Pulse Oximetry Screening for Congenital Heart Defects

No longer recruiting at 2 trial locations

Overseen By Elva Horath, IMG

Age: < 18

Sex: Any

Trial Phase: Academic

Sponsor: University of California, Davis

Must not be taking: Vasoactive medications

Disqualifiers: Echocardiogram, Cardiac surgery, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: Unknown

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial explores a new method to screen newborns for critical congenital heart disease (CCHD) using a machine learning (ML) algorithm. The approach combines data from pulse oximeters, which measure blood oxygen levels, to determine if the algorithm can accurately predict CCHD. This non-invasive method could enable earlier and easier detection of heart issues in newborns. The trial is suitable for babies less than 22 days old, particularly those suspected of having heart problems or undergoing routine oxygen level screenings. As an unphased trial, it offers a chance to contribute to groundbreaking research that could enhance early detection of heart issues in newborns.

Do I need to stop my current medications to join the trial?

The trial protocol does not specify if you need to stop taking your current medications. However, if you are on vasoactive medications other than prostaglandin therapy, you may not be eligible to participate.

What prior data suggests that this ML algorithm and SpO2/PIx measurement are safe for CCHD screening?

Research has shown that using pulse oximetry to screen newborns for serious heart defects is safe and effective. This method checks the oxygen levels in a baby's blood without causing harm. It has significantly reduced early infant deaths from these heart issues. Studies indicate that this simple, non-invasive test can detect heart problems that might otherwise go unnoticed.

Recently, trials combined pulse oximetry with a machine learning (ML) algorithm to enhance screening. This new approach helps identify heart defects soon after birth. Current evidence suggests that using pulse oximetry with an ML algorithm is safe for newborns and poses no significant risks. ^{1 2 3 4 5}

Why are researchers excited about this trial?

Researchers are excited about this trial because it leverages a machine learning (ML) algorithm combined with pulse oximetry to screen for congenital heart defects (CCHD) in newborns. Unlike traditional methods that rely primarily on physical examinations and echocardiograms, this approach offers a non-invasive and continuous measurement of oxygen levels (SpO2) and perfusion index (PIx). The ML algorithm provides real-time predictions, potentially increasing the accuracy and speed of detecting heart defects right after birth. This innovative use of technology could allow for earlier intervention and treatment, improving outcomes for affected infants.

What evidence suggests that this ML algorithm is effective for CCHD screening?

Research has shown that using pulse oximetry for screening can significantly reduce early infant deaths from critical congenital heart defects (CCHD) by 33%, saving about 120 lives each year. This screening also decreases the likelihood of discharging babies with undiagnosed heart problems by nearly six times. Pulse oximetry is a simple, painless test that can detect 50-70% of previously undetected heart issues in newborns.

In this trial, participants will undergo SpO2 and PIx Measurement, which involves non-invasive measurements of oxygenation and perfusion using pulse oximeters. Additionally, a machine learning (ML) algorithm will analyze the pulse oximetry data. This ML approach has shown promise in better detecting CCHD and related conditions within 48 hours after birth. ML models improve the accuracy of identifying heart defects compared to older methods. These findings suggest that combining traditional pulse oximetry with advanced ML could effectively catch heart issues early in newborns. ^{4 6 7 8 9}

Are You a Good Fit for This Trial?

This trial is for asymptomatic newborns under 22 days old being screened for critical congenital heart disease (CCHD), including those suspected of having CCHD or with a prenatal suspicion of heart defects. Newborns who've had certain cardiac interventions, are on specific medications, or have had an echocardiogram prior to enrollment are excluded.

Inclusion Criteria

I am less than 22 days old.

My newborn is healthy but needs a heart defect screening.

My newborn is suspected or confirmed to have a serious heart condition.

Exclusion Criteria

Newborns with certain types of heart defects or who have had heart surgery or current use of specific heart medications may not participate.

An echocardiogram needs to be done before enrolling the newborn.

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Inpatient Screening

Non-invasive measurements of oxygenation (SpO2) and perfusion (PIx) are taken using pulse oximeters and a machine learning CCHD screening algorithm assigns a prediction every minute.

0-48 hours

Continuous monitoring

Outpatient Follow-up

Repeated pulse oximetry measurements are conducted after 48 hours of age, potentially in an outpatient setting, to enhance detection of CCHD.

Up to 2 months

Follow-up

Participants are monitored for safety and effectiveness after screening, with health status confirmed to a minimum of 2 months of age.

4 years

What Are the Treatments Tested in This Trial?

