Predicting Extubation Outcomes-A Model Incorporating Heart Rate Characteristics Index

Heart rate characteristics predict risk of mortality in preterm infants in low and high target oxygen saturation ranges

Background

The Neonatal Oxygenation Prospective Meta-analysis found that in infants <28 weeks gestational age, targeting an oxygen saturation (S pO2 ) range of 85-89% versus 91-95% resulted in lower rates of retinopathy of prematurity but increased mortality. We aimed to evaluate the accuracy of the heart rate characteristics index (HRCi) in assessing the dynamic risk of mortality among infants managed with low and high target S pO2 ranges.

Methods

We linked the SUPPORT and HRCi datasets from one centre in which the randomised controlled trials overlapped. We examined the maximum daily HRCi (MaxHRCi24) to predict mortality among patients randomised to the lower and higher target S pO2 groups by generating predictiveness curves and calculating model performance metrics, including area under the receiver operating characteristics curve (AUROC) at prediction windows from 1-60 days. Cox proportional hazards models tested whether MaxHRCi24 was an independent predictor of mortality. We also conducted a moderation analysis.

Results

There were 84 infants in the merged dataset. MaxHRCi24 predicted mortality in infants randomised to the lower target S pO2 (AUROC of 0.79-0.89 depending upon the prediction window) and higher target S pO2 (AUROC 0.82-0.91). MaxHRCi24 was an important additional predictor of mortality in multivariable modelling. In moderation analysis, in a model that also included demographic predictor variables, the individual terms and the interaction term between MaxHRCi24 and target S pO2 range all predicted mortality.

Conclusions

Associations between HRCi and mortality, at low and high S pO2 target ranges, suggest that future research may find HRCi metrics helpful to individually optimise target oxygen saturation ranges for hospitalised preterm infants.

Artificial intelligence in the NICU to predict extubation success in prematurely born infants

OBJECTIVES

Mechanical ventilation in prematurely born infants, particularly if prolonged, can cause long term complications including bronchopulmonary dysplasia. Timely extubation then is essential, yet predicting its success remains challenging. Artificial intelligence (AI) may provide a potential solution.

CONTENT

A narrative review was undertaken to explore AI's role in predicting extubation success in prematurely born infants. Across the 11 studies analysed, the range of reported area under the receiver operator characteristic curve (AUC) for the selected prediction models was between 0.7 and 0.87. Only two studies implemented an external validation procedure. Comparison to the results of clinical predictors was made in two studies. One group reported a logistic regression model that outperformed clinical predictors on decision tree analysis, while another group reported clinical predictors outperformed their artificial neural network model (AUCs: ANN 0.68 vs. clinical predictors 0.86). Amongst the studies there was an heterogenous selection of variables for inclusion in prediction models, as well as variations in definitions of extubation failure.

SUMMARY

Although there is potential for AI to enhance extubation success, no model's performance has yet surpassed that of clinical predictors.

OUTLOOK

Future studies should incorporate external validation to increase the applicability of the models to clinical settings.

A Nomogram for Predicting Extubation Failure in Preterm Infants with Gestational Age Less than 29 Weeks

INTRODUCTION

How to avoid reintubations in prematurity remains a hard nut. This study aimed to develop and validate a nomogram for predicting extubation failure in preterm infants who received different modes of noninvasive ventilation as post-extubation support.

METHODS

This was a secondary analysis of pre-existing data from a large multicenter RCT combined with a multicenter retrospective investigation in three tertiary referral NICUs in China. The training cohort consisted of extubated infants from the RCT and the validation cohort included neonates admitted to the three NICUs in the last 5 years. The nomogram was developed through univariate and multivariate logistic regression analyses of peri-extubation clinical variables.

RESULTS

A total of 432 and 183 preterm infants (25 weeks ≤ gestational age [GA] <29 weeks) were, respectively, included in the training cohort and the validation cohort. Lower birth weight, lower Apgar 5-min score, lower postmenstrual age at extubation, lower PO2 and higher PCO2 before extubation, and continuous positive airway pressure rather than nasal intermittent positive pressure ventilation or noninvasive high-frequency oscillatory ventilation after extubation were associated with higher risks of extubation failure (p < 0.05), on which the nomogram was established. In both the training cohort and the validation cohort, the nomogram demonstrated good predictive accuracy (area under the receiver operating characteristic curve = 0.744 and 0.826); the Hosmer-Lemeshow test (p = 0.192 and 0.401) and the calibration curve (R2 = 0.195 and 0.307) proved a good fitness and conformity; and the decision curve analysis showed significant net benefit at the best threshold (p = 0.201).

CONCLUSION

This nomogram could serve as a good decision-support tool when predicting extubation failure in preterm infants with GA less than 29 weeks.

