Background: Infants born extremely preterm (<28 weeks’ gestation) are at high risk of neurodevelopmental impairment (NDI) with 50% of survivors showing moderate or severe NDI when at 2 years of age. ...We sought to develop novel models by which to predict neurodevelopmental outcomes, hypothesizing that combining baseline characteristics at birth with medical care and environmental exposures would produce the most accurate model. Methods: Using a prospective database of 692 infants from the Preterm Epo Neuroprotection (PENUT) Trial, which was carried out between December 2013 and September 2016, we developed three predictive algorithms of increasing complexity using a Bayesian Additive Regression Trees (BART) machine learning approach to predict both NDI and continuous Bayley Scales of Infant and Toddler Development 3rd ed subscales at 2 year follow-up using: 1) the 5 variables used in the National Institute of Child Health and Human Development (NICHD) Extremely Preterm Birth Outcomes Tool, 2) 21 variables associated with outcomes in extremely preterm (EP) infants, and 3) a hypothesis-free approach using 133 potential variables available for infants in the PENUT database. Findings: The NICHD 5-variable model predicted 3–4% of the variance in the Bayley subscale scores, and predicted NDI with an area under the receiver operator curve (AUROC, 95% CI) of 0.62 (0.56–0.69). Accuracy increased to 12–20% of variance explained and an AUROC of 0.77 (0.72–0.83) when using the 21 pre-selected clinical variables. Hypothesis-free variable selection using BART resulted in models that explained 20–31% of Bayley subscale scores and AUROC of 0.87 (0.83–0.91) for severe NDI, with good calibration across the range of outcome predictions. However, even with the most accurate models, the average prediction error for the Bayley subscale predictions was around 14–15 points, leading to wide prediction intervals. Higher total transfusion volume was the most important predictor of severe NDI and lower Bayley scores across all subscales. Interpretation: While the machine learning BART approach meaningfully improved predictive accuracy above a widely used prediction tool (NICHD) as well as a model utilizing NDI-associated clinical characteristics, the average error remained approximately 1 standard deviation on either side of the true value. Although dichotomous NDI prediction using BART was more accurate than has been previously reported, and certain clinical variables such as transfusion exposure were meaningfully predictive of outcomes, our results emphasize the fact that the field is still not able to accurately predict the results of complex long-term assessments such as Bayley subscales in infants born EP even when using rich datasets and advanced analytic methods. This highlights the ongoing need for long-term follow-up of all EP infants. Funding: Supported by the National Institute of Neurological Disorders and Stroke U01NS077953 and U01NS077955.
Both HFOV and HFJV are important adjuncts to the ventilatory care of sick infants and children. Today, it is important that neonatologists, pediatric intensivists, and respiratory care practitioners ...understand these ventilators and the options they provide. It is no longer necessary to continue the use of damaging pressures and volumes with CV simply because no other option is available. The clinician who understands not only the pathology and physiology of the underlying lung condition but also understands the available choices in ventilators, how each ventilator functions, and what potential advantage it may offer his patients is able to provide the best possible care to these critically ill patients.
Pulmonary disease is the most important cause of morbidity in preterm neonates, whose lungs are often physiologically and morphologically immature. Surfactant deficiency in immature lungs triggers a ...cascade of alveolar instability and collapse, capillary leak edema, and hyaline membrane formation. The term respiratory distress syndrome (RDS) has come to represent the clinical expression of surfactant deficiency and its nonspecific histologic counterpart, hyaline membrane disease. Historically, chest radiographs of infants with RDS predictably demonstrated decreased pulmonary expansion, symmetric generalized reticulogranular lung opacities, and air bronchograms. Refinements in perinatal medicine, including antenatal glucocorticoid administration, surfactant replacement therapy, and increasingly sophisticated ventilatory strategies have decreased the prevalence of RDS and air leak, altered familiar radiographic features, and lowered the threshold of potential viability to a gestational age of approximately 23 weeks. Alveolar paucity and pulmonary interstitial thickness in these profoundly premature neonates impair normal gas exchange and may necessitate prolonged mechanical ventilation, increasing the risk of lung injury. Bronchopulmonary dysplasia (BPD), alternatively termed chronic lung disease of infancy, is a disorder of lung injury and repair originally ascribed to positive-pressure mechanical ventilation and oxygen toxicity. Before the advent of surfactant replacement therapy, chest radiographs of infants with classic BPD demonstrated coarse reticular lung opacities, cystic lucencies, and markedly disordered lung aeration that reflected alternating regions of alveolar septal fibrosis and hyperinflated normal lung parenchyma. In the current era of surfactant replacement, BPD is increasingly a disorder of very low-birth-weight neonates with arrested alveolar and pulmonary vascular development, minimal alveolar septal fibrosis and inflammation, and more subtle radiographic abnormalities.
