Background: Diffusion capacity of the lungs for carbon monoxide (DLCO) represents a biomarker of oxygen uptake across the alveolar capillary interface. However, evidence on the predictive value of DLCO across heart failure (HF) phenotypes is lacking.

Methods: The MyoVasc-study (N=3,289; NCT04064450) is a prospective cohort study on HF. HF-phenotypes were defined according to the universal definition of heart failure; i.e. HF with preserved ejection fraction (HFpEF), HF with mildly reduced ejection fraction (HFmrEF), and HF with reduced ejection fraction (HFrEF). During a standardized 5-hour examination in a dedicated study centre pulmonary functional status was ascertained via body plethysmography (MasterScreen Body, Carefusion, Germany) using the single breath method for the determination of DLCO (test gas: CO; tracer gas: Helium). According to previous analysis DLCO < 60 % predicted was defined as reduced. Structured follow up investigations with subsequent adjudication and validation of clinical endpoints assessed information on clinical outcome. The primary study endpoint was worsening of heart failure, defined as a composite of HF-hospitalization and cardiac death.

Results: The analysis sample comprised 1,084 individuals (age 67.1±10.1 years, 33.7 % females) with HF of whom 528 had HFpEF, 299 HFmrEF, and 254 HFrEF. The median NT-proBNP was 345.5 pg/ml, and 438 subjects had a reduced DLCO. Participants with reduced DLCO had higher NT-proBNP-levels (135 pg/ml vs 408 pg/ml, P<0.001), a lower EF (55.3% vs 49.5%, P <0.001), and increased E/E’-ratio (7.9 vs 10.1, P<0.001) compared to a DLCO ≥ 60 %. During a median follow up period of 4.0 years 223 subjects experienced the primary study endpoint worsening of HF consisting of 181 HF-hospitalizations, and 83 cardiac deaths. In the overall analysis sample reduced DLCO was related to an increasing incidence for worsening of HF (20.8 % vs 35.6 %, P <0.001). This relationship was consistent acorss the two subcomponents HF-hospitalization (11.0 % vs 20.0 %, P<0.001), and cardiac death (14.2% vs 31.7 %, P<0.001). Cox regression analysis adjusting for age, sex and height revealed that DLCO <60 % predicted worsening of HF in HFpEF (hazard ratio (HR) 2.60 95% confidence interval (CI) 1.56-4.32, P<0.001), but not in HFmrEF (HR 1.23 95% CI 0.71-2.11, P=0.46) and HFrEF (HR 1.09 95% CI 0.73-1.61, P=0.69; P for interaction 0.009 vs HFpEF). Consistently DLCO <60 % was a HFpEF-specific predictor of HF-hospitalization (HR_HFpEF 2.27 95% CI 1.26-4.11; HR_HFmrEF 1.13 95% CI 0.63-2.03, HR_HFrEF 1.29 95% CI 0.84-1.99), while a relation between DLCO <60 % and cardiac death was observed across all HF-phenotypes (HR_HFpEF 5.04 95% CI 1.99-12.75; HR_HFmrEF 2.95 95% CI 1.13-7.74, HR_HFrEF 2.27 95% CI 1.15-4.49). After additional adjustment for the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) Score and NT-proBNP reduced DLCO remained an independent predictor for worsening of HF (HR 2.27 955 CI 1.35-3.80, P=0.002) and HF-hospitalization (HR 2.00 95 % CI 1.11-3.62, P=0.02) in subjects with HFpEF. In this fully adjusted model the strongest relationship between DLCO <60 % and cardiac death was observed in HFpEF: HR_HFpEF 3.87 95% CI 1.47-10.15; HR_HFmrEF 2.07 95% CI 0.84-5.07; HR_HFrEF 1.77 95% CI 0.87-3.59.

Conclusion: In HFpEF structural remodelling of the alveolar capillary interface seems of high clinical relevance as DLCO represents a strong and independent predictor of clinical outcome in these subjects.