Autonomic markers of extubation readiness in premature infants

BACKGROUND

In premature infants, extubation failure is common and difficult to predict. Heart rate variability (HRV) is a marker of autonomic tone. Our aim is to test the hypothesis that autonomic impairment is associated with extubation readiness.

METHODS

Retrospective study of 89 infants <28 weeks. HRV metrics 24 h prior to extubation were compared for those with and without extubation success within 72 h. Receiver-operating curve analysis was conducted to determine the predictive ability of each metric, and a predictive model was created.

RESULTS

Seventy-three percent were successfully extubated. The success group had significantly lower oxygen requirement, higher sympathetic HRV metrics, and a lower parasympathetic HRV metric. α₁ (measure of autocorrelation, related to sympathetic tone) was the best predictor of success-area under the curve (AUC) of .73 (p = 0.001), and incorporated into a predictive model had an AUC of 0.81 (p < 0.0001)-sensitivity of 81% and specificity of 78%.

CONCLUSIONS

Extubation success is associated with HRV. We show an autonomic imbalance with low sympathetic and elevated parasympathetic tone in those who failed. α₁, a marker of sympathetic tone, was noted to be the best predictor of extubation success especially when incorporated into a clinical model.

IMPACT

This article depicts autonomic markers predictive of extubation success. We depict an autonomic imbalance in those who fail extubation with heightened parasympathetic and blunted sympathetic signal. We describe a predictive model for extubation success with a sensitivity of 81% and specificity of 78%.

Oxygen saturation histogram classification system to evaluate response to doxapram treatment in preterm infants

BACKGROUND

An oxygen saturation (SpO₂) histogram classification system has been shown to enable quantification of SpO₂ instability into five types, based on histogram distribution and time spent at SpO₂ ≤ 80%. We aimed to investigate this classification system as a tool to describe response to doxapram treatment in infants with severe apnea of prematurity.

METHODS

This retrospective study included 61 very-low-birth-weight infants who received doxapram. SpO₂ histograms were generated over the 24-h before and after doxapram start. Therapy response was defined as a decrease of ≥1 histogram types after therapy start.

RESULTS

The median (IQR) histogram type decreased from 4 (3-4) before to 3 (2-3) after therapy start (p < 0.001). The median (IQR) FiO₂ remained constant before (27% [24-35%]) and after (26% [22-35%]) therapy. Thirty-six infants (59%) responded to therapy within 24 h. In 34/36 (94%) of the responders, invasive mechanical ventilation (IMV) was not required during the first 72 h of therapy, compared to 15/25 (60%) of non-responders (p = 0.002). Positive and negative predictive values of the 24-h response for no IMV requirement within 72 h were 0.46 and 0.94, respectively.

CONCLUSIONS

Classification of SpO₂ histograms provides an objective bedside measure to assess response to doxapram therapy and can serve as a tool to detect changes in oxygenation status around respiratory interventions.

IMPACT

The SpO₂ histogram classification system provides a tool for quantifying response to doxapram therapy. The classification system allowed estimation of the probability of invasive mechanical ventilation requirement, already within a few hours of treatment. The SpO₂ histogram classification system allows an objective bedside assessment of the oxygenation status of the preterm infant, making it possible to assess the changes in oxygenation status in response to respiratory interventions.

Cardiorespiratory measures shortly after extubation and extubation outcomes in extremely preterm infants

BACKGROUND

Nasal continuous positive airway pressure, nasal intermittent positive pressure ventilation, and non-invasive neurally adjusted ventilatory assist are modes of non-invasive respiratory support. The objective was to investigate if cardiorespiratory measures performed shortly after extubation are associated with extubation outcomes and predictors of extubation success.

METHODS

Randomized crossover trial of infants with birth weight (BW) ≤ 1250 g undergoing their first extubation. Shortly after extubation, electrocardiogram and electrical activity of the diaphragm (Edi) were recorded during 40 min on each mode. Measures of heart rate variability (HRV), diaphragmatic activity (Edi area, breath area and amplitude), and respiratory variability (RV) were computed on each mode and compared between infants with extubation success or failure (reintubation ≤ 7 days).

RESULTS

Twenty-three extremely preterm infants with median [IQR] gestational age 25.9 weeks [25.2-26.4] and BW 760 g [595-900] were included: 14 success and 9 failures. There were significant differences for HRV (very low-frequency power and sample entropy) and RV parameters (breath areas, amplitudes and expiratory times) between groups, with moderate strength (0.75-0.80 areas under ROC curves) in predicting success. Diaphragmatic activity measures were similar between groups.

CONCLUSIONS

In extremely preterm infants receiving non-invasive respiratory support shortly after extubation, several cardiorespiratory variability parameters were associated with successful extubation with moderate predictive accuracy.

