Evid Based Med 18:e26 doi:10.1136/eb-2012-100971
  • Diagnosis
  • Systematic review and meta-analysis

Confirming the value of pulse oximetry screening for diagnosing critical congenital heart disease

  1. Julien I E Hoffman
  1. Department of Pediatrics, University of California, San Francisco, California, USA
  1. Correspondence to : Dr Julien I E Hoffman
    925 Tiburon Boulevard, Tiburon, CA 94920-1525, USA; julien.hoffman{at}

Commentary on: [CrossRef][Medline][Web of Science]Google Scholar


Thangaratinam and colleagues, studied the value of pulse oximetry screening of neonates for critical congenital heart disease (CCHD). CCHD is usually defined as a congenital heart disease (CHD) that requires treatment in the neonatal period to prevent death or severe morbidity; most of these neonates have ductus-dependent lesions. About 25–40% of children born with CHD have CCHD, and 10–35% of these infants leave hospital undiagnosed. Some die at home, and others return to hospital critically ill, with resulting higher operative mortality and morbidity, increased hospital costs and suboptimal outcomes.

Pulse oximetry has been recommended to decrease these missed diagnoses, but the sensitivity and specificity of this screening method have not been thoroughly evaluated. Two large studies in Sweden1 and Norway2 ,3 indicated the value of pulse oximetry screening, but prior smaller studies sometimes supported and at other times refuted the use of this screening method.


This was a methodologically well-done systematic review and meta-analysis. The authors screened 552 reported studies, and identified 13 studies involving 229 421 neonates that met strict criteria for populations, tests and outcomes. Sophisticated statistical analyses were performed. Pulse oximetry was performed <24 h of age in six studies. Eight studies measured oxygen saturation only in the foot (postductal). Nine studies excluded infants suspected antenatally of having congenital heart disease. Positive results were verified by echocardiography, and negative results by clinical follow-up or registries of congenital anomalies or mortality.


Senisitivity was 76.5% (95% CI 67.7% to 83.5%). Specificity was 99.9% (95% CI 99.7% to 99.9%) with a false-positive rate of 0.14% (95% CI 0.06% to 0.33%). The positive likelihood ratio was 549, and the negative likelihood ratio 0.24. False-positive rates were lower (0.05%, 95% CI 0.02% to 0.12%) for screens performed > 24 h after birth (compared to 0.50%, 95% CI 0.29% to 0.86%), but were not affected by oximetry measurement site (foot alone vs foot and hand); however, in some individuals screening only on the foot would have led to misdiagnosis.


The relatively low sensitivity is not important, because nobody expects pulse oximetry to detect all CCHD. What is required of sensitivity is that it is better than the alternative tests, that is, antenatal ultrasound and standard clinical examination, and this is indeed true. In regard to specificity, the test is quite good. One extensive review of this subject found that the false-positive rate was sixfold lower for pulse oximetry screening than for clinical examination.4 The figures for false-positive rates are particularly useful because it was concern for the potentially high costs of performing echocardiograms on all positive results (particularly false positives) that delayed instituting universal pulse oximetry screening of neonates. Furthermore, a false-positive rate of 0.6% would in a hospital with 2000 births annually lead to one unnecessary echocardiogram for a false-positive screen each month,5 so that with a false-positive rate of 0.05% there would be one unnecessary echocardiogram per year. This is a very small price to pay for an improved method of diagnosing CCHD.

The authors neither addressed cut-off oxygen saturations, nor the value of a repeat screen, if the first was borderline. Nine of the 13 studies used 95% as a cut-off value, two studies used 94% and one each used 92% and 96%. Eight studies repeated oximetry after 2–12 h for saturations from 90% to 95%, eliminating many false-positive results due to transient pulmonary hypertension or delayed clearance of lung fluid. A difference between hand and foot probes was regarded as abnormal if >2% or 3%, except for one study that used 7%.

Another main contribution of this study has been to show that the low false-positive rate is not confined to one or two institutions, and is low enough to allow this method to be used as a universal screening test. Their conclusions add support to the results of the large Scandinavian studies of this subject, and suggest that there is no excuse for withholding a valuable way of reducing the burden of serious CCHD.


  • Competing interests None.


Free Sample

This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of EBM.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.