Introduction

Neonates requiring intensive care are exposed to a variety of interventions and treatments that improves survival but can affect their neurological system. Epidemiological studies indicate that every year 1.15 million infants suffer a hypoxic-ischemic event at birth and 13 million are born prematurely [1–5]. Of these, a substantial number will develop moderate or severe neurological impairment. Furthermore, other disorders such as congenital heart disease, cerebral malformations, congenital and nosocomial infections, intraventricular hemorrhages, inborn errors of metabolism, seizures, and severe hemodynamic and ventilatory instability can compromise brain integrity and function [6–8]. Modern neonatal intensive care units (NICU) can provide strategies to prevent neurological impairment using standardized clinical protocols and training, neuromonitoring, and neuroimaging; the so called neonatal neurocritical care units or neuroNICUs.

Amplitude Integrated Electroencephalography (aEEG) is a non-invasive methodology for continuous monitoring of brain function at the bedside, specifically for determination of electrical background activity and sleep-wake cycle (SWC) patterns, and detection of seizures [9, 10]. Indeed, the use of a two-channel aEEG with raw EEG readings (aEEG/EEG) has showed improved accuracy for the detection of seizures on real-time [11, 12], an important effect as seizure burden is an independent risk factor for neurodevelopmental delay [13–16] and treatment of subclinical seizures is associated with improved neurodevelopmental outcomes [17]. Moreover, in infants with hypoxic ischemic encephalopathy (HIE), aEEG findings have good predictive value for neurological short- and long-term outcomes [18–21].

Another available methodology for brain monitoring at the bedside is the Near Infrared Spectroscopy (NIRS), a non-invasive tool that provides continuous measurements of cerebral tissue oxygenation and perfusion (rSO2). NIRS evaluates the balance between tissue oxygen delivery and consumption allowing the early identification of perfusion abnormalities and hemodynamic changes [22, 23]. Therefore, NIRS monitoring is important in several clinical scenarios including HIE, prematurity, shock, anemia, congenital heart disease and other cases that lead to hemodynamic and ventilatory instability [24–26].

According to the World Health Organization (WHO), telemedicine is defined as the delivery of health care services, where distance is a critical factor, by all health care professionals using information and communication technologies for the exchange of valid information for diagnosis, treatment and prevention of disease and injuries, research and evaluation, and for the continuing education of health care providers, all in the interest of advancing the health of individuals and their communities [25]. WHO also listed four important elements to telemedicine: provide clinical support; intention to overcome geographic barriers; use of several types of information and communication technologies; and improve health outcomes [26]. This model portrays the natural evolution of healthcare in the digital era, resulting in reduction of the distance between two or more locations, promoting access to specific methodologies in addition to reducing the cost of structure. It can also work through live interactive audiovisual links, live video transmission or online viewing of stored education material [27–29].

Brazil is the largest South America country with approximately 3 million live births per year and a significant proportion of infants at high-risk for brain injury [30]. Like any middle-income country, the available resources for neonatal care are unequal. Therefore, telemedicine can be used as a tool for education, consultation, patient care and research. This model allows low-cost access of remote and underserved areas to specific methodologies and expertise [27–29]. Considering that less than 5% of Brazilian NICUs provide well-structured neurocritical care [31], and the implementation of that requires specialized resources, adequate equipment, and qualified personnel, the use of an advanced and protocolized telemedicine model may be a better solution to decrease these disparities. For that, in order to provide neurocritical care to a broader range of Brazilian NICUs, a protocolized telemedicine system called Protecting Brains and Saving Futures (PBSF) Guidelines was developed and implemented in some NICUs. The goals of the PBSF Guidelines are to establish remote neurocritical care units by providing adequate equipment, specialized training, implementation of standardized protocols, continuous neuromonitoring and expert support, quickly and at low cost.

The aim of this prospective multicenter and observational study is to describe the centers’ adherence to the PBSF Guidelines and clinical outcomes over a 3-years period. We hypothesize that the adherence to the PBSF Guidelines will increase over time and this will be associated with improvement of clinical outcomes. Results of this study may provide the necessary background information on the use of telemedicine aiming to protect neonatal brains and save futures, and for the development of larger studies and initiatives.

