rScO 2 , especially in cases of low arterial oxygen levels (SpO 2 ). 19The role of NIRS is still uncertain in pre-term neonatal intensive care, but is used as clinical practice in other contexts such as pediatric cardio-surgery monitoring; however, many studies have demonstrated that NIRS is useful for detecting different alterations that are potentially associated with brain injury, and it is widely used in many neonatal intensive care units (NICU). 20,21In our unit, NIRS has been part of standard monitoring in pre-term infants in the first hour of life since 2019.The levels of rScO 2 and cFTOE depend on many variables such as blood pressure, patent ductus arteriosus (PDA), and anemia; in addition, these indices especially depend on respiratory factors such as SpO 2 , pCO 2 , and intrathoracic pressure, 22 which are strictly linked to the ventilation mode. 23n this observational study, we evaluate any postextubation changes in cerebral oxygenation and perfusion in a pre-term population.

Methods
We retrospectively reviewed all consecutive patients who were monitored using NIRS at our institution between October 2021 and October 2022.The inclusion criteria were: GA <33 weeks, need for ventilatory support (either nasal continuous positive airway pressure [CPAP] or endotracheal tube [ETT]) at birth (within the first 5 h of life) and withdrawal of ventilatory support within the first 120 h of life (i.e., 5 days).
The decision to stop the ventilatory support (i.e., extubation) was in accordance with the institutional protocol based on ventilatory parameters; namely, peak inspiratory pressure (PIP) £21 cm H 2 O, positive end expiratory pressure (PEEP) £6 cm H 2 O, an FiO 2 £ 25%, and good respiratory drive.In all cases, the extubation maneuver was performed by maintaining the NIRS sensor in place, and aspiration of secretions from the endotracheal tube was allowed at the discretion of the treating physician.Following extubation, CPAP or BiPAP was used as non-invasive ventilatory support based on the discretion of the treating physician.
All patients underwent NIRS monitoring as per internal protocol using the MASIMO ROOT system (Masimo Corporation, Irvine, CA, USA); the MASIMO Radical 7 device was used to monitor peripheral blood oxygen saturation (SpO 2 ), whereas the MASIMO O3 sensor was used to monitor cerebral regional oxygen saturation (rScO 2 ).The MASIMO O3 sensor provides readings every 2 sec.It has been approved for use in pediatric patients and has a reported a bias, standard deviation (SD), standard error (SE), and root mean squared error (RMSE) for rSO 2 of 2.6%, 4.5%, 0.3%, and 4.3%, respectively, and a trend accuracy with a relative mean error of -1.4%, SD of 4.3%, SE of 0.2%, and RMSE of 3.9%. 24In all patients, the adhesive sensor for cerebral rScO 2 monitoring was attached to the left frontoparietal region, and monitoring was started within the first 5 h of life.Peripheral SpO 2 monitoring was performed in the right upper limb.For the current study, rScO 2 and SpO 2 data within 1 h before and 1 h after the extubation were collected.To minimize errors caused by movement artifacts, rScO 2 and SpO 2 continuous readings were averaged over a 10-min time frame within 1 h before and after extubation.Such a brief time frame and a similar approach were already validated in other studies. 25The chosen time frame was based on the best stability of the rScO 2 data (lower SD), greater availability of data, and best proximity to the extubation.Readings pertaining to the 15 min immediately before and after extubation were excluded from the current analysis (Figure 1).In addition to rScO 2 and SpO 2 , post-extubation change in rScO 2 (DrScO 2 ) and SpO 2 (DSpO 2 ) were also calculated (i.e., rScO 2 and SpO 2 after extubation -rScO 2 and SpO 2 before extubation, respectively), and cFTOE was calculated using the formula: SpO 2 -ScO 2 ) / SpO 2 .The cFTOE is a useful index for assessment of cerebral regional O 2 extraction and perfusion, which are independent of the variations in peripheral SpO 2 .
Clinical charts of all patients were reviewed, and general demographics along with GA, birth weight, 1-min and 5-min Apgar scores, presence (and treatment) of PDA at 48-72 h, need for inotropic therapy, ventilatory parameters before and after extubation (i.e., PIP, PEEP, and FiO 2 ) were collected.All patients underwent brain ultrasound examination as per the institutional protocol at days 1, 2, and 3 after birth, and once a week thereafter until 36 weeks of age.Further, all patients underwent cerebral magnetic resonance imaging (MRI) at term age with a 3 Tesla MR system using an internal protocol that includes T1, T2, and susceptibility weighted imaging (SWI) sequence, which is highly sensitive to hemosiderin deposits. 26resence of GMH-IVH was reported, and graded according to Volpe classification. 27Patients were divided into subroups according to clinical characteristics that could influence cerebral perfusion (GA < 28 weeks, birth weight <1000 g, significant PDA in treatment, presence of GMH-IVH, extubation timing <24h, and need for inotropic therapy). 20

