Should Dual Pulse Oximetry Screening Be Used in Neonates? The Insights from the CHOOSE Study Authors: Neha Singh, Jaya Upadhyay, Monica Lazarus, Shishir Soni Published: 2026 ABSTRACT ABSTRACTBackground:Critical congenital heart disease (CCHD) is a leading cause of neonatal morbidity and mortality. Early detection is crucial to prevent life-threatening complications. Pulse oximetry screening (POS) is a noninvasive method to identify hypoxemia indicative of CCHD, but the optimal timing for screening remains debated.Aim:To evaluate the diagnostic accuracy of dual POS performed at 24–36 h and 48–72 h of life for early detection of CCHD in neonates.Methods:This prospective observational study was conducted over 20 months at a tertiary care neonatal intensive care unit (NICU) in India. A total of 1431 neonates, including term and preterm, underwent POS at two time points. Screening followed the modified American Academy of Pediatrics algorithm, with echocardiographic follow-up for positive screens. Sensitivity, specificity, positive predictive value, and negative predictive value (NPV) were calculated.Results:Of the enrolled neonates (n = 1431), dual screening identified 95 positives (6.64%): 11 CCHD, 38 noncritical CHD and 46 noncardiac hypoxemic conditions. Dual POS had higher sensitivity (73.33% vs. 60%) and NPV (99.7% vs. 99.54%), and a lower false-negative rate (26.66% vs. 40%). POS 2 detected 2 additional CCHDs missed by POS1. Among 1336 negatives, clinical follow-up found 4 CCHDs.Conclusion:Dual POS significantly improves CCHD detection over single screening while identifying treatable noncardiac hypoxemic conditions. This nurse-led protocol is feasible even in resource-constrained NICUs, warranting routine adoption alongside clinical examination to minimize missed diagnoses of CCHD. INTRODUCTION Critical congenital heart disease (CCHD) represents a significant cause of neonatal morbidity and mortality worldwide, with a prevalence of approximately 1–2 per 1000 live births.[1] Many CCHDs remain asymptomatic or exhibit subtle signs in the early postnatal period, often evading detection through routine clinical examination alone, leading to delayed diagnosis and potential circulatory collapse upon ductal closure. In the resource-limited settings such as India, where prenatal detection rates are low and hospital stays are short, effective postnatal screening strategies are essential to improve the outcomes.[2] Pulse oximetry screening (POS) has emerged as a simple, noninvasive tool for detecting CCHD by identifying hypoxemia, with meta-analyses reporting sensitivities of around 76.5% and specificities exceeding 90%.[345] However, the optimal timing remains controversial; screening before 24 h yields higher false positives due to transitional circulation, while late screening (>24 h) reduces them but may delay detection.[678] Dual screening protocols, such as 24–36 and 48–72 h, aim to balance sensitivity and specificity, minimizing false negatives and capturing noncardiac hypoxemic conditions such as pulmonary disorders or infections. This Critical Heart Disease Oximetry Observation in Screening Examination (CHOOSE study) for neonates evaluates the diagnostic accuracy of dual POS (POS1:24–36 h and POS2:48–72 h) in a tertiary care neonatal intensive care unit (NICU) in Central India, addressing a gap in evidence from the low-resource settings where such protocols could enhance newborn care feasibility by the timely detection of CCHD, noncritical CHDs, and noncardiac hypoxemic conditions. METHODS Study design and setting This prospective observational study was conducted from September 1, 2022, to June 30, 2024, at a tertiary care hospital in Central India after obtaining ethical approval from the Institutional Ethics Committee (IEC no. 2022/8929-40). Participants Eligible participants included term (>37 weeks gestational age) and late preterm (34–36 weeks gestational age) neonates who were born at this center or referred within 24 h of birth, clinically stable without the need for oxygen support beyond the first 24 h. Exclusion criteria were neonates requiring prolonged respiratory support after 24 h of life, prenatally diagnosed CHD, major congenital anomalies or chromosomal syndromes, and discharge from the hospital before 48 h of life. After fulfilling the inclusion and exclusion criteria, 1431 eligible neonates were enrolled in this study [Figure 1]. Flow diagram Variables The primary variable was the