Materials and methods

Ethics approval and consent to participate

The study followed the precepts of good clinical practice and has been performed in accordance with the Declaration of Helsinki. The study proposal was approved by the Research Ethics Committee of the Irmandade da Santa Casa de Misericórdia de São Paulo under the following approval numbers and dates: 3.142.318 on February 12^th, 2019 (S1 and S2 Files); 3.357.239 on May 30^th, 2019 (S3 and S4 Files); and 3.506.106 on August 13^th, 2019 (S5 and S6 Files). In addition, it received formal authorization from the Clinical and Administrative Board of each center. The original protocol approved by ethics is available at S7 File. The written informed consent from parents or guardians will be obtained by trained research investigators prior to inclusion in the study. The same care under the PBSF Guidelines will be provided to patients who decline study entry. All data will be treated anonymously and confidentially, and in no phase of the study any name, image, or data that allows participants identification will be disclosed. The study was registered in ClinicalTrials.gov (NCT03786497). It is reported using the Standard Protocol Items Recommendations for Interventional Trials (SPIRIT). The SPIRIT schedule is showed in Fig 1 and the completed SPIRIT checklist is available in S1 Appendix. Any proposed amendments will be discussed with the Research Ethics Committee and each institution’s Clinical and Administrative Board and communicated with all investigators, sponsor, trial participants and trial registry (Clinicaltrials.gov).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schedule of enrollment, interventions, and assessments.

https://doi.org/10.1371/journal.pone.0262581.g001

Discussion

This large prospective, multicenter, and observational study will describe the use and adherence to a unique telemedicine neonatal neurocritical care guidelines and clinical outcomes. As a multidisciplinary approach to care for newborns with high-risk of brain injury has increased over the last 10 years, the implementation of an organized, low-cost, and high-expertise telemedicine guidelines has the potential to streamline clinical care improving outcomes.

The PBSF Guidelines defines high-risk neonates, with clear indications for neonatal neurocritical care. Among the etiologies associated with impaired neurodevelopmental outcomes, neonatal asphyxia represents the third cause of neonatal death worldwide (23%) [2, 3, 32–34]. In newborns with moderate or severe HIE the risk of death, cerebral palsy and/or disability among survivors is high. In addition, some disability might also occur in mild HIE infants [4, 6, 7, 35]. Neonatal seizures have been reported in approximately 3.5 per 1000 live births of term neonates [17] and several conditions are associated with increased risks for developing seizures such as perinatal asphyxia and ischemic stroke. Indeed, it is estimated that perinatal asphyxia is the leading cause of seizures in the neonatal period, accounting for 40 to 60% of all seizures in term newborns [36, 37]. Seizure burden is associated with more severe brain injury and worse neurodevelopmental outcomes after adjusting for underlying severity of brain injury. Status epilepticus is also associated with post neonatal epilepsy [17]. Since seizures are difficult to diagnose clinically, continuous monitoring by using aEEG, EEG or video aEEG/EEG will help the identification of sub-clinical events allowing immediate recognition and treatment, which can reduce seizure burden and improve neurodevelopmental outcomes [13–16].

Preterm infants are also at high risk for neurodevelopmental problems, and different areas may be affected such as neurosensory, cognitive, and behavioral functioning. The more immature, the greater the risks [8, 38] and extreme prematurity has been associated with moderate or severe neurocognitive deficits during childhood [39]. Moderate and late preterm infants (MLPT) can also have worse cognitive, language, motor development, social-emotional competence at 2 years of age when compared to term infants [40]. Other disorders such as cerebral malformations, infections, inborn errors of metabolism, and severe hemodynamic and ventilatory instability also carry risks of brain injury and seizures. Moreover, CHD affects from 4 to 75 per 1,000 live births and has been recognized as high risk for neurodevelopmental problems [41, 42]. In CHD, infants brain injury may follow episodes of cerebral hypoxia or hypoperfusion that can occur during the pre-, intra- and/or postoperative periods, in addition to biological and environmental factors [43–45]. Indeed, more than 50% of CHD patients have been reported with neurodevelopmental deficiencies [44–47] that may persist into adolescence [47]. All these situations are included as indications for the use of the PBSF Guidelines.

In recent decades, brain monitoring has become increasingly common in NICUs [48] but not yet a consolidated practice in all services, especially in low- and middle- income countries (LMIC), where the lack of structure, resources, and professional training is a major problem. Brazil is a large country with a population of > 200 million people and approximately 3 million births per year, with great financial disparity. This discrepancy directly affects the Brazilian health system, with differences in availability of financial, material, and personal resources, including adequate training of professionals. A prospective cross-sectional national survey conducted by our group evaluated the assessment of HIE and TH practices in Brazil. A total of 1,092 professionals answered and 62% reported using TH in their units. However only 13% received specific training for HIE assessment, 22% did not use any neurologic score, and only 12% had brain monitoring readily available, showing the need for well-established neurocritical care guidelines [31].

