|
Indian Pediatr 2014;51: 913-917 |
|
How Useful is Pulse Oximetry for Screening of
Congenital Heart Disease in Newborns?
|
Source Citation:
Zhao QM, Ma XJ, Ge XL, Liu F, Yan WL, Wu L, et al. Pulse oximetry with
clinical assessment to screen for congenital heart disease in neonates
in China: A prospective study. Lancet. 2014;384:747-54.
Section Editor: Abhijeet Saha
|
Summary
In the initial pilot study at three hospitals in
Shanghai, the authors assessed the accuracy of pulse oximetry plus
clinical assessment for detection of congenital heart disease. They then
undertook a large, prospective, and multicenter screening study in all
consecutive neonates (aged 6–72 h) born at 18 hospitals in China between
August 1, 2011, and November 30, 2012. Newborns with positive screen
results (either an abnormal pulse oximetry or abnormal clinical
assessment) were referred for echocardiography within 24 h of screening.
Sensitivity, specificity, positive and negative predictive values, and
positive and negative likelihood ratios for pulse oximetry alone, and in
combination with clinical assessment, for detection of major and
critical congenital heart disease were calculated.
In the pilot study, 6785 consecutive newborns were
screened; 46 of 49 cases of asymptomatic major congenital heart disease
and eight of eight cases of asymptomatic critical disease were detected
by pulse oximetry and clinical assessment. In the prospective
multicenter study, they screened 122 738 consecutive new born babies
(120 707 asymptomatic and 2031 symptomatic), and detected congenital
heart disease in 1071 (157 critical and 330 major). In asymptomatic
newborns, the sensitivity of pulse oximetry plus clinical assessment was
93·2% (95% CI 87·9–96·2) for critical congenital heart disease and 90·2%
(86·4–93·0) for major disease. The addition of pulse oximetry to
clinical assessment improved sensitivity for detection of critical
congenital heart disease from 77·4% (95% CI 70·0–83·4) to 93·2%
(87·9–96·2). The false-positive rate for detection of critical disease
was 2·7% (3298 of 120 392) for clinical assessment alone and 0·3% (394
of 120 561) for pulse oximetry alone.
The authors concluded that pulse oximetry plus
clinical assessment is feasible and reliable for the detection of major
congenital heart disease in newborns in China. They recommended this
combined method to be used in maternity hospitals to screen for
congenital heart disease.
Commentaries
Evidence-based-medicine Viewpoint
Relevance: Early diagnosis of congenital
heart disease (CHD) is important for appropriate management of
potentially critical (yet often treatable) conditions. An ideal
diagnostic test is expected to be accurate, reliable, reproducible, and
applicable at the point-of-care (place of birth). Echocardiography fits
the bill, but is expensive, requires considerable expertise, and is not
easily available at most centers. Meticulous clinical examination alone
can miss cases, and is highly dependent on the observer’s training,
skill and experience. Further, most delivery centers do not apply a
standardized protocol of newborn examination, resulting in missing
congenital heart disease. Against this backdrop, the study by Zhao,
et al. [1] appears to be of great relevance.
Critical appraisal: The study was a large
prospective, multi-center investigation, screening newborn infants for
CHD using clinical examination, pulse oximetry, and both (diagnostic
tests). Echocardiography was performed in all asymptomatic infants who
had an abnormal screen result (defined as any of the three screening
tests being abnormal), as well as a smaller cohort of symptomatic
infants. The investigators undertook a pilot study in three Shanghai
hospitals representative of Chinese delivery centers, and enrolled over
6700 infants to confirm the feasibility of the study procedures.
Thereafter all consecutively born infants (>122,000) in 18 hospitals
across the country, underwent the screening tests (between 6-72 h of
life). Standard definitions were used to define a positive screen. The
authors analyzed asymptomatic and symptomatic infants separately.
Asymptomatic babies with a negative screen were followed clinically at 6
weeks and by parental feedback, to determine if a diagnosis of
congenital heart disease emerged. However all symptomatic infants
underwent echocardiography. CHD was categorized into four groups viz
critical, serious, significant and non-significant; the first two groups
were combined as major CHD and the latter two as minor CHD. The outcome
of interest was critical and major CHD. The investigators reported that
the screening tests were highly sensitive and specific for detecting the
outcomes of interest. The combination of clinical examination plus pulse
oximetry performed better than either alone.
