This was a cross-sectional analytical study with prospective data collection. The study was conducted over a six-month period, from 1 March to 31 August 2022, in the maternity wards of two facilities in Brazzaville: the Brazzaville University Hospital Centre (CHU-B) and the Talangaï Reference Hospital (HRT).

The target population was newborns of at least 34 weeks’ gestational age born in Brazzaville.

The source population included mother-newborn pairs in the immediate postpartum period within the delivery and obstetrics units of the CHU-B and the HRT.

2.3.1. Inclusion Criteria

We included in the study newborns with a gestational age of at least 34 weeks of amenorrhea and a birth weight of at least 2000 grams.

2.3.2. Exclusion Criteria

We excluded newborns with poor adaptation to extrauterine life (requiring immediate advanced resuscitation or transfer to neonatal intensive care), those born to mothers with known neuropsychiatric disorders, and those born to mothers who did not provide consent for participation.

2.4.1. Type of Sampling

A two-stage sampling procedure was used. In the first stage, a simple random sample was drawn from the eight hospitals in Brazzaville, resulting in the selection of two hospitals: Brazzaville University Hospital and Talangaï Referral Hospital. These hospitals constituted the primary sampling units. In the second stage, for each selected hospital, the days of visit were randomly chosen. Finally, all mother-newborn pairs present on the days of visit were consecutively included in the study, constituting the secondary sampling units.

2.4.2. Sample Size

The minimum sample size was calculated using the SCHWARTS formula:

N = P(1 − P)(Zα)²/(m)²

N: minimum sample size.

P: Estimated proportion of the population with the characteristic under study, which is 7.4% based on data from Philippe Mahouna ADJAGBA’s Beninese study on the contribution of pulse oximetry in the screening for cyanogenic congenital heart disease in newborns at CNHU-HKM Cotonou [21].

Zα = 1.96 (corresponding to a 95% confidence level).

m = 5% (tolerated margin of error).

Thus, the minimum sample size should be 105. After accounting for an anticipated non-response rate of 15%, the final minimum required sample size was increased to 121 newborns.

2.5.1. Collection Tools and Sources

Data were collected using a standardized, anonymous survey form. Information was obtained through individual interviews with mothers, antenatal care records, medical records, hospitalization and delivery registers, surgical reports, clinical examinations, pulse oximetry measurements, and cardiac Doppler ultrasound examinations of newborns

2.5.2. Collection Procedure

Face-to-face interviews with mothers were conducted by the same interviewer in a private setting, after obtaining parental consent. The interview was followed by a physical examination of the newborn, during which the interviewer also took pulse oximetry readings.

Pulse oximetry was performed using a PC60E pulse oximeter equipped with a neonatal sensor. Measurements were taken while the newborn was calm and at rest. Two measurements were obtained: one from the right upper limb (finger or wrist) and one from a lower limb. The difference between the two saturation values was then calculated.

2.6.1. Interpretation of Pulse Oximetry Results

Pulse oximetry screening was considered positive if oxygen saturation was ≤90% in either the upper or lower limb, or if the difference in saturation between the upper and lower limbs exceeded 3%.

Screening was considered negative if oxygen saturation was ≥95% in both the upper and lower limbs and the difference between the two measurements was less than 3%.

Screening was considered inconclusive if both saturation values were between 90% and 95% and the difference was ≤3%. In such cases, a second measurement was performed. If the results remained unchanged, the screening was ultimately classified as positive.

2.6.2. Doppler Echocardiography Procedure

Cardiac Doppler ultrasound was performed in all newborns using a multi-frequency pediatric probe connected to a tablet running the Philips Lumify application (version 1950105.428118487). All examinations were carried out by the same pediatric cardiologist, who was blinded to the pulse oximetry results.

2.6.3. Cardiac Doppler Ultrasound Results

Findings were categorized as follows:

No congenital heart disease: absence of structural cardiac abnormalities.

Transitional congenital heart disease: this referred to congenital heart conditions observed during the transition period between foetal circulation and mature circulation, namely patent foramen ovale and restrictive ductus arteriosus. These heart conditions were considered non-pathological and as absence of congenital heart disease in the data analysis.

Congenital heart disease: all other structural congenital cardiac anomalies.

2.6.4. Interpretation of Combined Oximetry and Ultrasound Results

True positive: positive pulse oximetry screening and presence of congenital heart disease on ultrasound.

False positive: positive pulse oximetry screening with no congenital heart disease on ultrasound.

True negative: negative pulse oximetry screening and no congenital heart disease on ultrasound.

False negative: negative pulse oximetry screening with congenital heart disease confirmed on ultrasound.

The dependent variable or variable of interest was pulse oximetry.

The independent (explanatory) variables included maternal, neonatal, clinical, and ultrasound-related factors:

Maternal variables: maternal age, household socioeconomic status, history of chronic maternal disease, family history of congenital heart disease or other congenital malformations, medication use during pregnancy, and type of pregnancy (singleton or twin).