1 Department of Pediatrics, Robert Wood Johnson Medical
School at Camden, The Children's Regional Hospital at Cooper
Hospital/University Medical Center, Camden, New Jersey 08103; and
2 Mercy ...Children's Hospital and Department of Pediatrics,
Medical College of Ohio, Toledo, Ohio 43608
Positive airway
pressure (Paw) during high-frequency oscillatory ventilation (HFOV)
increases lung volume and can lead to lung overdistention with
potentially serious adverse effects. To date, no method is available to
monitor changes in lung volume ( V L ) in HFOV-treated
infants to avoid overdistention. In five newborn piglets (6-15
days old, 2.2-4.2 kg), we investigated the use of direct
current-coupled respiratory inductive plethysmography (RIP) for
this purpose by evaluating it against whole body plethysmography. Animals were instrumented, fitted with RIP bands, paralyzed, sedated, and placed in the plethysmograph. RIP and plethysmography were simultaneously calibrated, and HFOV was instituted at varying Paw
settings before (6-14 cmH 2 O) and after (10-24
cmH 2 O) repeated warm saline lung lavage to induce
experimental surfactant deficiency. Estimates of V L from
both methods were in good agreement, both transiently and in the steady
state. Maximal changes in lung volume ( V L max ) from all piglets were highly
correlated with V L measured by RIP (in ml) = 1.01 × changes measured by whole body plethysmography 0.35; r 2 = 0.95. Accuracy of RIP was
unchanged after lavage. Effective respiratory system compliance (Ceff)
decreased after lavage, yet it exhibited similar sigmoidal dependence
on V L max pre- and postlavage. A decrease in
Ceff (relative to the previous Paw setting) as
V L max was methodically increased from low to
high Paw provided a quantitative method for detecting lung
overdistention. We conclude that RIP offers a noninvasive and
clinically applicable method for accurately estimating lung recruitment
during HFOV. Consequently, RIP allows the detection of lung
overdistention and selection of optimal HFOV from derived Ceff data.
respiratory inductance plethysmography; infants; mechanical
ventilation; lung mechanics; respiratory distress syndrome
1 Department of Pediatrics, The Children's
Regional Hospital at Cooper Hospital and Robert Wood Johnson Medical
School, Camden 08103; 2 Department of Pediatrics,
Robert Wood Johnson Medical School, ...New Brunswick, New Jersey 08901;
and 3 Department of Pediatrics, Mercy Children's
Hospital at St. Vincent Mercy Medical Center, Medical College of Ohio,
Toledo, Ohio 43608
Reported values of lung resistance
(R L ) and elastance (E L ) in spontaneously
breathing preterm neonates vary widely. We hypothesized that this
variability in lung properties can be largely explained by both inter-
and intrasubject variability in breathing pattern and demographics.
Thirty-three neonates receiving nasal continuous positive airway
pressure weight 606-1,792 g, gestational age (GA) of
25-33 wk, 2-49 days old were studied. Transpulmonary pressure was measured by esophageal manometry and airway flow by face
mask pneumotachography. Breath-to-breath changes in R L and
E L in each infant were estimated by Fourier analysis of
impedance (Z) and by multiple linear regression (MLR).
R L MLR (R L MLR = 0.85 × R L Z 0.43; r 2 = 0.95) and E L MLR
(E L MLR = 0.97 × E L Z + 8.4; r 2 = 0.98) were
highly correlated to R L Z and
E L Z , respectively. Both R L
(mean ± SD; R L Z = 70 ± 38, R L MLR = 59 ± 36 cmH 2 O · s · l 1 )
and E L (E L Z = 434 ± 212, E L MLR = 436 ± 210 cmH 2 O/l)
exhibited wide intra- and intersubject variability.
Regardless of computation method, R L was found to decrease
as a function of weight, age, respiratory rate (RR), and tidal volume
(V T ) whereas it increased as a function of
RR · V T and inspiratory-to-expiratory
time ratio (T I /T E ). E L decreased
with increasing weight, age, V T and female gender and
increased as RR and T I /T E increased. We
conclude that accounting for the effects of breathing pattern
variability and demographic parameters on estimates of R L
and E L is essential if they are to be of clinical value.