The study provides an opportunity to critically
appraise a well-designed investigation of diagnostic test accuracy. Most
tools available for the purpose [2-7] examine reports as shown in
Table I. The sample size was not calculated a priori, but
post hoc analysis suggested that the pilot study was adequately
powered. Appropriate statistical tests were used in the study and data
were presented using all parameters for diagnostic tests.
Table I Critical Appraisal of the Study
Validity |
|
Are the
results of the study valid? |
The
investigators applied the diagnostic test (clinical examination,
pulse oximetry, and the combination) in a large cohort of
consecutively born infants in three hospitals. Only a few babies
who were antenatally diagnosed were excluded from analysis. Thus
there is very low risk of selection bias in this study.
|
Was the
reference standard applied regardless |
All newborn
infants underwent the three screening tests. In the pilot study
of the index test result (n=6785) all infants underwent
echocardiography as the reference standard, which was performed
regardless of the screening test results. However in the main
study, echocardiography was performed only in those with an
abnormal result and in symptomatic babies. Further, this is a
passive system of detection based on parents bringing sick
infants. Obviously, there is a danger that several infants
(including all that died) would be missed by this method. This
may explain why the investigators detected 1071 CHD among 122738
enrolled infants (0.87%) whereas the expected baseline
prevalence was estimated to be 1.25-1.62%. |
Was there an
independent, blind comparison between the index test and an
appropriate reference (‘gold’) standard of diagnosis?
|
The reference
test (echocardiography) used in this study is the current ‘gold
standard’ believed to be as close to the ‘truth’ as possible.
However, in the main study, it was performed only in 2031
symptomatic infants and asymptomatic infants with a positive
screening test (exact number unclear). Diagnosis of CHD in
asymptomatic infants with a negative screen was sought by
clinical follow-up and parental feedback to the reference
standard. |
Results |
Test
characteristics and measures |
The authors
presented the results for clinical assessment alone, pulse
oximetry alone, and the combination. The specific outcomes of
interest were critical CHD (fatal or requiring correction before
28d) and major CHD (critical CHD plus cases requiring
intervention before 1y of age). Results for the combination are
summarized below. |
|
Critical CHD:
Sn=93·2% (87·9–96·2); Sp=97·1% (97·1–97·2); PPV=3·8% (3·2–4·5);
NPV= 99·99% (99·98–100); LR+ = 32·6% (32·5–32·6); LR- = 0·07%
(0·06–0·09) |
|
Major CHD: Sn=90·2%
(86·4–93·0); Sp=97·3% (97·2–97·4); PPV=7·9% (7·1–8·9);
NPV=99·97% (99·96–99·98); LR+ = 32·9% (32·9–33·0); LR- = 0·10%
(0·10–0·11). |
|
The data
suggest that the combination was fairly accurate, and could
predict the absence of critical or major CHD reliably. The
likelihood ratios suggest that the test could be very useful in
that setting. For all parameters (except PPV), the combination
performed better than either clinical examination or pulse
oximetry alone. |
Applicability |
Do the methods described permit replication? |
The investigators have provided detailed descriptions of the
clinical examination protocol which included four indicators
(family history of CHD, special facial characteristics, cardiac
murmur after the first day, and non-cardiac malformations).
Likewise, the method and device for pulse oximetry measurement
are also well described. Overall, the diagnostic test appears
applicable in diverse settings. Successful pilot testing in a
few thousand infants before initiating the study confirmed the
replicability of the procedures and results. However, the
diagnostic test was applied at 6-72h of age; whereas cardiac
murmur was considered only after 24 h; this apparent discrepancy
is not elaborated by the authors. There is no description of the
accuracy, reliability and reproducibility of the pulse oximeters
used in this study. |
LR- = Likelihood ratio of a positive test, LR- = Likelihood
ratio of a negative test, NPV=Negative predictive value. Sn=Sensitivity,
SP=Specificity, PPV=Positive predictive value. |
Extendibility: At first glance, the
diagnostic test described in this study appears simple to use at various
levels of clinical care, as in the Indian health-care delivery system.