Statistical analysis
Continuous variables were expressed as average -SD (range) unless stated otherwise; categorical variables were expressed as frequency (percentage).Preand post-extubation NIRS parameters were compared using the Wilcoxon test.Mann-Whitney U test was used for comparison between subgroups.P levels £0.05 were considered indicative of statistical significance.Data were analyzed using the SPSS Statistics software, version 23 (SPSS, Chicago, IL, USA) and Microsoft Office Excel 2016 Professional (Microsoft, Redmond, WA, USA).
The studies involving human participants were reviewed and approved by Giannina Gaslini Hospital, Genoa, Italy.Parental informed consent was required and obtained before recruitment.

Results
Out of 46 newborns who underwent NIRS monitoring at our institution during the study period, 27 patients were excluded because they were not extubated during the monitoring period and 6 patients had movement artifacts.A total of 19 patients (14 males) were included in the current study.Mean GA at birth was 29.4 -2.39 weeks (range 24.9-32.6)and average birth weight was 1353 -472 g (range 680-2170).On ultrasound screening, there were five cases of GMH-IVH (26.3%), two infants with stage I hemorrhage limited to the germinal matrix (IVH I, 10.5%), and three infants with stage II IVH (15.8%).Minor lesions in the periventricular white matter were confirmed in all three patients with stage II IVH on MRI performed at term-corrected age; SWI sequences (SWI + periventricular white matter lesions [PWML]) were positive for the presence of hemosiderin deposits in all three infants.All IVH cases were confirmed on ultrasound screening prior to extubation.A significant PDA requiring treatment was observed in six patients (31.6%).PDA was treated with paracetamol in three cases, ibuprofen in two cases, and both drugs in one case; none of the patients underwent surgical treatment.None of the patients were treated with inotropes (Table 1).