There are a variety of clinical applications of video aEEG/EEG monitoring and NIRS in neonates at high-risk for brain injury:

1. Seizures: clinical signs of seizures are easy to evaluate and overdiagnosis occurs frequently, resulting in unnecessary treatment [21, 33, 36]. Neonatal seizures are frequently underdiagnosed since more than 80–90% of seizures are subclinical, resulting in undertreatment. Indeed, the accuracy for seizure detection with clinical evaluation alone is < 9% [37]. New devices associate aEEG with raw EEG readings and even video imaging (video aEEG/EEG), providing higher accuracy for the detection of seizures [12, 36, 48, 49]. Shah et al. evaluated term newborns with clinical seizures monitored with aEEG with simultaneous continuous conventional EEG for seizure detection. In 41 non-status epilepticus seizures, 31 were correctly identified (sensitivity, 76%; specificity, 78%; PPV, 78%; NPV, 78%). However, they found a lower sensitivity (27–56%) when aEEG was assessed without the conventional EEG findings [11].
2. Hypoxic Ischemic Encephalopathy: Three of the six major clinical trials of TH used an abnormal aEEG background activity as entry criteria [50–52]. Indeed, an abnormal aEEG may improve selection of neonates to be treated [53]. However, the presence of a normal early aEEG should not exclude indication of TH in infants with neurological evidence of moderate to severe HIE [54]. After the implementation of TH, the predictive value of aEEG has been evaluated. Chandrasekaran et al. found that persistently abnormal trace on aEEG at ≥ 48 hours of life was associated with adverse long-term outcomes and a systematic review [55] found that maximum predictive reliability was achieved at 72 h of life with a post-test probability of 95.7% (95% CI: 84.4 to 98.5%). Therefore, aEEG is a valuable tool in infants with HIE to be used for the classification of the level of encephalopathy and as a predictor of long-term outcomes. The persistence of elevated values of rScO2 was also associated with abnormal neuroimaging and adverse neurodevelopmental outcomes as low cerebral metabolism indicates low oxygen use, cerebral hyperperfusion and impaired autorregulation [56–58].
3. Preterm infants: the incidence of seizures in preterm infants has been reported between 4 to 48%, and seizures were associated with adverse outcomes including IVH, white matter injury (WMI), cognitive impairment and neonatal death [13, 16, 59, 60]. The presence of SWC and continuous background pattern in the first week of life were associated to good neurodevelopmental outcomes, while absence of SWC and abnormal background activity were associated with neonatal death, severe IVH, WMI and neurological impairment at 2 years of age [61–64]. In extremely premature newborns during the first days of life, the occurrence of episodes of hypoxia/hyperoxia, hemodynamic and ventilatory instability, and systemic hypotension with decreased cerebral flow are common. For that, NIRS monitoring has been used extensively used in this population. A phase II randomized controlled trial conducted by the SafeBoosC Consortium evaluated newborns ≤ 72 h of life to demonstrate the feasibility and effectiveness of implementing continuous NIRS monitoring with a clinical management protocol. Neonates monitored with NIRS had lower burden of cerebral hypoxia or hyperoxia compared to controls and a trend towards lower IVH and mortality rates [65]. A hemodynamically significant patent ductus arteriosus (hsPDA) may be associated with increased pulmonary blood flow and decreased systemic blood flow. Chock et. al found that low renal oxygenation values (rSrO2), with rSrO2 < 66%, were associated with the presence of hemodynamically significant PDA [24]. Other studies have successfully demonstrated association between PDA closure and significant increases on cerebral and renal saturation [66, 67].
4. Other Applications: ventilatory instability (respiratory acidosis/alkalosis), metabolic disorders, inborn errors of metabolism, cerebral malformations, stroke, congenital infections, sepsis, meningitis and need of extracorporeal circulation (ECMO) [48, 68, 69]. In 20% of infants with CHD electrographic seizures were detected, including repetitive seizures and status epilepticus [70]. Early recovery to normal aEEG background activity and presence of SWC was associated with good neurological outcomes after cardiac surgery [46, 71]. Perioperative electrical seizures were found in 30% of infants, of which only ¼ showed any clinical correlation. Failure to recover the background activity to a continuous pattern ≥ 48 h postoperatively was associated with neurodevelopmental delay. The persistence of abnormal aEEG for seven days after surgery was also highly associated with death [46].
   NIRS monitoring may help to evaluate efficacy or need for additional interventions such as ventilator changes, use of diuretics, change in prostaglandin (PGE) dose, need for blood transfusion or early surgical intervention. A rSrO2 < 30% was correlated with increased Intensive Care Units (ICU) stay, and rScO2 and rSrO2 < 45% and < 40% respectively, were correlated with need of ECMO and death [72]. NIRS monitoring was also useful to diagnose early renal insufficiency after surgical intervention [73–75].
5. Anemia and shock: tissue oxygen delivery, somatic and cerebral saturations are directly affected by hemoglobin levels, especially in scenarios of important anemia, suggesting that NIRS monitoring may play a role in the indication of blood transfusion. A significant increase in rScO2 values was reported in children with rScO2 <55% treated with blood transfusion [76]. NIRS has also been demonstrated as a good early indicator of abnormal somatic perfusion and shock in different populations including newborns [77–79]. In the setting of shock, two site NIRS monitoring (cerebral and somatic) may be particularly useful in cases where autoregulation remains intact and somatic tissue oxygenation may be compromised earlier [80].