Interventions

• ML Algorithm
• SpO2/PIx Measurement

Trial Overview The study tests a machine learning algorithm that uses pulse oximetry data to screen for CCHD in infants. It aims to validate the effectiveness of combining oxygen saturation and perfusion index measurements from standard screenings with this new technology.

How Is the Trial Designed?

Group I: SpO2 and PIx Measurement Experimental Treatment 1 Intervention

Non-invasive measurements of oxygenation (SpO2) and perfusion (PIx) will be measured with pulse oximeters and a ML CCHD screening algorithm will be assigning a prediction every minute.

ML Algorithm is already approved in United States, European Union, Canada, United Kingdom for the following indications:

Approved in United States as AI/ML Algorithm for CCHD Screening for:

• Critical Congenital Heart Disease Screening

Approved in European Union as AI/ML Algorithm for CCHD Screening for:

• Critical Congenital Heart Disease Screening

Approved in Canada as AI/ML Algorithm for CCHD Screening for:

• Critical Congenital Heart Disease Screening

Approved in United Kingdom as AI/ML Algorithm for CCHD Screening for:

• Critical Congenital Heart Disease Screening

Who Is Running the Clinical Trial?

University of California, Davis

Lead Sponsor

Trials

958

Recruited

4,816,000+

National Institutes of Health (NIH)

Collaborator

Trials

2,896

Recruited

8,053,000+

Published Research Related to This Trial

In a study of 5487 newborns screened for congenital heart disease (CHD), only 19% sensitivity was found for detecting all CHDs and 20% for major CHDs using clinical evaluation and pulse oximetry, indicating a need for improved detection methods.

Despite the low sensitivity, the combination of clinical signs and pulse oximetry showed high specificity (88%), suggesting that while many cases may be missed, those identified as abnormal should receive prompt echocardiography for further evaluation.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Clinical screening for Congenital heart disease at birth: a prospective study in a community hospital in Kerala. Vaidyanathan, B., Sathish, G., Mohanan, ST., et al.[2022]

In a study of 6,329 newborns, pulse-oximetry was found to be a safe and effective screening tool for detecting critical congenital heart disease (CCHD), with a sensitivity of 87.5% and specificity of 99.8%.

The use of pulse-oximetry identified 4 out of 7 cases of CCHD that would have been missed by clinical examination alone, highlighting its importance in routine newborn checks.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Effectiveness of pulse-oximetry in addition to routine neonatal examination in detection of congenital heart disease in asymptomatic newborns. Oakley, JL., Soni, NB., Wilson, D., et al.[2015]

Current screening for Critical Congenital Heart Disease (CCHD) using only oxygen saturation (SpO2) misses about 900 newborns in the US each year, highlighting the need for improved detection methods.

By integrating machine learning algorithms that analyze additional pulse oximetry features, the proposed screening system increased sensitivity by approximately 10 percentage points without significantly affecting the false positive rate, demonstrating a promising advancement in CCHD detection.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Enhanced Critical Congenital Cardiac Disease Screening by Combining Interpretable Machine Learning Algorithms. Lai, Z., Vadlaputi, P., Tancredi, DJ., et al.[2022]

Pulse oximetry screening for critical congenital cardiovascular malformations (CCVM) in asymptomatic newborns showed a sensitivity of 60% and a specificity of 99.95%, indicating it is a reliable method for detecting serious heart conditions in newborns.

The study screened 11,281 newborns and detected 3 cases of CCVM, demonstrating that while the test is effective, there is a risk of false negatives for certain heart defects that may not be detectable through oximetry alone.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Effectiveness of pulse oximetry screening for congenital heart disease in asymptomatic newborns. Koppel, RI., Druschel, CM., Carter, T., et al.[2022]

In a study of 200 newborns with cyanotic congenital heart disease (CCHD), significant differences in pulse oxygen saturation (SpO2) levels were observed, with the lowest levels found in those with transposition of the great arteries (TGA).

Respiratory interventions were most frequently required for TGA infants, with 66% receiving continuous positive airway pressure, 33% requiring positive pressure ventilation, and 53% needing intubation, highlighting the critical need for tailored respiratory support in these high-risk newborns.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Delivery room oxygen physiology and respiratory interventions for newborns with cyanotic congenital heart disease. Thomas, AR., Ma, AL., Weinberg, DD., et al.[2023]

Machine learning algorithms demonstrated high accuracy in categorizing adult congenital heart disease (ACHD) patients by diagnosis, disease complexity, and functional class, achieving accuracies of 91.1%, 97.0%, and 90.6% respectively, based on a dataset of over 44,000 medical records from 10,019 patients.