detection of CHD and the secondary variable was the identification of noncritical CHD and noncardiac hypoxemic conditions. Screening protocol All recruited neonates underwent POS twice; POS1: between 24–48 h of life and POS2: between 48–72 h of life. Screenings were performed by trained registered nurses using a Masimo Radical-7 pulse oximeter. Sensor placement was preductal (right hand) and postductal (left foot), with readings taken in a resting state until a stable pulse waveform was achieved. The highest stable SpO2 value was recorded for each site. Interpretation criteria (per modified American Academy of Pediatrics algorithm)[9] Positive screen SpO2 <90% in right hand or left foot, SpO2 <95% in both sites on two measurements 1 h apart, persistent pre- and postductal difference >3% on two measurements 1 h apart. Negative screen SpO2 ≥95% in right hand or left foot and pre /post ductal difference ≤3%. Newborns with positive screens were referred for echocardiographic evaluation by a pediatric cardiologist. Neonates diagnosed with CCHD were referred to the cardiac center for further management. Neonates with noncritical CHD were counseled regarding follow-up in the cardiac center after discharge. Noncardiac hypoxemia prompted targeted investigations, including chest X-ray, blood cultures, sepsis work-up, and were managed accordingly. Data sources Maternal and neonatal details were recorded, including demographics, gestational age (determined via clinical examination and antenatal records), birth weight, mode of delivery, oximetry values (POS1 and POS2), screening outcomes, and echocardiographic findings. Management records of noncardiac hypoxemic conditions included length of stay, need for O2 support/continuous positive airway pressure (CPAP)/high-flow nasal cannula (HFNC)/mechanical ventilation (MV), and length of stay (LOS) in the hospital. Statistical methods Data were entered in Microsoft Excel and analyzed using the SPSS software version 22 (IBM Corp., Armonk, NY, USA). The continuous variables were expressed as mean (standard deviation) or median (interquartile range [IQR]) and categorical variables as frequencies or percentages. Diagnostic accuracy metrics-sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), false positive rate (FPR) and false negative rate (FNR) using echocardiography confirmed cases of CCHD. Primary outcome: CCHD (ductal-dependent lesions or those requiring intervention <28 days). Secondary outcomes: noncritical CHD (cardiac lesion other than CCHD) and noncardiac hypoxemic conditions. RESULTS Baseline characteristics A total of 1431 neonates were enrolled, comprising 652 males (45.56%) and 779 females (54.44%). Most were born out of normal vaginal deliveries (n = 905; 63.24%), with cesarean sections accounting for 518 cases (36.21%) and assisted deliveries numbering 8 (0.56%). Median gestational age was 38 weeks (IQR: 37–39) and mean birth weight was 2.68 + 0.42 kg. Of these, 1242 (86.79%) were term (>37 weeks), while 189 (13.20%) were late preterm (34–36 weeks). Screening outcomes POS1 detected 9 CCHD, 23 noncritical CHD and 17 noncardiac hypoxemic conditions. The sensitivity and specificity of POS1 for detecting CCHD were 60% and 96.54%, respectively, while PPV was 15.51%, and the NPV was 99.56%. FPR and FNR were 3.46% and 40%, respectively. POS2 detected 2 additional CCHD, 15 noncritical CHD and 29 noncardiac hypoxemic conditions. Thus, dual screening (combined POS1 and POS2) resulted in the identification of 11 CCHD, 38 noncritical CHD and 46 hypoxemic conditions. The sensitivity and specificity of dual screening for detecting CCHD were 73.33% and 94.06%, respectively while the PPV was 11.57% and the NPV was 99.7%. Dual screening had an FPR of 5.93% and an FNR of 26.66%. Congenital heart disease in neonates with positive pulse oximetry screening Of 95 neonates found to be positive after dual screening, 58 had positive POS1, while POS2 detected 37. Dual screening identified 11 cases of CCHD and 38 cases of noncritical CHD. Details of the CCHD and noncritical CHDs identified with POS are mentioned in Table 1. Table 1Echocardiographic findings in neonates with positive pulse oximetry screening testCHDsCombined POS1 and POS2POS1POS2Critical CHDs DTGA440 Unbalanced AVSD, severe PS101 Tricuspid atresia330 Cardiac TAPVC101 HLH220Non-critical CHDs Dysplastic pulmonary valve with PFO101 Dextrocardia with PFO110 ASD alone220 Bicuspid aortic valve with ASD101 Cardiomyopathy with