Article Text

This article has a correction. Please see:

Download PDFPDF

Pulse oximetry as a screening tool for detecting major congenital heart defects in Indian newborns
  1. Anita Saxena1,
  2. Anurag Mehta1,
  3. Sivasubramanian Ramakrishnan1,
  4. Mamta Sharma2,
  5. Sudha Salhan3,
  6. M Kalaivani4,
  7. Rajnish Juneja1
  1. 1Department of Cardiology, All India Institute of Medical Sciences, New Delhi, India
  2. 2Department of Pediatrics, Columbia Asia Hospital, Gurgaon, Haryana, India
  3. 3Department of Obstetrics and Gynaecology, Hindu Rao Hospital, New Delhi, India
  4. 4Department of Biostatistics, All India Institute of Medical Sciences, New Delhi, India
  1. Correspondence to Professor Anita Saxena, Department of Cardiology, All India Institute of Medical Sciences, Room # 29, Ansari Nagar, New Delhi 110029, India; anitasaxena{at}hotmail.com

Abstract

Objective To evaluate the use of pulse oximetry as a screening tool for detecting major congenital heart defects (CHDs) in Indian newborns.

Design Cross-sectional observational study.

Patients In a community hospital of north India, babies born during a specific 8 h period of the day were recruited over a period of 3 years. Newborns with incomplete documentation were excluded.

Intervention Routine clinical examination, pulse oximetry and bedside echocardiography.

Outcome measures Any abnormalities in clinical examination and pulse oximetry were recorded. CHDs were diagnosed using bedside echocardiography. Accuracy of pulse oximetry, clinical examination and their combination for detecting major CHDs was calculated.

Results Among the 19 009 newborns screened, 70 had major CHDs at birth (44 serious, 26 critical). Pulse oximetry detected 39 major (sensitivity 55.7%, 95% CI 44.1% to 66.8%; specificity 68.3%, 67.6% to 68.9%) and 22 critical CHDs (sensitivity 84.6%, 66.5% to 93.9%; specificity 68.3%, 67.6% to 68.9%). Addition of pulse oximetry to clinical examination significantly improved sensitivity for major CHDs (35.7% (25.5% to 47.4%) to 75.7% (64.5% to 85.3%), p<0.01) and critical CHDs (11.5% (4.0% to 29.0%) to 84.6% (66.5% to 93.9%), p<0.01).

Conclusions Pulse oximetry is a sensitive screening tool for detecting major CHDs in Indian newborns. It adds significant value to the current practice of using clinical examination as a sole screening tool for detecting major CHDs. However, specificity of pulse oximetry was much lower in our study. Possible reasons for low specificity could be non-repetition of pulse oximetry in newborns with initial lower saturations, high prevalence of infections and respiratory issues in our cohort and use of non-motion tolerant oximeter.

  • Neonatology
  • Screening
  • Cardiology
  • Congenital Abnorm
View Full Text

Statistics from Altmetric.com

What is already known on this subject?

  • Many studies from high-income countries have shown that pulse oximetry is a sensitive screening tool for detecting major congenital heart defects (CHDs) in newborns.

  • A few countries have implemented the approach of universally screening newborns for CHDs using pulse oximetry and clinical examination.

  • Pulse oximetry done after 24 h of birth has lower false-positive rates.

What this study adds?

  • Pulse oximetry combined with clinical examination is a sensitive screening tool for detecting major CHDs in middle-income countries as well.

  • Pulse oximetry screening can be of more value in low-income and middle-income countries because of higher disease burden and very low antenatal detection rates for CHDs.

  • Further studies are needed to determine the correct screening regimen before pulse oximetry can be universally used for CHD screening in the developing world.

Introduction

Congenital heart defects (CHDs) are a leading cause of neonatal and infant mortality in the developed world, but data from low-income and middle-income countries are scarce.1 CHDs are among the commonest congenital malformations with nearly 15% being life threatening during the neonatal period, and almost one-third to half requiring surgical or transcatheter intervention within the first year of life.1–3 The birth prevalence of CHDs worldwide is regarded as 8 per 1000 live births.4 Using this prevalence data, approximately 180 000 children in India are born with CHDs annually. Of these, nearly 60 000–90 000 suffer from critical CHDs that require early intervention.5 Missed and delayed diagnosis of these critical defects account for a significant increase in morbidity and mortality.6 ,7 This is especially true for newborns with duct-dependent critical defects where closure of the duct leads to acute cardiovascular decompensation and death.8–10