IMPACT

Measures of cardiorespiratory variability, performed in extremely preterm infants while receiving NCPAP, NIPPV, and NIV-NAVA shortly after extubation, were significantly different between patients that succeeded or failed extubation. Cardiorespiratory variability measures had a moderate predictive accuracy for extubation success and can be potentially used as biomarkers, in recently extubated infants. Future investigations in this population may also consider including cardiorespiratory variability measures when assessing types of post-extubation respiratory support and promote individualized care.

A Simple Weaning Model Based on Interpretable Machine Learning Algorithm for Patients With Sepsis: A Research of MIMIC-IV and eICU Databases

Background

Invasive mechanical ventilation plays an important role in the prognosis of patients with sepsis. However, there are, currently, no tools specifically designed to assess weaning from invasive mechanical ventilation in patients with sepsis. The aim of our study was to develop a practical model to predict weaning in patients with sepsis.

Methods

We extracted patient information from the Medical Information Mart for Intensive Care Database-IV (MIMIC-IV) and the eICU Collaborative Research Database (eICU-CRD). Kaplan–Meier curves were plotted to compare the 28-day mortality between patients who successfully weaned and those who failed to wean. Subsequently, MIMIC-IV was divided into a training set and an internal verification set, and the eICU-CRD was designated as the external verification set. We selected the best model to simplify the internal and external validation sets based on the performance of the model.

Results

A total of 5020 and 7081 sepsis patients with invasive mechanical ventilation in MIMIC-IV and eICU-CRD were included, respectively. After matching, weaning was independently associated with 28-day mortality and length of ICU stay (p &lt; 0.001 and p = 0.002, respectively). After comparison, 35 clinical variables were extracted to build weaning models. XGBoost performed the best discrimination among the models in the internal and external validation sets (AUROC: 0.80 and 0.86, respectively). Finally, a simplified model was developed based on XGBoost, which included only four variables. The simplified model also had good predictive performance (AUROC:0.75 and 0.78 in internal and external validation sets, respectively) and was developed into a web-based tool for further review.

Conclusions

Weaning success is independently related to short-term mortality in patients with sepsis. The simplified model based on the XGBoost algorithm provides good predictive performance and great clinical applicablity for weaning, and a web-based tool was developed for better clinical application.

Neonatal heart rate variability: a contemporary scoping review of analysis methods and clinical applications

BACKGROUND

Neonatal heart rate variability (HRV) is widely used as a research tool. However, HRV calculation methods are highly variable making it difficult for comparisons between studies.

OBJECTIVES

To describe the different types of investigations where neonatal HRV was used, study characteristics, and types of analyses performed.

ELIGIBILITY CRITERIA

Human neonates ≤1 month of corrected age.

SOURCES OF EVIDENCE

A protocol and search strategy of the literature was developed in collaboration with the McGill University Health Center's librarians and articles were obtained from searches in the Biosis, Cochrane, Embase, Medline and Web of Science databases published between 1 January 2000 and 1 July 2020.

CHARTING METHODS

A single reviewer screened for eligibility and data were extracted from the included articles. Information collected included the study characteristics and population, type of HRV analysis used (time domain, frequency domain, non-linear, heart rate characteristics (HRC) parameters) and clinical applications (physiological and pathological conditions, responses to various stimuli and outcome prediction).

RESULTS

Of the 286 articles included, 171 (60%) were small single centre studies (sample size <50) performed on term infants (n=136). There were 138 different types of investigations reported: physiological investigations (n=162), responses to various stimuli (n=136), pathological conditions (n=109) and outcome predictor (n=30). Frequency domain analyses were used in 210 articles (73%), followed by time domain (n=139), non-linear methods (n=74) or HRC analyses (n=25). Additionally, over 60 different measures of HRV were reported; in the frequency domain analyses alone there were 29 different ranges used for the low frequency band and 46 for the high frequency band.

CONCLUSIONS

Neonatal HRV has been used in diverse types of investigations with significant lack of consistency in analysis methods applied. Specific guidelines for HRV analyses in neonates are needed to allow for comparisons between studies.

Clinical study of Dysfunctional Ventilatory Weaning Response in critically ill patients

Objective:

to clinically validate the nursing diagnosis of Dysfunctional Ventilatory Weaning Response in adult patients admitted to Intensive Care Units.

Method:

a concurrent cohort performed with 93 patients admitted to Intensive Care Units. The incidence and incidence density of the diagnosis were estimated, its related factors were identified based on bivariate analysis and clinical indicators for determining its occurrence, according to the global and temporal presentation.