However, it should be noted that the results presented were obtained in
a highly controlled research setting, wherein specially trained
observers and meticulous procedures were involved. It is debatable
whether such promising results would be obtained in the operational
setting where deliveries are not always conducted by physicians, level
of training of personnel is variable, accuracy of oximetry devices is
unconfirmed, and supervision/monitoring may be limited. The other
important issue is what could/should be done in the event of an abnormal
test result. Most maternity facilities have no access to confirmatory
echocardiography, and positive test results would obviously translate to
referral. This could create unwarranted anxiety and inconvenience to
families, as well as an over-burden on the limited centers with
facilities for neonatal echocardiography.
Conclusions: This study suggests that meticulous
clinical examination of newborn infants supplemented with pulse oximetry
could be a useful diagnostic test to detect critical and major
congenital heart disease in diverse clinical settings. The methods can
be extended to the Indian context, although the results may be variable.
Joseph L Mathew
Department of Pediatrics,
PGIMER, Chandigarh, India.
Email:
[email protected]
Neonatologist’s Viewpoint
Among all birth defects, CHD is single most common
and important cause of infant mortality [8]. Critical congenital heart
diseases consist of a subgroup which needs surgery or catheter
intervention in the neonatal period [9]. Timing of manifestation of
critical CHD is dependent on fall in pulmonary vascular resistance and
closure of ductus arteriosus. Neonates with critical CHDs may not
manifest clinical signs like murmur, tachypnea or cyanosis in first
48-72 h of birth and therefore discharged undiagnosed. Delay in
detection and presentation with circulatory collapse or severe cyanosis
is associated with worse outcome. In pre-symptomatic period, pulse
oximetry can detect subclinical hypoxemia resulting from decreased
pulmonary blood flow or intracardiac shunting. With high-quality
evidence from high-resource settings demonstrating efficacy of pulse
oximetry in early detection of critical CHD, many developed countries
have implemented a universal pre-discharge screening program [10]. In
this study [1] on a large cohort from China, pulse oximetry combined
with clinical assessment was able to detect more than 90% of critical
CHD. Sensitivity of clinical assessment alone was lower than pulse
oximetry. Pulse oximetry was especially useful in detection of cyanotic
CHD like transposition of great vessels and total anomalous venous
connection. However, true benefit of screening strategy lies not in
early and accurate diagnosis, but in preventing morbidity and mortality
due to the target condition. The study does not present data on outcome
of neonates diagnosed with the screening strategy. Early detection may
not translate into survival if post-screening procedures are not in
place.
Universal metabolic screening is a well-established
practice in high-income countries. Despite being home to second largest
population in world and therefore with potentially huge burden of
metabolic disorders, no state in India has been able to introduce and
sustain universal neonatal screening program. Lack of laboratory set up,
trained manpower, non-availability of confirmatory tests and high-cost
of treatment of metabolic disorders have precluded implementation of
universal screening program in India and other low- and middle-income
countries. A screening program to diagnose critical CHD needs to be
assessed in this context. Equipment, training and time needed to screen
and interpret results of pulse oximetry are less resource-intensive than
metabolic screening. CHD are more common than most of metabolic
disorders, making a pulse oximetry screening program potentially more
cost-effective. However, difficulties start once an infant is labeled as
‘suspect’ on the basis of screening program.
1. Availability of confirmatory test:
Suspected patients of CHD need quick confirmation by
echocardiography either at site of birth or referral to a center
where a trained pediatric cardiologist is available. Majority of
births occur in health facilities where pediatric cardiologist is
not available. Even in centers where a cardiologist is available,
the screening program can significantly increase the work-load.
Pulse oximetry has high false-positivity rate resulting in large
number of neonates labeled ‘suspect’, and therefore necessitating
echocardiographic confirmation.
2. Availability of treatment: Early
diagnosis of critical CHD is useless unless immediate palliative or
corrective intervention is done. In most settings of developing
world, Pediatric cardiac intervention set-ups are rarely available.