Routine clinical examination of newborns lacks sensitivity for detecting CHDs.11 ,12 Pulse oximetry is an accurate, non-invasive test used for quantifying hypoxaemia that has been used in several large-scale studies to screen for CHDs.13–19 All of these studies, except one from China, are from high-income countries. The rationale for using pulse oximetry is that most critical CHDs produce some degree of hypoxaemia that is detectable by oximetry, but does not produce clinically visible cyanosis.16 Pulse oximetry has been reported as a highly specific test with moderate sensitivity for detecting critical CHDs, making it an ideal test for CHD screening.20 Also, it has a lower false-positive rate for detecting critical CHDs when testing is done after 24 h of birth.20

Though echocardiography remains the gold standard for CHD diagnosis, using pulse oximetry with clinical examination as a screening regimen can aid in the early detection of critical CHDs. Presuming the availability of paediatric cardiology centres close to the maternity hospitals, this regimen can be useful in low-income and middle-income countries where prenatal and postnatal CHD detection rates are low.19 The purpose of this study is to determine the value of pulse oximetry for detecting major CHDs in babies born in a community hospital of north India.

Methods

Patients and setting

The study was a cross-sectional observational survey conducted at the Safdarjung Hospital in New Delhi, India, over a period of 3 years (May 2009–April 2012). Approximately 20 000 deliveries occur in this hospital annually, but for logistic reasons, babies born over a period of 8 h every day were included. Newborns were screened for CHDs by a study team from the All India Institute of Medical Sciences (AIIMS), which is a tertiary referral centre located close to the Safdarjung Hospital. The team comprised of two paediatric cardiologists, a research officer, a field investigator and a social worker.

Clinical examination

Routine clinical examination was performed by the research officer within 24 h of birth. It was recorded in a pro forma that included the following clinical parameters: central cyanosis, cardiac murmur and respiratory distress.

Pulse oximetry

Non-invasive arterial oxygen saturation was obtained for all newborns within 48 h of life. The test was carried out by the field investigator (who was supervised in the initial 3 months of the study period) using the Mindray PM-60 handheld pulse oximeter. The oximeter is US Food and Drug Administration approved and calibrated for functional saturation only. Oximetry values were obtained from one of the feet of the baby. Oxygen saturation less than 95% was considered as abnormal.

Echocardiography

All newborns were screened using bedside echocardiography after the clinical examination and pulse oximetry screening, but no later than 48 h of birth. Bedside echocardiography was performed by a paediatric cardiologist using the FUJIFILM SonoSite M-Turbo ultrasound system. The cardiologist was unaware of the clinical findings or the results of oximetry. All loops and images were stored in digital format for later review. Newborns with CHDs were referred to AIIMS for further management or intervention if needed. In case of any difference in diagnosis after review, the test was repeated at AIIMS by the second paediatric cardiologist using the Philips iE 33 echocardiography equipment.

Definitions of echocardiographic findings

Normal

No echocardiographic abnormality or any of the following: patent ductus arteriosus <2 mm in size without volume overload of left ventricle, interatrial communication (patent foramen ovale or atrial septal defect) <5 mm without volume overload of right ventricle and mild turbulence at branch pulmonary arteries.

Non-significant CHD

Very small muscular ventricular septal defects, which are likely to close spontaneously.

Minor CHD

Atrial septal defect >5 mm with right ventricle volume overload, patent ductus arteriosus >2 mm with left ventricle volume overload, restrictive ventricular septal defect, valvular aortic/pulmonary stenosis with gradients <25 mm Hg.

Major CHD

These were subdivided into two subgroups:

Serious CHD

Any CHD that was likely to require intervention within the first year of life, but was not defined as critical.

Critical CHD

Any CHD that was likely to require an intervention within the first 28 days of life.

Informed consent

An informed consent sheet with details of the study protocol was provided to the parents, and newborns were recruited into the study after one of the parents signed this sheet. None of the parents refused to provide consent for the study.

Statistical analysis

The sample size was calculated by using the estimates from a systematic review of pulse oximetry.21 For an assumed sensitivity of 60% and specificity of 99%, a sample size of 160 true positives provided 80% power to prove that the sensitivity was at least 49% with 5% level of significance. Based on a prevalence of CHDs of 8 per 1000, the number of newborns to be screened to get 160 true positives was calculated to be 20 000.

We calculated the sensitivity, specificity, positive and negative predictive values, positive and negative likelihood ratios and diagnostic accuracy of pulse oximetry and clinical examination and their combination for detecting major CHDs. Ninety-five per cent CIs were calculated using the Wilson method. The differences between sensitivities of pulse oximetry, clinical examination and their combination were compared using the Z-test. The false-positive rates of pulse oximetry done before and after 24 h of birth for critical CHDs were also compared using the Z-test. Statistical significance was defined as p value <0.05.