PFO101 AVSD413 VSD16115 PDA1284ASD: Atrial septal defect, AVSD: Atrioventricular septal defect, CHD: Congenital heart disease, DTGA: Dextro-transposition of the great arteries, HLH: Hypoplastic left heart syndrome, PDA: Patent ductus arteriosus, PFO: Patent foramen ovale, POS1: Pulse oximetry screening at 24–36 h, POS2: Pulse oximetry screening at 48–72 h, PS: Pulmonary stenosis, VSD: Ventricular septal defect, TAPVC: Total anomalous pulmonary venous connection Echocardiographic findings in neonates with positive pulse oximetry screening test ASD: Atrial septal defect, AVSD: Atrioventricular septal defect, CHD: Congenital heart disease, DTGA: Dextro-transposition of the great arteries, HLH: Hypoplastic left heart syndrome, PDA: Patent ductus arteriosus, PFO: Patent foramen ovale, POS1: Pulse oximetry screening at 24–36 h, POS2: Pulse oximetry screening at 48–72 h, PS: Pulmonary stenosis, VSD: Ventricular septal defect, TAPVC: Total anomalous pulmonary venous connection Congenital heart disease in neonates with negative pulse oximetry screening Of 1336 neonates with negative POS after dual screening, 4 cases of CCHD and 17 cases of noncritical CHD were identified on echocardiography done after getting abnormalities such as heart murmur, pulse abnormality, positive family history, and abnormal chest X-ray suggesting cardiac abnormality [Tables 2 and 3]. Table 2Presentation of neonates with congenital heart diseases and a negative pulse oximetry screening testDetectionCritical CHDNon-critical CHDHeart murmur014Family history of CHD13Abnormal Chest X-ray10Abnormality in pulse20CHD: Congenital heart disease Presentation of neonates with congenital heart diseases and a negative pulse oximetry screening test CHD: Congenital heart disease Table 3Echocardiographic findings in neonates with negative pulse oximetry screening testCHDs n Critical CHDs Nonobstructive TAPVC-CS type1 Single ventricle physiology with Severe PAH, no PS1 COA2Non-critical CHDs Bicuspid aortic valve2 PM VSD4 Mid muscular VSD1 PDA5 Isolated severe PS2CHD: Congenital heart disease, COA: Coarctation of the aorta, PAH: Pulmonary arterial hypertension, PDA: Patent ductus arteriosus, PS: Pulmonary stenosis, VSD: Ventricular septal defect, TAPVC-CS: Total anomalous pulmonary venous connection-coronary sinus, PM: Perimembranous Echocardiographic findings in neonates with negative pulse oximetry screening test CHD: Congenital heart disease, COA: Coarctation of the aorta, PAH: Pulmonary arterial hypertension, PDA: Patent ductus arteriosus, PS: Pulmonary stenosis, VSD: Ventricular septal defect, TAPVC-CS: Total anomalous pulmonary venous connection-coronary sinus, PM: Perimembranous Noncardiac conditions with positive pulse oximetry screening Of 95 neonates with positive POS after dual screening, 46 had a noncardiac abnormality. POS1 detected 17 and POS2 identified an additional 29 such cases. Of 46 noncardiac hypoxemic conditions, 2 with pneumothorax, 16 with sepsis, and 28 with lower respiratory tract infection were diagnosed on further workup. Out of these patients, 1 required MV, 4 CPAP, 3 HFNC, and 12 O2 support [Table 4]. Table 4Neonates with noncardiac abnormalities (n=46) detected after positive pulse oximetry screeningCondition n Median LOS (days)InterventionPneumothorax2101 MVSepsis167AntibioticsLRTI2854 CPAP, 3HFNC, 12 O2 support, antibioticsTotal4641MV, 4 CPAP, 3HFNC, 12 O2 supportCPAP: Continuous positive airway pressure, HFNC: High flow nasal cannula, LOS: Length of stay, LRTI: Lower respiratory tract infection, MV: Mechanical ventilation, O2: Oxygen Neonates with noncardiac abnormalities (n=46) detected after positive pulse oximetry screening CPAP: Continuous positive airway pressure, HFNC: High flow nasal cannula, LOS: Length of stay, LRTI: Lower respiratory tract infection, MV: Mechanical ventilation, O2: Oxygen DISCUSSION This study highlights the concept of dual screening of neonates with POS, addressing the debate on timing of neonatal screening with oximetry. This study adds meaningful evidence for a practical screening tool, particularly in low- and middle-income countries (LMICs). POS is already established as an effective tool, with known influences of timing (early screening yields more false positives, while later screening reduces false positives but risks missing early disease), limitations in detecting left-sided obstructive lesions, and potential to identify noncardiac hypoxemia.