Treatment needs even greater infrastructure in the form of Pediatric
cardiothoracic surgeon, neonatal anesthesiologist and pediatric
cardiac intensive care unit.
High-quality evidence now exists favoring role of
pulse oximetry screening program in early diagnosis of critical CHD [3].
Nevertheless developing countries first need to bring down infant and
neonatal mortality by implementing measures to save infants from
infections, birth asphyxia and prematurity. Time-consuming process of
creating a pool of trained manpower and establishing infrastructure to
diagnose and manage complex birth defects like CHDs need to be started
meanwhile.
Deepak Chawla
Department of Pediatrics,
Government Medical College Hospital,
Chandigarh,
India
Email: [email protected]
Pediatric Cardiologist’s Viewpoint
CHD is an important public health issue, with an
incidence of 2 to 3 cases of critical CHD per 1000 live births, and has
been shown to be responsible for more than 40% of infant deaths related
to congenital malformations [11,12]. Without a prenatal and postnatal
screening test, even severe forms of CHD commonly go undetected on usual
clinical examination until after discharge to home, leading to avoidable
morbidity and mortality [13]. Prenatal or postnatal detection of major
forms of CHD, may improve preoperative conditions and survival after
surgery. Prenatal diagnosis is the optimum need as it offers several
options, including termination of pregnancy, in utero treatment,
and planning the timing, mode and place of delivery to a better equipped
tertiary care center with facilities for cardiac surgery. However,
current approaches to prenatal screening for CHD remains flawed,
commonly missing more than half of the cases of severe CHD [14-16].
Routine neonatal examination fails to diagnose more
than half of babies with heart disease; examination at 6 weeks misses
one-third [17]. Spending more time on physical examination is
unrewarding as only milder cases of pulmonic stenosis and other
relatively benign forms of CHD are diagnosed, which has little impact on
the morbidity and mortality from undiagnosed CHD. Pulse oximetry can
pick up lesions producing low oxygen saturation levels consequent to
substantial abnormal mixing of systemic and pulmonary blood streams or
critical obstructive duct-dependent lesions (mostly cyanotic CHD).
Although it may fail to detect acyanotic CHD and critical CHD with
non-critical obstruction or mixing, these lesions do not contribute to
early mortality and morbidity.
This paper expresses the sensitivity of a combination
of both the physical examination and pulse oximetry over pulse oximetry
alone. Improving the diagnosis of CHD by a physical examination is
unrewarding as there is a poor correlation between the ease of diagnosis
and the severity of the cardiac lesion [18]. However, physical
examination for CHD in this study was not only based on presence of
murmurs, cyanosis or congestive heart failure, but on the whole gamut of
history – including family history of CHD – and examination findings,
including syndromic facies and extra cardiac anomalies, thereby
increasing the sensitivity and specify of picking up more of critical
and major CHD. The paper does not report the competency of the
pediatrician, and on the time spent to auscultate the heart.
A recent systematic review [19] of data from 229 421
new born babies reported high specificity and acceptable sensitivity for
detection of critical CHD. The false-positive rate for detection of
critical congenital heart defects was particularly low when new born
pulse oximetry was done after 24 h from birth than when it was done
before 24 h.
Although pulse oximetry is inexpensive and without
side effects, it cannot detect CHD in every neonate with congenital
heart disease, before they leave the hospital. Pulse oximetry is highly
specific for detection of critical congenital heart defects with
moderate sensitivity that meets criteria for universal screening [20].
In a country with a huge population and poor prenatal diagnostic
infrastructure, it is probably the best thing to do and should become a
recommendation. We still need to address several issues including
overall costs of screening, delayed diagnoses because of false-negative
screen results, the costs of evaluation and the iatrogenic anxiety/
fears generated in families of children with false positive screen
results .
Dinesh Kumar
Department of Pediatrics,
PGIMER, Dr RML Hospital,
New Delhi, India.
Email:
[email protected]
References
1. Zhao Q, Ma X, Ge X, Liu F, Yan W, Wu L, et al.
Pulse oximetry with clinical assessment to screen for congenital heart
disease in neonates in China: a prospective study. Lancet. 2014;
384:747-54.