Discussion
Transition from invasive mechanical ventilation to non-invasive ventilatory support is a critical event in the life of pre-term infants.Early extubation or extubation failure exposes pre-term neonates to the risk of apnea, hypercapnia, and need for reintubation.These factors can lead to acute fluctuations in cerebral perfusion and increase the risk of IVH. 9,28On the other hand, prolonged or excessive mechanical ventilation can lead to hypocapnia and vasoconstriction leading  to cerebral hypoperfusion and inflammation of white matter and increased risk of periventricular leukomalacia (PVL) and impaired brain development. 29,30Several guidelines and clinical criteria have been developed to help clinicians decide on the correct timing of extubation, which is typically exclusively based on the analysis of respiratory difficulties.Nevertheless, the decision making is often challenging, and it is a matter of intense debate among neonatologists.
In our study, no significant changes in rScO 2 and cFTOE were found after extubation when looking at the whole population.This finding confirms what was previously described by Ericksen and coworkers in a smaller population of 12 pre-term patients. 25eripheral SpO 2 levels remained substantially stable (i.e., between 91% and 95%) throughout the study period, demonstrating that the extubation maneuver was well tolerated. 31There was no post-extubation brain bleeding event and we recorded only one case of failed extubation, which had no clinical consequences for the patient.No correlation was detected between extubation timing and DrScO 2 or DcFTOE, even in the setting of early extubation.Nevertheless, in patients with GA <28 weeks, extubation was associated with a significant decrease in regional brain SaO 2 compared with infants with GA ‡28 weeks (DrScO 2 À4.6 -3.4 vs. 2.2 -4.7, p = 0.01).However, there was no significant difference between the two groups with respect to cFTOE index.Decrease in rScO 2 in extremely pre-term infants can be explained by impaired cerebral autoregulation in these patients.Wong and coworkers showed how tissue oxygenation index and mean arterial blood pressure have high correlation (as a sign of lack autoregulation) in patients with lower GA. 8urther, Roche-Labarbe and coworkers also observed lower levels of cerebral oxygenation during the first 7 weeks of life of infants with GA <31 weeks using frequency domain NIRS monitoring, and the same authors also reported lower blood flow index in infants with GA between 24 and 27 weeks. 32It is unlikely that this may be the result of an increase in cerebral metabolism, as brain metabolism has been shown to increase with GA. 33 On the other hand, newborns with GMH-IVH showed a significant change in the cFTOE index following extubation (DcFTOE 0.040 -0.035 vs. À0.010-0.055, p = 0.05).Increase in the cFTOE index can be considered to be a sign of decreased brain perfusion. 19erhagen and coworkers demonstrated that pre-term infants with GMH-IVH or periventricular hemorrhagic infarction (PVHI) had lower rScO 2 and higher cFTOE during the first 2 weeks after birth than infants without GMH-IVH.Lower rScO 2 and higher FTOE occurred irrespective of the grade of GMH-IVH. 34his was confirmed by Lin and coworkers, who focused on low-grade GMH-IVH, and by Vesoulis and coworkers more recently. 35,36Because FTOE reflects the balance between cerebral oxygen supply and cerebral oxygen consumption, increased FTOE can be explained either by reduced oxygen supply or increased oxygen consumption.A lower oxygen supply may result from lower CBF. 32Our study highlighted worsening of the cFTOE index following extubation, which is a potential sign of further impaired CBF autoregulation in these patients. 7,33,37NIRS evaluates the oxygenation of most superficial cerebral tissue (* 2 cm of depth from the skin). 38,39GMH-IVH bleeding is located in the deep regions of the brain; therefore, this correlation suggests an influence on oxygenation and perfusion of more superficial areas of the brain, as previously demonstrated with MRI techniques by our group. 381][42] This further effect on the frontal areas following extubation in infants already having GMH-IVH may have greater adverse effect on the development of local neural circuitry.In addition, if a similar effect is likely to happen also in the posterior part of the brain, where brain maturation is known to be more advanced than in the frontal part, 43,44 impaired development of the visual cortex may also be postulated, although NIRS is unlikely to allow assess-ment of the posterior part of the brain for technical reasons.Vasospasm secondary to the increase in locally produced pro-inflammatory cytokines has been cited as the underlying mechanism for cerebral cortex hypoperfusion associated with GMH-IVH. 45This could be part of a complex pathophysiological pattern that includes other pathogenic noxae such as increase in pro-inflammatory factors and reduced cell proliferation of the germinative matrix. 46,47This may also help explain why even low-grade bleeding is associated with impaired neurological outcomes 3,4 and impaired periventricular white matter maturation in pre-term infants. 2,48,49Further, the positioning of the NIRS sensor in our study was over the front-parietal portions of the brain that are known to exhibit delayed maturation compared with posterior portions. 50Reduction in CBF in these more immature areas could have a greater impact on brain maturation.At least, in our population, three out of the five patients with GMH-IVH also had PWML with a hemorragic appearance (positive on SWI at MRI), which are minor lesions of white matter. 51hese lesions have been described to be more represented in cases of GMH-IVH, 51 and may be correlated with altered venous blood flow and perfusion in these patients.
Some limitations of our study should be considered.The relatively small sample size may affect the generalizability of our results, in particular for subgroup analysis, where only a limited number of patients had GMH-IVH and low GA.Although rScO 2 data are known to be affected by blood pressure variations, it is worth noting that no inotropes were used in our patients and that blood pressure values were stable throughout the study period.Finally, although we detected significant changes in the cFTOE index after extubation, our monitoring data were limited to few minutes after extubation.

Conclusion
Our study shows that NIRS monitoring can help detect subclinical cerebral hypoperfusion events in pre-term infants.Our cohort confirms that extubation was not related to significant CBF fluctuation, even in the setting of early extubation.A significantly increased cerebral tissue extraction fraction of oxygen caused by reduction of brain perfusion was observed following extubation only in those patients with GMH-IVH.This is an important finding, as extubation can aggravate brain hypoperfusion described to follow GMH-IVH.Further, without NIRS monitoring, these hypoperfusion events would have been completely undetected because of the stability of the other parameters.These findings suggest the need for a more cautious approach when mechanically ventilated newborns with GMH-IVH are considered for extubation because of changes in ventilatory strategies or improvements in respiratory conditions.Larger studies are required to confirm these findings.

FIG. 1 .
FIG.1.Squares show the time frame selected for the analysis.Vertical line represents the extubation time.

Table 1 .
Demographic and Baseline Characteristics of the Study Population Data are expressed as averagestandard deviation (range) unless stated otherwise.IVH, intraventricular hemorrhage; GMH, germinal matrix hemorrhage; ETT, endotracheal tube; PIP, peak inspiratory pressure; PEEP, positive end-expiratory pressure; FiO2, fraction of inspired oxygen.