Results

A total of 20 307 newborns were recruited during the study period. Among these, complete data were available for 19 009 babies (94%); of which, 10 183 were males, and 8826 were females. Only newborns with complete data were included in the analysis. The birth weight ranged from 0.85 to 4.75 kg (mean 2.74±0.46 kg). The study profile is shown in figure 1.

Figure 1

Study profile. CHD, congenital heart defect.

A CHD was diagnosed by echocardiography in 159 newborns, giving a prevalence rate of 8.36 per 1000 live births (95% CI 7.17% to 9.76%). Eighty-nine had a minor CHD, and 70 had a major CHD (major CHDs prevalence rate, 3.68 per 1000 live births; 95% CI 2.92% to 4.65%). Among the 70 with major defects, 44 had serious CHDs, and 26 had critical CHDs.

Among the 44 newborns with serious CHDs, 33 had large ventricular septal defects, three had tetralogy of Fallot, two had non-duct-dependent coarctation of aorta, two had persistent truncus arteriosus, and there was one case each of severe aortic stenosis, congenitally corrected transposition of great arteries with ventricular septal defect and pulmonary stenosis, single ventricle with pulmonary hypertension and atrioventricular septal defect with pulmonary stenosis. Twenty-six newborns had critical CHDs. Their details are given in table 1.

Table 1

Details of newborns with critical congenital heart defects (n=26)

Accuracy of pulse oximetry and clinical examination for detecting CHDs

The sensitivity of pulse oximetry for detecting any CHD (minor and major) was 47.2% (95% CI 39.6% to 54.9%). The sensitivity of clinical examination for detecting any CHD was 25.2% (95% CI 19.1% to 32.4%), the difference was statistically significant (p<0.01). Combining pulse oximetry with clinical examination improved the sensitivity to 65.4% (95% CI 57.7% to 72.4%). This was significantly higher than both, using pulse oximetry alone (p<0.01) or using clinical examination alone (p<0.01). The specificities for detection of any CHD were 68.3% (95% CI 67.7% to 69%) for oximetry, 97.3% (95% CI 97.1% to 97.6%) for clinical examination and 66.7% (95% CI 66% to 67.4%) for the combination.

For major CHDs (table 2), pulse oximetry detected 39 patients (55.7%), and clinical examination detected 25 patients (35.7%). The difference in sensitivity was statistically significant (p=0.02). The combination of pulse oximetry and clinical examination detected 53 patients (75.7%); hence, adding clinical examination to pulse oximetry significantly increased the sensitivity (p=0.01). But the combination still failed to diagnose 17 cases (24.3%).

Table 2

Accuracy of pulse oximetry, clinical examination and their combination for detecting major congenital heart defects (n=70)

For critical CHDs (table 3), pulse oximetry detected 22 patients (84.6%). Clinical examination detected only three cases (11.5%), and again pulse oximetry had a significantly higher sensitivity (p<0.01). The combination of pulse oximetry and clinical examination also detected 22 patients (84.6%), which could have been detected by using pulse oximetry alone. Thus, in this case, adding clinical examination to pulse oximetry showed no increase in sensitivity (p=1.000).

Table 3

Accuracy of pulse oximetry, clinical examination and their combination for detecting critical congenital heart defects (n=26)

Overall, adding pulse oximetry to the standard practice of using clinical examination alone for screening newborns had a significant increase in sensitivity for both major (p<0.01) and critical CHDs (p<0.01). Also, the sensitivity of pulse oximetry increased and that of clinical examination decreased as the severity of CHDs progressed (tables 2 and 3).

Effect of timing of pulse oximetry on detection of critical CHDs

The timing of pulse oximetry test did not have a significant impact on the sensitivity of detecting critical CHDs (table 4). A total of 4261 newborns in the cohort had testing done before the age of 24 h. Among these, pulse oximetry detected four of five critical CHDs (80%). For the remaining 14 748 newborns that were tested after the age of 24 h, pulse oximetry detected 18 of 21 critical CHDs (85.7%).The false-positive rate between the groups was also not significantly different.

Table 4

Accuracy of pulse oximetry for detecting critical congenital heart defects (CHDs) by timing of test

Discussion

The prevalence of CHDs in our study was 8.36 per 1000. This is slightly higher than the reported worldwide prevalence, but is consistent with the high prevalence of CHDs reported from Asia.4 The impact of antenatal detection of CHDs on the prevalence could not be calculated as the rate of fetal echocardiography in our region is very low.