2. Jaesche R, Guyatt GH, Sackett DL. Users’ Guides to
the Medical Literature, V1. How to use an article about a diagnostic
test. JAMA. 1994; 271:389-91.
3. No authors listed. Critical Appraisal Skills
Programme (CASP) Diagnostic Test checklist_14.10.10. Available from:
http://www.caspinternational.org/mod_product/uploads/CASP_Diagnostic_Checklist_14.10.10.pdf.
Accessed October 10, 2014.
4. No authors listed. Exercise: Critical Appraisal of
a Diagnostic Test Study. Available from:
http://www-users.york.ac.uk/~mb55/msc/critappr/bursitis.pdf.
Accessed October 10, 2014.
5. Bossuyt P, Irwig L, Glasziou P. Diagnostic test
appraisal form. Screening and test evaluation program. Available from:
http://sydney.edu.au/medicine/public-health/step/about/appraisal/form.pdf.
Accessed October 10, 2014.
6. No authors listed. Diagnostic Test Studies:
Assessment and Critical Appraisal. Available from:
http://clinicalevidence.bmj.com/x/set/static/ebm/toolbox/665061.html.
Accessed October 10, 2014.
7. No authors listed. Diagnostic Study Appraisal
Worksheet. Available from:
http://www.cebm.net/wp-content/uploads/2014/04/diagnostic-study-appraisal-worksheet.pdf.
Accessed October 10, 2014.
8. Heron MP, Smith BL. Deaths: leading causes for
2003. Natl Vital Stat Rep. 2007;55:1-92.
9. Ewer AK, Middleton LJ, Furmston AT, Bhoyar A,
Daniels JP, Thangaratinam S, et al. Pulse oximetry screening for
congenital heart defects in newborn infants (PulseOx): A test accuracy
study. Lancet. 2011;378:785-94.
10. Mahle WT, Newburger JW, Matherne GP, Smith FC,
Hoke TR, Koppel R, et al. Role of pulse oximetry in examining
newborns for congenital heart disease: A scientific statement from the
American Heart Association and American Academy of Pediatrics.
Circulation. 2009;120:447-58.
11. Hoffman JIE, Kaplan S. The incidence of
congenital heart disease. J Am Coll Cardiol 2002;39:1890-1900.
12. Boneva RS, Botto LD, Moore CA, Yang Q, Correa A,
Erickson JD. Mortality associated with congenital heart defects in the
United States: Trends and racial disparities, 1979–1997. Circulation.
2001;103:2376-81.
13. Kuehl KS, Loffredo CA, Ferenez C. Failure to
diagnose congenital heart disease in infancy. Pediatrics.
1999;103:743-7.
14. Simpson JM. Impact of fetal echocardiography. Ann
Pediatr Cardiol. 2009; 2:41-50.
15. Ogg‘e, G. Gaglioti P, Maccanti S, Faggiano F,
Todros T. Prenatal screening for congenital heart disease with
four-chamber and outflow-tract views: A multicenter study. Ultrasound
Obstet Gynecol. 2006;28:779-84.
16. Sklansky MS, Berman DP, Pruetz JD, Chang RK.
Prenatal screening for major congenital heart disease superiority of
outflow tracts over the 4-chamber view. J Ultrasound Med. 2009;
28:889-99.
17. Wren C, Richmond S, Donaldson L. Presentation of
congenital heart disease in infancy: Implications for routine
examination. Arch Dis Child Fetal Neonatal Ed. 1999;80:F49-53.
18. Reich JD. Clinical screening for congenital heart
disease at birth: a long way to go. Indian Pediatr. 2011;48:17-8.
19. Thangaratinam S, Daniels J, Ewer AK, Zamora J,
Khan KS. Accuracy of pulse oximetry in screening for congenital heart
disease in asymptomatic new-borns: A systematic review. Arch Dis Child
Fetal Neonatal Ed. 2007; 92:F176-80.
20. Cuzzi S, Bradshaw E. The road to universal pulse-oximetry
screening: Are we there yet? Pediatrics. 2011;128:e1271-2.
|
|
|
|