Pregnancy and delivery variables: number and quality of antenatal care visits, gestational age at delivery, and mode of delivery.

Neonatal variables: Apgar score, birth weight, postnatal age at screening, sex, skin colour, respiratory rate (RR), heart rate (HR), pulse symmetry, heart rhythm, presence of a cardiac murmur, and characteristics of the first and second heart sounds (S1 and S2).

Ultrasound variables: number of echocardiographic examinations and results, including the presence or absence of congenital heart disease (CHD), type of CHD, and CHD classification (cyanotic or non-cyanotic).

Non-cyanotic congenital heart disease: congenital heart disease in which haemodynamic occur without mixing of arterial and venous blood in the systemic circulation (e.g. VSD, ASD, PCA, atrioventricular canal, coarctation of the aorta).

Cyanotic congenital heart disease: congenital heart disease with mixing of arterial and venous blood in the systemic circulation. (e.g., Tetralogy of Fallot, Transposition of the Great Arteries).

Microsoft Excel (version 10) and SPSS Statistics (version 17.0, 2010) were used for data entry, database management, and the production of tables and graphs. Results for categorical variables are presented as frequencies (n) and percentages (%). For quantitative variables, data are expressed as medians with interquartile ranges (Q1 - Q3) for non‑normal distributions, and as means ± standard deviations for normally distributed variables

We used the student’s t-test to compare means and the Mann-Whitney U test to compare medians. the variable of interest (pulse oximetry) was correlated with each explanatory variable. Odds ratios (OR) with their 95% confidence intervals (CI) were calculated, with statistical significance set at p < 0.05.

The diagnostic performance of pulse oximetry was assessed by calculating sensitivity, specificity, and positive and negative predictive values.

The study was conducted in compliance with patient anonymity and privacy. All participants provided informed consent prior to data collection. Ethical clearance was obtained from the Health Sciences Research Ethics Committee (N˚ 0036-28/MESRSIT/DGRST/CERSSA/-22).

During the study period, congenital heart disease (CHD) was identified in 12 out of 300 newborns screened at CHU-B and HRT, yielding a prevalence of 4%.

The sociodemographic characteristics of the mothers and newborns are reported in Table 1.

Gestational characteristics, parity, and quality of prenatal care:Fifty-six percent (56%) of the mothers were primigravida. All mothers who underwent obstetric ultrasound had normal findings.

Gestational characteristics, parity, and the quality of prenatal care are summarized in Table 2.

Table 1. Sociodemographic characteristics of mothers and newborns.

Table 2. Pregnancy, parity and quality of mothers’ prenatal contacts.

Pulse oximetry screening was negative (SpO₂ ≥ 95%) in 85.7% of newborns. An inconclusive result (SpO₂ between 91% and 94%) was observed in 8% of cases, while 6.3% of newborns had a positive screening result (SpO₂ ≤ 90%).

All newborns (8%) with an initial inconclusive pulse oximetry reading (SpO₂ 91% - 94%) had a positive result upon repeat measurement.

Regarding cardiac Doppler ultrasound, 72% of newborns had normal results, 24% had transitional abnormalities, and 4% of newborns had congenital heart disease.

The congenital heart defects identified included atrial septal defect (ASD) in 33.3% of cases, followed by ventricular septal defect (VSD), complete atrioventricular canal defect (CAVD), and tetralogy of Fallot, each accounting for 16.7%. Truncus arteriosus and double-outlet right ventricle each represented 8.3% of cases.

3.5.1. Performance of Pulse Oximetry for Screening All Congenital Heart Diseases

The performance of pulse oximetry for detecting all CHD are summarized in Table 3.

Table 3. Performance of pulse oximetry for the screening of all congenital heart diseases.

1) Sensitivity: 41.7%; 2) Specificity: 86.8%; 3) Positive predictive value: 11.6%; 4) Negative predictive value: 97.2%; 5) Positive likelihood ratio: 4; 6) Negative likelihood ratio: 0.5.

3.5.2. Performance of Pulse Oximetry According to Congenital Heart Disease Group

Of the 12 congenital heart diseases diagnosed, 8 (66.7%) were non-cyanotic and 4 (33.3%) were cyanotic.

Performance of pulse oximetry for screening for non-cyanotic congenital heart disease

The performance of pulse oximetry in screening for non-cyanotic CHD is detailed in Table 4.

Table 4. Performance of pulse oximetry for non-cyanogenic congenital heart disease.

1) Sensitivity: 12.5%; 2) Specificity: 85.4%; 3) Positive predictive value: 2.3%; 4) Negative predictive value: 97.2%.

Performance of pulse oximetry for screening cyanotic congenital heart disease

The performance of pulse oximetry in screening for cyanotic CHD is detailed in Table 5.

Table 5. Performance of pulse oximetry for cyanogenic congenital heart disease.

1) Sensitivity: 100%; 2) Specificity: 86.7%; 3) Positive predictive value: 9.3%; 4) Negative predictive value: 100%.