Implementation of a NeuroNICU requires specialized education, training, use of validated protocols and implementation of specific methodologies. Financial resources are needed for equipment, such as servo-controlled temperature and brain monitoring devices. Lack of specialized assistance and resources may result in inadequate interventions with possible lack of benefit or even harm to the patients. A systematic review found no reduction in mortality with the use of TH in LMIC countries [81] and possible reasons to justify this finding include inefficiency of the low technology cooling devices and lack of optimal care. Brain monitoring using conventional EEG requires skilled technologists, experienced neurologist interpretation which may not be readily available in many centers, especially for continuous bedside monitoring. aEEG and NIRS monitoring can be interpreted at the bedside but also requires training and experienced users [82, 83]. Unfortunately, the majority of neonatal units lack adequate longitudinal training, specialized staff, and access to all these resources, especially in LMIC. This is what motivated our group to consider telemedicine as a solution in this scenario.

Differently from the previous studies, the PBSF Guidelines (S3 Appendix) is unique for the use of an advanced telemedicine system to provide neurocritical care. The guidelines include continuous neuromonitoring with data recording and real time feedback from experts, provision of equipment and resources to the NICUs, and short- and long-term teaching to the professionals working at the units (S2 Appendix). Telemedicine has several applications including education, consultation, practice, and research [28]. This type of communication has shown results in continuing education among physicians and nurses, and has assisted healthcare professionals to improve health disparities, promoting access to specialized care [27, 29]. Live interactive audiovisual materials, such as live streaming video and viewing stored educational material, are important to deliver tele-education; it promotes opportunities for continuous learning and development by dissemination of information and training health-care professionals [26, 28]. Telemedicine may also be helpful in case discussions for decision making which is useful to promote interaction between physicians, nurses, and patients. Udeh et al. exemplified some benefits of telemedicine including [84]: a) promote evidence-based best practices through checklists and prompting; b) enhance monitoring, early identification, and treatment of critical illness; c) improve coordination of care; and d) increase night-time vigilance. The use of telemedicine in ICUs is associated with decreased mortality and length of stay [84, 85]. Surprisingly, telemedicine has not been used on a significant scale worldwide, although this methodology is low-cost, feasible, clinically useful, and sustainable. It could be even more useful in LMIC countries due to the limited structure and resources.

Brazil allows several opportunities for telemedicine implementation, expanding the actions of healthcare professionals and integrating healthcare services [86]. We developed the PBSF Guidelines with a surveillance and intelligence center to support clinical decisions across the country. Telemedicine has been used by our group in the past few years to promote educational training and remote neuromonitoring but its impact on clinical outcomes needs to be formally evaluated. Thus, results of this 3 years prospective, multicenter, and observational study will help to determine the incidence of neonatal problems associated with brain injury and disabilities, the incidence of clinical and subclinical seizures, detection of brain hemodynamic/oxygenation instability and associated interventions. The study will be limited to the 20 NICUs using the PBSF Guidelines and variability in clinical practices, experience of health care providers and staffing of the units, will be analyzed as it may affect some of the outcomes. Enrollment of such large number of patients will require a strong coordination between the central unit and all 20 centers, which has already been established by the CSI. Practical or operational issues will be to assure that all eligible patients are identified in a daily basis and approached for consent. Data collection requires careful training of data extractors and timely and precise data entry. A computerized and safe double-checking system has been developed. Biological signals will be continuously collected and stored at the CSI and multimodal signal analysis has been developed. Although low, there are costs for the implementation and use of the PBSF Guidelines. Thus, an adjusted analysis based on the profile of the units will be performed as costs might represent a limitation on the generalizability of the study findings. A future cost/benefit analysis is under planning.

Supporting information