I read with interest the study by Cassady and colleagues comparing healthcare utilization among patients with pulmonary hypertension associated with interstitial lung disease (PH-ILD) who initiated inhaled treprostinil versus matched patients who remained untreated in a large US administrative claims dataset. The authors report fewer all-cause and ICU-related hospitalizations among treated patients, including a 30% decreased risk of hospitalization compared with untreated controls (relative risk 0.70; 95% CI 0.59–0.83; p < 0.01) 1. In routine care, it helps to name what a claims-based estimate can and cannot represent. Patients who ultimately receive inhaled treprostinil are a selected subset. They must survive long enough for PH-ILD to be recognized, reach a clinician comfortable prescribing the therapy, navigate authorization, and remain stable enough to attempt inhaled titration. Selection is visible even in the matched cohorts. After matching, right heart catheterization was far more common in treated patients than in untreated patients (79% vs 16%), which suggests meaningful differences in diagnostic certainty and care pathways between groups 1. Time-zero alignment is a second issue. In this analysis, treated patients are indexed at treprostinil initiation, whereas untreated patients are indexed at the first observed PH diagnosis. Events that occur between diagnosis and treatment start are not counted in the treated group by design. In a condition with high short-term morbidity and competing risks, this can amplify apparent benefit through immortal time bias 2. Sensitivity analyses that treat therapy as a time-varying exposure or apply a landmark design would help bound the magnitude of this effect 2. Generalizability is another hinge, particularly outside referral centers. Many patients I see in community-facing and safety net settings carry overlapping HFpEF, COPD, obesity-related hypoventilation, chronic kidney disease, and recurrent exacerbations that drive admissions independent of pulmonary vascular disease. Claims-based oxygen proxies and diagnostic codes struggle to separate severe precapillary pulmonary vascular disease from mixed physiology. This heterogeneity is central to group 3 pulmonary hypertension and shapes both treatment selection and outcomes 3. Finally, claims data cannot capture determinants of success that clinicians navigate day to day. Tolerability, adherence, inhalation technique, and clinic capacity for education and titration often decide whether therapy is sustained. The INCREASE trial established efficacy in PH-ILD and also highlighted the practical reality of prostacyclin-related adverse effects and dose escalation 4. Real-world evaluation and follow-up frameworks for inhaled treprostinil have been proposed, which reinforces the importance of structured phenotyping and longitudinal reassessment in practice 5. I include a simple one-panel pathway (Figure 1) to reflect how these steps can be operationalized in a clinic workflow. None of these considerations diminishes the value of the authors’ work. They sharpen how I would apply it. I view this analysis as encouraging, hypothesis-generating evidence that inhaled treprostinil may reduce acute-care utilization among patients who can access and persist with therapy. Prospective registries that pair utilization outcomes with physiologic and echocardiographic markers of right ventricular adaptation, reasons for non-initiation and discontinuation, and titration patterns would strengthen the bridge from association to implementation. Teja Narra conceived the letter, drafted the manuscript, approved the final version, and is accountable for all aspects of the work. The author has nothing to report. The author has nothing to report. The author has nothing to report. The author declares no conflicts of interest. The author, Teja Narra, was the guarantor of this work. Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.
Teja Narra (Wed,) studied this question.