The algorithms also effectively predicted the need for multidisciplinary team discussions and were linked to patient survival outcomes, indicating their potential to guide therapy and improve prognosis in ACHD patients over a median follow-up of 8 years.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Machine learning algorithms estimating prognosis and guiding therapy in adult congenital heart disease: data from a single tertiary centre including 10 019 patients. Diller, GP., Kempny, A., Babu-Narayan, SV., et al.[2022]

Pulse oximetry screening before discharge significantly improved the detection rate of duct dependent circulation in newborns, achieving a sensitivity of 82.8% and detecting 100% of cases, compared to a much lower detection rate in regions without this screening.

The introduction of pulse oximetry led to a substantial reduction in the risk of leaving the hospital undiagnosed, with only 8% of babies in West Götaland having undetected duct dependent circulation compared to 28% in other regions, ultimately preventing deaths from undiagnosed conditions.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Impact of pulse oximetry screening on the detection of duct dependent congenital heart disease: a Swedish prospective screening study in 39,821 newborns. de-Wahl Granelli, A., Wennergren, M., Sandberg, K., et al.[2022]

In a study of 963 newborns in Quito, pulse oximetry screening did not identify any critical congenital heart diseases (CHD), but it successfully detected other heart conditions like atrial septal defect (ASD) and patent ductus arteriosus (PDA).

The findings suggest that the current saturation cutoff points for identifying critical cardiac anomalies may need to be adjusted for high-altitude environments, as the mean preductal and postductal saturations were lower than typical values.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Is pulse oximetry helpful for the early detection of critical congenital heart disease at high altitude? González-Andrade, F., Echeverría, D., López, V., et al.[2019]

In a study involving 102 healthcare providers, the accuracy of interpreting pulse oximetry results for screening critical congenital heart disease was significantly higher when using a computer-based tool (98.3%) compared to a manual algorithm (81.6%).

The computer-based tool particularly improved accuracy for 'fail' scenarios (96.7% vs. 65.4%) and 'retest' scenarios (98.7% vs. 80.7%), highlighting its potential to reduce human error in critical health screenings.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Quality improvement in screening for critical congenital heart disease. Oster, ME., Kuo, KW., Mahle, WT.[2013]

Citations

1. ahajournals.org ahajournals.org/doi/10.1161/JAHA.123.033786

Machine Learning–Based Critical Congenital Heart ... ML pulse oximetry that combines oxygenation, perfusion data, and pulse delay at 2 time points may improve detection of CCHD and CoA within 48 hours after birth.

2. pubmed.ncbi.nlm.nih.gov pubmed.ncbi.nlm.nih.gov/38879455/

Machine Learning-Based Critical Congenital Heart Disease ... ML pulse oximetry that combines oxygenation, perfusion data, and pulse delay at 2 time points may improve detection of CCHD and CoA within 48 hours after birth.

3. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC10841295/

A machine learning model for predicting congenital heart ... Conclusions: Applying ML algorithms improved the accuracy of identifying TP CHD cases in comparison to ICD codes alone. Use of this technique to identify CHD ...

4. sciencedirect.com sciencedirect.com/science/article/pii/S1386505624004040

Development of machine learning-based models to predict ... This study underscores the feasibility and efficacy of employing a machine learning-based approach for CHD prediction.

5. arxiv.org arxiv.org/html/2501.04493v1

The Role of Machine Learning in Congenital Heart ... This paper presents a systematic review of congential heart disease recognition using machine learning, conducting a meta-analysis of 432 ...

6. publications.aap.org publications.aap.org/pediatrics/article/155/1/e2024069667/200337/Newborn-Screening-for-Critical-Congenital-Heart

Newborn Screening for Critical Congenital Heart Disease Newborn CCHD screening using pulse oximetry plays an important role in the timely identification of children with CCHD, in conjunction with ...

7. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC4946827/

Pulse oximetry screening: a review of diagnosing critical ... Pulse oximetry is a tool to measure oxygen saturation, and based on the presence of hypoxemia, many cardiac lesions are detected.

8. cdc.gov cdc.gov/heart-defects/hcp/screening/index.html

Clinical Screening and Diagnosis for Critical Congenital ... Mandated CCHD screening using pulse oximetry reduces early infant deaths from CCHD by 33%, or 120 early infant deaths from CCHD averted per year ...

9. wisconsinshine.org wisconsinshine.org/handouts/Pulse-Oximetry-Toolkit.pdf

Pulse Oximetry Screening for Congenital Heart Disease Pulse oximetry screening is used to detect low blood oxygen levels that cannot be recognized by the naked eye. In some newborns, this abnormal blood oxygen.