Sex and age at screening for congenital heart disease using pulse oximetry:In our series, a slight male predominance was observed, with 54.3% of cases (sex ratio 1.2). This finding is consistent with the results of Majani in Tanzania and Janjua Zakarimanana, who documented male proportions of 52.4% and 52.6%, respectively [24][26].

No scientific explanation has been put forward to account for this slight overrepresentation of males.

The timing of pulse oximetry screening is flexible after birth. However, the optimal timing depends mainly on each country’s health and socio-cultural context, particularly the usual length of stay in maternity wards, the availability of healthcare staff, and parental acceptance of the test. In settings where early discharge from maternity wards is common, screening within the first 24 hours of life is recommended to maximise coverage.

This approach also facilitates earlier detection of serious congenital heart disease (CHD), nearly half of which becomes symptomatic within the first 24 hours of life. Among these cases, 9% present with cardiogenic shock, as reported by de-Wahl Granelli et al. and Riede et al. [27][28]. However, early screening is associated with a higher rate of false-positive results [29][30].

Despite this limitation, the American Academy of Pediatrics (AAP) recognised in 2020 the clinical value of pulse oximetry screening within 24 hours of birth in contexts of early postnatal discharge [31]. This recommendation is also endorsed by the European Pulse Oximetry Screening Workgroup [17]. However, in China, the screening window is broader wider, ranging from six to 72 hours after birth [32].

In line with this strategy, and to account for local organizational constraints and a high rate of early maternity discharges, over half of the newborns in our study (55.3%) were screened between 12 and 24 hours after birth. This timing aimed to optimize coverage.

Sensitivity and specificity of pulse oximetry in screening for congenital heart disease:Although pulse oximetry is a valuable screening tool for CHD, it cannot detect all forms of the disease [15]. Pulse oximetry’s principle of detecting hypoxemia means it is most effective in identifying cyanotic congenital heart diseases. In our study, the overall sensitivity and specificity of pulse oximetry for detecting all CHD were 41.7% and 86.8%, respectively. These figures are lower than the ranges reported by most authors (sensitivity: 67% - 72%; specificity: 90% - 100%) [30][33]. This discrepancy may be partly attributed to our smaller sample size. Furthermore, reference studies often focus specifically on critical CHD, which is predominantly hypoxemic. When our analysis was restricted to cyanotic heart disease, our results improved markedly (sensitivity: 100%; specificity: 86.7%), aligning with or surpassing published data [29][32]. Conversely, the diagnostic performance for non-cyanotic CHDs was significantly lower (sensitivity of 12.5%, specificity of 85.4%). Similarly, the positive predictive value in our cohort was 11.6%, which is lower than values reported elsewhere [30][33].

In our study, screening for all congenital heart diseases yielded 38 false positive cases (12.8%). This rate increased slightly to 13.1% (39 cases) for cyanotic heart disease screening specifically, and to 14.2% (42 cases) for non-cyanotic heart disease screening. These rates are significantly higher than those reported in the literature. For instance, Adjagba in Benin and Plana and al. reported rates of 0.04% (one case) and 0.14%, respectively [16][21].

Several factors may explain this discrepancy. In our cohort, more than half of the newborns were screened between 12 and 24 hours of life. This period is characterized by persistent transitional circulation and physiological pulmonary arterial hypertension, which can temporarily lower oxygen saturation. Other studies support this hypothesis, showing an association between early screening (conducted within the first 24 hours) and a higher false-positive rate. For instance, a systematic review by Thangaratinam et al. found that screening performed before 24 hours of live was associated with a 10% increase in the false-positive rate [30][34][35]. However, the timing of screening alone is unlikely to fully explain the markedly elevated false-positive rate observed in our study. Additional contributing factors may include technical conditions during measurement, such as neonatal movement, inadequate peripheral perfusion, low skin temperature, and the type or placement of the sensor.

Regarding false negatives, our study identified six cases. In contrast, the studies by Arlettaz et al. and Adjagba et al. reported none [21][23]. This discrepancy is likely attributable to our study’s inclusion of non-cyanotic congenital heart disease. These conditions often evade detection by pulse oximetry, as they typically do not cause hypoxemia.

A total of 12 cases of congenital heart disease were diagnosed in our study. Atrial septal defect (ASD) was the most frequent lesion, accounting for 33.3% of cases, followed by ventricular septal defect (VSD) (16.7%), complete atrioventricular septal defect (16.7%), tetralogy of Fallot (16.7%), truncus arteriosus (8.3%), and double outlet right ventricle (8.3%).The spectrum of congenital heart disease observed in this study is comparable to that reported by Bakr et al. [36]. However, comparisons with other studies are limited, as most investigations focus exclusively on screening for critical congenital heart disease, which restricts the diversity of lesions identified.

Six of the congenital heart defects diagnosed in our study were associated with a risk of haemodynamic decompensation within the first month of life. Pulse oximetry successfully detected five of these cases. Overall, pulse oximetry identified all cases of cyanotic congenital heart disease but detected only one of the eight cases of non-cyanotic congenital heart disease.