Late-presenting left-to-right shunts and operability scoring: Finer nuances

Dear Sir

I read with interest the article by Kadiyani et al.[1] in the recent issue of this journal, an examination of a large dataset assessing operability in late-presenting shunt lesions. I commend the authors for bringing out tangible data on a subject that continues to challenge even the most rigorous clinical decision-making. I wish to highlight certain aspects of the study design, statistical analysis, and interpretation that may be relevant for clinicians seeking to apply these findings in practice.

Although shunt lesions are often grouped together given the similar principles underlying hemodynamic data interpretation, the differing pathophysiology and natural history of pre-and posttricuspid shunt lesions are well reflected in this dataset. The authors note a statistically significant difference in mean age between the two groups (32 years vs. 9 years, P < 0.01). While a common operability cutoff of >2 years is reasonable, the richness of this database might have been better utilized by analyzing the two pathophysiological groups separately and by treating age as a continuous rather than categorical variable in correlating operable/inoperable and good/poor prognosis outcomes.[2]

The results report 86% of positive vasoreactivity responders by the Barst criteria, yet only 74 (47%) cases were “deemed operable” by criteria that are not further clarified in the manuscript. It is also noted that the score predicts lack of vasoreactivity, but data correlating score performance with actual postoperative outcomes are absent. In this light, “operability score” may be a somewhat misleading term; “acute vasoreactivity predictor score” might be more precise. Sharing the score distribution across the full study population – and its agreement with precatheterization clinical assessment of likely operable, borderline, or likely inoperable cases – would add considerable value for the reader. A comparison of scores between patients with good versus adverse postsurgical outcomes would further strengthen its clinical utility.[3] This data may also support a three-tier risk stratification (low, intermediate, and high) rather than a binary classification.

In late-presenting shunt lesions, vasoreactivity testing serves as a crucial screening step before proceeding with or declining surgery. Since a favorable vasoreactivity result carries significant prognostic implications for the individual patient, the primary goal of any bedside screening tool should be to avoid missing potentially operable patients. An ideal predictive score for “nonvasoreactivity” or “inoperability” should therefore prioritize near-100% specificity and high positive predictive value, minimizing false negatives – that is, operable patients incorrectly classified as nonvasoreactors. However, the authors state that their score was designed to “balance sensitivity and specificity,” yielding a specificity of 75% for predicting vasoreactivity. One must ask whether a score with this level of specificity risks denying a meaningful proportion of patients their last opportunity for surgical correction.

A scoring tool for operability is most valuable when it improves upon standard clinical judgment. While this score quantifies clinical parameters that have conventionally informed operability decisions,[4] a system that misclassifies one in four operable patients warrants careful scrutiny. Notably, in this very dataset, clinical judgment achieved a higher specificity for predicting inoperability – only 7 of 45 “likely inoperable” patients were found operable on catheterization data, yielding a specificity of 85%.

A deeper engagement with these aspects would help readers better contextualize the findings and further strengthen the study’s contribution to the wider pediatric cardiac community.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.