Results:

the overall incidence of the diagnosis was 44.09% and the incidence density was 14.49 occurrences for every 100 extubations/day. The factors related to the diagnosis were the following: age, clinical severity, fluid balance, oliguria, hemodialysis, edema in upper/lower limbs, anasarca, number of antibiotics, hypothermia, hyperthermia, amount of secretion, muscle retraction, anxiety score, heart rate, use of vasopressors and non-invasive ventilation after extubation. The clinical indicators most frequently identified for determining the diagnosis were the following: tachypnea, drop of saturation and tachycardia. Temporal progression in the severity of these manifestations was found.

Conclusion:

the Dysfunctional Ventilatory Weaning Response is a common finding in critically ill patients. Some components of the diagnosis of the NANDA-International (2018) version could be clinically validated. It is noteworthy that there are variables not yet described in the taxonomy, demonstrating the need to review this nursing diagnosis.

Predicting extubation outcomes using the Heart Rate Characteristics index in preterm infants: a cohort study

A strategy of early extubation to noninvasive respiratory support in preterm infants could be boosted by the availability of a decision support tool for clinicians. Using the Heart Rate Characteristics index (HRCi) with clinical parameters, we derived and validated predictive models for extubation readiness and success.Peri-extubation demographic, clinical and HRCi data for up to 96 h were collected from mechanically ventilated infants in the control arm of a randomised trial involving eight neonatal centres, where clinicians were blinded to the HRCi scores. The data were used to produce a multivariable regression model for the probability of subsequent re-intubation. Additionally, a survival model was produced to estimate the probability of re-intubation in the period after extubation.Of the 577 eligible infants, data from 397 infants (69%) were used to derive the pre-extubation model and 180 infants (31%) for validation. The model was also fitted and validated using all combinations of training (five centres) and test (three centres) centres. The estimated probability for the validation episodes showed discrimination with high statistical significance, with an area under the curve of 0.72 (95% CI 0.71-0.74; p<0.001). Data from all infants were used to derive models of the predictive instantaneous hazard of re-intubation adjusted for clinical parameters.Predictive models of extubation readiness and success in real-time can be derived using physiological and clinical variables. The models from our analyses can be accessed using an online tool available at www.heroscore.com/extubation, and have the potential to inform and supplement the confidence of the clinician considering extubation in preterm infants.

Continuous vital sign analysis for predicting and preventing neonatal diseases in the twenty-first century: big data to the forefront

In the neonatal intensive care unit (NICU), heart rate, respiratory rate, and oxygen saturation are vital signs (VS) that are continuously monitored in infants, while blood pressure is often monitored continuously immediately after birth, or during critical illness. Although changes in VS can reflect infant physiology or circadian rhythms, persistent deviations in absolute values or complex changes in variability can indicate acute or chronic pathology. Recent studies demonstrate that analysis of continuous VS trends can predict sepsis, necrotizing enterocolitis, brain injury, bronchopulmonary dysplasia, cardiorespiratory decompensation, and mortality. Subtle changes in continuous VS patterns may not be discerned even by experienced clinicians reviewing spot VS data or VS trends captured in the monitor. In contrast, objective analysis of continuous VS data can improve neonatal outcomes by allowing heightened vigilance or preemptive interventions. In this review, we provide an overview of the studies that have used continuous analysis of single or multiple VS, their interactions, and combined VS and clinical analytic tools, to predict or detect neonatal pathophysiology. We make the case that big-data analytics are promising, and with continued improvements, can become a powerful tool to mitigate neonatal diseases in the twenty-first century.

Diffusing an Innovation: Clinician Perceptions of Continuous Predictive Analytics Monitoring in Intensive Care

BACKGROUND

The purpose of this article is to describe neonatal intensive care unit clinician perceptions of a continuous predictive analytics technology and how those perceptions influenced clinician adoption. Adopting and integrating new technology into care is notoriously slow and difficult; realizing expected gains remain a challenge.

METHODS

Semistructured interviews from a cross-section of neonatal physicians (n = 14) and nurses (n = 8) from a single U.S. medical center were collected 18 months following the conclusion of the predictive monitoring technology randomized control trial. Following qualitative descriptive analysis, innovation attributes from Diffusion of Innovation Theory-guided thematic development.

RESULTS

Results suggest that the combination of physical location as well as lack of integration into work flow or methods of using data in care decisionmaking may have delayed clinicians from routinely paying attention to the data. Once data were routinely collected, documented, and reported during patient rounds and patient handoffs, clinicians came to view data as another vital sign. Through clinicians' observation of senior physicians and nurses, and ongoing dialogue about data trends and patient status, clinicians learned how to integrate these data in care decision making (e.g., differential diagnosis) and came to value the technology as beneficial to care delivery.

DISCUSSION

The use of newly created predictive technologies that provide early warning of illness may require implementation strategies that acknowledge the risk-benefit of treatment clinicians must balance and take advantage of existing clinician training methods.