Diagnosis Scenario

An infant is born at 30 weeks gestation, weighing 1560 grams. She requires endotracheal intubation and mechanical ventilation shortly after birth for respiratory distress syndrome. Ventilator pressure and the fraction of inspired oxygen are weaned nicely over the first three days following treatment with exogenous surfactant. However, on day four, the infant’s need for ventilator and oxygen support increases. You suspect a patent ductus arteriosus (PDA) with left to right shunting. Treatment with indomethacin will be started if a PDA is confirmed. The senior resident wants to order an echocardiogram. You aren’t sure that this is necessary and you wonder if the clinical exam is sufficiently accurate to diagnose a PDA.

You pose the clinical question, “In a preterm infant with a birth weight of 1560 grams, what is the accuracy of the clinical exam as a diagnostic test for PDA with left to right shunting?”

You perform a PubMed search using the MeSH terms “patent ductus arteriosus”, “infant, premature”, and “sensitivity and specificity”. Of the 7 articles identified, only 2 compare clinical signs with echocardiography. You review both abstracts in PubMed and decide to copy the original article for the larger and more comprehensive study Arch Pediatr Adolesc Med 1995; 149: 1136-1141.

Read the article and decide:

Is the evidence from this randomised trial valid?
If valid, is this evidence important?
If valid and important, can you apply this evidence in caring for your patient?

Completed Diagnosis Worksheet for Evidence-Based Neonatal Medicine

Citation

Davis P., Turner-Gomes S., Cunningham K., et al. Precision and Accuracy of Clinical and Radiological Signs in Premature Infants at Risk of Patent Ductus Arteriosus. Arch Pediatr Adolesc Med 1995; 149: 1136-1141.

Are the results of this diagnostic study valid?

Was there an independent, blind comparison with a reference (“gold”) standard of diagnosis?: Yes. Comparison to Doppler flow echocardiogram.
Was the diagnostic test evaluated in an appropriate spectrum of patients (like those in whom it would be used in practice)?: Yes. Infants had birth weights less than 1750 grams, and were between 3 and 7 days old. A third of the cohort was intubated and mechanically ventilated at the time of study.
Was the reference standard applied regardless of the diagnostic test result?: Yes

Are the valid results of this diagnostic study important?

As infants in this study were examined independently by at least three and up to five assessors, a single 2×2 table cannot be generated based on published averages. Below are the mean sensitivities, specificities, and positive predictive values. Likelihood ratios are reported for the presence of all three clinical signs (increased pulse volume, active precordium, and cardiac murmur), and for the absence of any clinical sign.

Clinical Sign	Sensitivity	Specificity	+ Likelihood Ratio
Increased pulse volume	43%	74%	1.6
Active Precordium	26%	85%	1.7
Cardiac Murmur	42%	87%	3.0

$$ text{Positive Predictive Value}\
= 22% (text{pulse volume}); 36% (mathit{precordium}); 51% (mathit{murmur}). $$

$$ text{Pre-test Probability ($prevalence$) of PDA}\
= 23% $$

$$ mathit{Pre-test-odds} = mathit{prevalence}/(1-mathit{prevalence})\
= 0.299 $$

$$ text{Likelihood ratio for the presence of PDA with all three clinical signs}\
= 3.7 $$

$$ text{Post-test odd}\
= text{Pre-test odds} times text{Likelihood Ratio}\
= 1.106 $$

$$ text{Post-test Probability of PDA with all three clinical signs}\
= text{Post-test odds}/(text{Post-test odds} + 1)\
= 53% $$

$$ text{Likelihood ratio for the presence of PDA without any clinical sign}\
= 0.6 $$

$$ text{Post-test odds} \
= text{Pre-test odds} times text{Likelihood Ratio} \
= 0.179 $$

$$ text{Post-test Probability of PDA without any clinical sign}\
= text{Post-test odds}/(text{Post-test odds} + 1)\
= 15% $$

Can you apply this valid, important evidence about a diagnostic test in caring for your patient?

Do these results apply to your patient?