[4591011] Most prior studies have compared single screenings before versus after 24 h.[111213141516] In contrast, this study provides novel evidence for sequential dual screening at 24–36 h and 48–72 h – both after transitional circulation has stabilized – a strategy that remains underexplored yet holds substantial promise, especially in LMIC NICU settings. This is one of the few prospective studies evaluating two post-24-h screens (especially in LMIC NICU settings). Key findings and comparisons This study provides compelling evidence for the efficacy of dual POS performed at 24–36 h of life (POS1) and 48–72 h of life (POS2) in detecting CCHD among 1431 neonates including both term and late-preterm, identifying 11 CCHDs, 38 noncritical CHD, and 48 noncardiac hypoxemic conditions, yielding a combined sensitivity of 73.33%, specificity of 94.06%, PPV of 11.57%, and NPV of 99.7% – a marked improvement over single POS1 (sensitivity 60%, specificity 96.54%). POS2 uniquely detected 2 additional CCHDs, 15 noncritical CHDs and 29 noncardiac hypoxemic conditions, highlighting how sequential screening mitigates false negatives from transitional circulation, where elevated pulmonary vascular resistance can mask hypoxemia early on. These results resonate with previous literature on POS.[1011121314151617] Thangaratinam et al.’s meta-analysis of 13 high-quality studies encompassing 229,441 newborns, which established POS sensitivity at 76.5% (95% CI: 67.7%–83.5%) and specificity >99% for CCHD.[417] Notably, in most of the studies, sensitivity remained consistent before or after 24 h, but false-positive rates dropped significantly later. However, in this study, both the screening (POS1 and POS2) was done after 24 h, and therefore improvement in sensitivity and reduction in false negatives have been evident. False positives were higher with dual screening, a pattern suggesting the detection of an evolving noncardiac hypoxemic condition by sequential screening. Large scale validations, such as Poland’s POLKARD program (51,698 first day postductal screens) and China’s prospective study (122,738 combined pre/post screens at 6–72 h), reported comparable metrics, reinforcing POS’s robustness across populations.[1819] In our cohort, the 95 positive screens (6.6%) prompted timely cardiology referrals and subsequent echocardiography, detecting 11 CCHDs, thus averting potential ductal closure related collapses, a risk affecting up to 26%–33% of undiagnosed CCHDs per Wren et al. and Mellander studies.[2021] Feasibility and cost-effectiveness in resource-limited settings Dual POS represents a highly feasible, low-cost intervention requiring minimal workforce beyond routine nursing duties already performed in NICUs. Using widely available pulse oximeter devices for NICUs, this nurse-led protocol adds negligible additional cost while leveraging existing echocardiographic pathways. The incremental echocardiography burden remains low at <2.1% (21 additional echoes per 1000 live births), far outweighed by preventing ductal-dependent collapses that demand emergency interventions costing 10–50 times more. In LMICs where single screening misses ~40% of CCHDs (as seen with POS1 alone), this sequential approach delivers substantial value – detecting 2 extra CCHDs and 29 noncardiac emergencies missed by day-2 discharge protocols – without specialized training or infrastructure upgrades. Based on this dual POS offers a cost-effective approach for CCHD identification. Noncardiac hypoxemia detection A key strength and broader utility of POS lies in its detection of noncardiac hypoxemia, which outnumbered CCHDs (46 vs. 11) in this study. Of these, 28 lower respiratory tract infections, 16 sepsis cases and 2 pneumothorax cases required interventions (1 MV, 4 CPAP, 3 HFNC, and 12 on O2 support), echoing Meberg et al.’s findings, where 41% of first-day false positives signaled urgent extra-cardiac issues such as transient tachypnea, Group B streptococcal sepsis, or pulmonary hypertension.[622] Ewer et al. noted 27% of positives needed intervention, while Zhao et al. and others affirmed POS as a “side benefit” screener for hidden pulmonary pathology, often outnumbering CCHDs.