Echocardiography has been the gold standard for diagnosing CHDs, and can be performed accurately by neonatologists.22 Its use as a universal screening tool in low-income and middle-income countries is not feasible because of economic constraints, low proportion of institutional deliveries and lack of adequately trained specialists.23–25 Given the higher burden of CHDs in low-income and middle-income countries due to high birth rates, there is a pressing need of developing an effective screening regimen.26

The sensitivity of pulse oximetry for detecting any CHD in our study was low (47.2%). This may be because of the fact that most newborns with CHDs had non-life-threatening acyanotic disorders like ventricular septal defects and minor CHDs like atrial septal defects and patent ductus arteriosus, which are not usually associated with hypoxaemia. Thus, their oximetry result was not abnormal as has been defined in previous studies.16 Also, human and technical errors could have contributed to the low sensitivity.27 What is interesting is the increase in sensitivity as the severity of CHDs progressed. This increase highlights the potential utility of oximetry in screening newborns for major CHDs.

Clinical examination had a significantly lower sensitivity (25%) than pulse oximetry for detecting any CHD. This is low when compared with the best potential (sensitivity above 90%) of using clinical examination alone as a tool for detecting CHDs.28 Detection of a clinical examination abnormality has been suggested as a predictor of CHD, warranting a prompt echocardiographic evaluation.23 In our study, the sensitivity of clinical examination declined with the severity of defects. This highlights the limitation of using clinical examination as a sole screening method for detecting critical CHDs.

In our study, if pulse oximetry had not been used along with clinical examination, 28 of 70 patients with major and 19 of 26 patients with critical CHD would have been discharged from the hospital undiagnosed. These figures would be higher in the real world as the balance between examination and oximetry is affected by the level of training of the examining doctor. In our study, the research officer was trained in paediatric cardiology, and was available for the entire duration of the study, while other doctors might not be well trained because of limited exposure to paediatric cardiology during their training.19 The consequences of missing minor defects are important, but not as potentially devastating as missing life-threatening critical CHDs.

The timing of the pulse oximetry test had a minimal impact on the sensitivity and false-positive rate for detecting critical CHDs. It has been shown that pulse oximetry during the first 24 h of life detects more false-positive cases.20 This may be due to the transition from fetal to neonatal circulation.29 However, in our study, there was no significant difference between the false-positive rate of pulse oximetry for detecting critical CHDs in newborns screened before and after 24 h of life.

The major strength of our study is the large number of newborns screened. This study is the largest Indian investigation of the use of pulse oximetry for detection of CHDs. Also, in our study, only postductal saturation was taken into account for the sake of simplicity, which has recently been suggested as a good screening method.30

Several limitations of our study merit attention. First, the specificity of pulse oximetry was low as compared with clinical examination and the specificity of oximetry observed in recent similar studies.15–19 This very high false-positive rate can be attributed to a combination of factors. Newborns with oxygen saturation less than 95% did not have the test repeated. Also, there is a high prevalence of infections and respiratory issues in our newborns as most of the deliveries are unbooked pregnancies of women belonging to the lower socioeconomic status who present to the hospital for the first time during labour. The use of a non-motion tolerant oximeter could have also contributed to the increased false-positive rate. However, the exact reasons for false-positive saturation results were not recorded in our study. Second, consecutive newborns were not included in the study due to logistic reasons. Third, the exact timing of oximetry testing was not noted, and the cohort was divided into newborns tested before and after 24 h of birth. Finally, follow-up echocardiography was done only for newborns that were diagnosed with a CHD at birth.

Conclusions

The increasing availability of treatment opportunities for newborns with major CHDs makes early detection crucial to decrease mortality and long-term morbidities. The results of our study indicate that pulse oximetry is a sensitive tool for detecting major CHDs in Indian newborns. Adding pulse oximetry to clinical examination significantly increases the detection of newborns with major CHDs before they are discharged from the hospital. The major limitation is the high false-positive rate of oximetry as compared with clinical examination. Further studies need to be conducted to implement an efficient and cost-effective screening system for detecting CHDs in newborns from low-income and middle-income countries.

References

View Abstract

Footnotes

  • Contributors AS: guarantor; conceptualised and designed the study, obtained funding for the project, performed echocardiograms of newborns, reviewed and critically revised the manuscript and approved the final manuscript as submitted. AM: data analysis and interpretation; drafted the manuscript and approved the final manuscript as submitted. SR: contributed to design of the study, performed echocardiograms of newborns and approved the final manuscript as submitted. MS, SS: contributed to design of the study, supervised the study at primary hospital and approved the final manuscript as submitted. MK: contributed to design of the study, data analysis and approved the final manuscript as submitted. RJ: contributed to data analysis and approved the final manuscript as submitted.

  • Funding The study was funded by a restricted grant from the Indian Council of Medical Research, New Delhi, India. The grant ID was 5/4/1-15/08-NCD-II, and it was used only for conducting the study.

  • Competing interests None declared.

  • Ethics approval Ethics committees of Safdarjung Hospital and All India Institute of Medical Sciences.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and Royal College of Paediatrics and Child Health