Is the diagnostic test available, affordable, accurate, and precise in your setting?: Clinical exam is readily available but its accuracy and precision are poor.
Can you generate a clinically sensible estimate of your patient’s pre-test probability (from practice data, from personal experience, from the report itself, or from clinical speculation): In the US National Collaborative Study, a PDA was detected in 20% of all infants who weighed 1750 grams or less at birth. Within the study cohort, the prevalence of PDA increased with decreasing birth weights (Pediatrics 1983; 71:364-72).
Will the resulting post-test probabilities affect your management and help your patient? (Could it move you across a test-treatment threshold?; Would your patient be a willing partner in carrying it out?): No. The absence of any clinical signs does not rule out PDA; the presence of all three clinical signs results in a post-test probability of only 53%, too low to justify the use of therapeutic indomethacin with its potentially adverse effects.
Would the consequences of the test help your patient?: No.

Additional Notes

The investigators also examined the interobserver variability in identifying the clinical signs. In general, the precision was poor (weighted kappas 0.15 to 0.41) indicating great variability in the ability of clinicians to identify the clinical signs.
And what about that second paper which was retrieved from PubMed: Its authors come to the same conclusion: “Echocardiography is required for the reliable early diagnosis of a PDA in ventilated preterm infants” J Paediatr Child Health 1994; 30:406-11

Patent Ductus Arteriosus: Clinical exam is inaccurate and imprecise in diagnosing PDA with left to right shunting in premature infants

Clinical Bottom Line

In preterm infants with birth weights 750 to 1750 grams, the clinical examination on days 3 to 7 of life is inadequate to diagnose a PDA with left to right shunting.

Citation

Clinical Question

In preterm infants, what is the accuracy of the clinical examination as a diagnostic test for patent ductus arteriosus (PDA) with left to right shunting?

Search Terms

PubMed search using MeSH terms: “patent ductus arteriosus”, “infant, premature” and “sensitivity and specificity”.

The Study

The Study Patients: Infants with birth weights between 750 and 1750 gm and who were between 3 and 7 days old. Infants were enrolled from one Canadian neonatal intensive care unit.
Independent blind: reference standard applied regardless of test result and performed in an appropriate spectrum of patients
Target disorder and Gold Standard: Patent ductus arteriosus with left to right shunting. Gold standard was echocardiography.
Diagnostic test: Each infant was assessed independently by at least three and up to five assessors for the pulse quality, precordial activity, and cardiac murmur.

The Evidence

Clinical Sign	Sensitivity	Specificity	+ Likelihood Ratio
Increased pulse volume	43%	74%	1.6
Active Precordium	26%	85%	1.7
Cardiac Murmur	42%	87%	3.0

$$ text{Positive Predictive Value}\
= 22% (text{pulse volume}); 36% (mathit{precordium}); 51% (mathit{murmur}).
$$

$$ text{Pre-test Probability ($prevalence$) of PDA}\
= 23%
$$

$$ mathit{Pre-test-odds} = mathit{prevalence}/(1-mathit{prevalence})
= 0.299
$$

$$ text{Likelihood ratio for the presence of PDA with all three clinical signs}\
= 3.7
$$

$$ text{Post-test odds}\
= text{Pre-test odds} times text{Likelihood Ratio}\
= 1.106
$$

$$ text{Post-test Probability of PDA with all three clinical signs}\
= text{Post-test odds}/(text{Post-test odds} + 1)\
= 53%
$$

$$ text{Likelihood ratio for the presence of PDA without any clinical sign}\
= 0.6
$$

$$ text{Post-test odds}\
= text{Pre-test odds} times text{Likelihood Ratio}\
= 0.179
$$

$$ text{Post-test Probability of PDA without any clinical sign}\
= text{Post-test odds}/(text{Post-test odds} + 1)\
= 15%
$$

Comments

Interobserver variability of the clinical signs was examined and found to be poor (weighted kappas 0.15 to 0.41)
2×2 table cannot be generated based on the published results which were averaged amongst the different assessors.
See also J Paediatr Child Health 1994; 30: 406-11 which reached similar conclusions.

Appraised By

Aaron Chiu