Multivariate statistical models of R L and E L
may facilitate the interpretation of lung mechanics measurements in
spontaneously breathing infants.
impedance; multiple linear regression; frequency dependence; amplitude dependence
High-Frequency Ventilation Thome, Ulrich H; Pohlandt, Frank
The New England journal of medicine,
03/2003, Letnik:
348, Številka:
12
Journal Article
Recenzirano
To the Editor:
In the studies of high-frequency ventilation reported in the August 29 issue by Johnson et al.
1
and Courtney et al.,
2
there were important differences in the application of ...conventional ventilation in the control groups, which may explain why these trials had conflicting results. The trial by Johnson et al. required conventional ventilation to be started at a rate of 60 breaths per minute, in accordance with previous studies of optimal mechanical ventilation.
3
,
4
Courtney et al., however, set 60 breaths per minute as the maximal rate, which made it likely that lower rates and therefore higher tidal . . .
To evaluate the feasibility of conducting a prospective, randomized trial comparing early high-frequency oscillatory ventilation (HFOV) to synchronized intermittent mandatory ventilation (SIMV) in ...very low birth weight (VLBW) premature infants. This pilot study evaluated two ventilator management protocols to determine how well they could be implemented in a multicenter clinical trial. Although this pilot study was not powered to detect differences in outcome, we also collected outcome data.
Prospective, multicenter, randomized pilot study.
Seven tertiary-level intensive care nurseries with previous experience with both HFOV and flow-triggered SIMV.
Fifty infants weighing 501 to 1200 g, less than 4 hours of age, who had received one dose of surfactant and required ventilation with mean airway pressure > or =6 cm H2O and F(I)O2 > or =0.25, and had an anticipated duration of ventilation greater than 24 hours.
Patients were stratified by birth weight and prenatal steroid status, then randomized to either HFOV or SIMV with tidal volume monitoring. Ventilator management for patients in both study arms was strictly governed by protocols that included optimizing lung inflation and blood gases, weaning strategies, and extubation criteria.
Data were collected using the tools planned for the larger collaborative study. Protocol compliance was closely monitored, with successive changes in the protocol made as necessary to improve clarity and increase compliance. The incidence of major neonatal adverse outcomes was recorded.
Data are presented for 24 HFOV and 24 SIMV infants (two infants, twins, were withdrawn from the study at parent's request). Nineteen of the 24 HFOV infants and 20 of the 24 SIMV infants survived to 36 weeks corrected age. Age at final extubation for survivors was 16+/-16 (mean+/-SD) days for HFOV infants and 24+/-24 days for SIMV infants. At 36 weeks corrected age, 14 of the 19 HFOV survivors were extubated and in room air, whereas 5 required supplemental oxygen. In comparison, 6 of the 20 SIMV survivors were extubated and in room air, whereas 14 required supplemental oxygen. Grade III/IV IVH and/or periventricular leukomalacia occurred in 2 HFOV and 2 SIMV patients. Overall compliance with the ventilator protocols was 82% for the SIMV protocol, and 88% for the HFOV protocol.
The preliminary outcome data supports conducting the large randomized trial, which began in July of 1998. The protocols for the ventilator management of VLBW infants, both with HFOV and with SIMV were easily implemented and consistently followed, and are presented here.
To assess the accuracy of a pneumotachometer (PN) for tidal volume (VT) measurements during high-frequency oscillation (HFO), we determined simultaneously VT using a PN and a full body plethysmograph ...(PL) in 12 rabbits. HFO was delivered with an oscillator at a frequency of 10 Hz, mean airway pressure of 8 cm H2O, and inspiratory time of 50%. Pressure amplitude (delta P) was varied as follows: 40, 60, 80, 20, 100, 40 cm H2O. Finally, in ten rabbits a spacer equal in deadspace (VD) to that of the PN (15 ml) was left in-line for 5 min. Blood gases were obtained before and after the spacer was added. We found that VT-PN correlates well with VT-PL (r = .92), although the difference between VT-PN and VT-PL is greater at large VT. Significant respiratory acidosis developed with the spacer in-line. PN may be used to trend VT during HFO but PN must not be left in-line, as increased VD seriously affects ventilation.
PDF