[1223] In this study, FPR was higher for dual screening (5.93% vs. 3.46%), probably indicating the detection of evolving hypoxemic noncardiac conditions, as in the early course of the disease, these conditions are likely to be missed; moreover, this study recruited relatively healthy newborns, thus highlighting an important aspect of missing these conditions with single screening or with early (<48 h) discharge. Detection of noncardiac hypoxemic conditions was higher with dual screening, as compared to POS1 alone, with 29 additional such abnormalities detected by POS2. This translates to an additional burden of echocardiography equivalent to <2.1% (21 per 1000 live births) in this study. Minimizing false negatives with clinical examination Several studies have been published on the CCHD detection in patients with positive screening; however, this study additionally assessed the CCHD detection by echocardiography prompted by any clinical clue in the history, physical examination or abnormal chest X-ray in those with negative POS. This highlighted the limitations in POS sensitivity for specific lesions, including 4 missed CCHDs (nonobstructive TAPVC-CS type, single ventricle with severe PAH, 2 cases of coarctation of the aorta) and 17 noncritical CHDs, identified via clinical clues (heart murmur, pulse abnormalities, positive family history, and abnormal chest X-ray). Left-sided obstructive defects such as coarctation of the aorta (CoA) show only 29%–50% detection rates across studies due to subtle pre-/post-ductal gradients early on, with 82% of Norwegian misses being such lesions in a study.[12242526] Noncritical CHDs missed on POS typically exhibit left-right shunting with normal SpO2, detectable clinically, as in this study, 14 such cases were screened by heart murmur. Integrating POS with routine examinations boosts pre-discharge CCHD detection to ~90%, as in Swedish (8%) and Norwegian (12%) cohorts.[242526] Prenatal ultrasound gaps (~30% detection in unselected populations) position POS as a vital postnatal safety net, capturing antenatally missed CHDs.[626] Limitations There are certain limitations of this study. Although this study reflects the real-world challenges, it limits generalisability. This study, with 45.56% males, 63.24% born out of vaginal deliveries, and exclusion of prenatally diagnosed CHD or clinical instability ensured a stable cohort. However, no long-term morbidity and mortality data were tracked. High-altitude adjustments for SpO2 cutoffs were unnecessary here, but relevant for broader adoption. This study excluded the sick neonates to focus more on screening and not diagnosis; moreover, frequent pulse oximetry along with close monitoring of vital parameters is routinely practised in sick neonates, thus providing a more suitable study population for this study by excluding them. Compared to U.S (>24 h American Academy of Paediatrics (AAP) protocols) or European first-day strategy (Nordic, Polish), our dual mid-period timing balances the advantages and disadvantages of the protocols, although first-day screening has a faster detection rate (median detection 6–17 h) Future directions and implications Single-center data limits generalizability; multi-center trials should validate dual POS in preterm/high-risk groups, preferably by incorporating perfusion index for left-heart lesions (promising per Granelli), or computerized algorithms to curb interpretation errors (81.6% manual vs. 98.3% automated accuracy).[2425262728] Standardized training, motion-stable devices, and database integration would enhance quality assurance. Routine adoption of this strategy, paired with clinical exams, promises reduced late diagnoses, fewer collapses, and holistic neonatal surveillance – critical amid global CCHD burdens. CONCLUSION Dual POS at 24–36 h and 48–72 h of life markedly improves the detection of CCHD (73.33% sensitivity) with high specificity (94.06%) and uncovers treatable noncardiac hypoxemic conditions in resource-limited NICU settings. These findings support routine adoption of sequential screening protocols in LMICs, where prenatal diagnostics remain limited (~30% detection) and early discharge is common. Combined with physical examination, this approach could reduce ductal-dependent collapses by bridging gaps in single-screen strategies. Adopting dual pulse oximetry as standard newborn care promises fewer missed diagnoses and better neonatal outcomes worldwide. Ethics declarations This study was approved by the institutional ethical committee of NSCB Medical College, Jabalpur, MP (IEC no. 2022/8929-40). Written informed consent was obtained from parents or legal guardians prior to enrollment, in accordance with the Declaration of Helsinki. Conflicts of interest There are no conflicts of interest.