Significance of diffusing capacity of the lungs for carbon monoxide on chronic thromboembolic pulmonary hypertension Authors: Shun Minatsuki, Masaru Hatano, Kouta Funakoshi, Yu Taniguchi, Shiro Adachi, Takumi Inami, Kazuya Hosokawa, Jun Yamashita, Hitoshi Ogino, Ichizo Tsujino, Nobuhiro Yaoita, Nobutaka Ikeda, Nobuhiro Tanabe, Hiroto Shimokawahara, Kayoko Kubota, Ayako Shigeta, Koichiro Tatsumi, Koshin Horimoto, Yoshito Ogihara, Yoshihiro Dohi, Takahiro Hiraide, Takashi Kawakami, Hidekazu Ikemiyagi, Yuichi Tamura, Yoshihiro Fukumoto, Kohtaro Abe, Satoshi Ikeda Published: 2026 ABSTRACT AbstractReduced diffusing capacity of the lungs for carbon monoxide (DLco) reflects microvasculopathy in chronic thromboembolic pulmonary hypertension, yet its clinical value is uncertain. In a Japanese nationwide registry (2018–2023) we studied 1270 patients: 486 formed an event cohort and 299 a treatment cohort who underwent pulmonary endarterectomy or balloon pulmonary angioplasty. Lower baseline DLco was indicative of smaller postprocedural improvements in mean pulmonary artery pressure, pulmonary vascular resistance and cardiac index (all p≤0.023) and a higher risk of clinical events (HR 0.971, p=0.005). Outcomes deteriorated below 59.6%, indicating DLco may help stratify prognosis and treatment benefit. Introduction In chronic thromboembolic pulmonary hypertension (CTEPH), the availability of treatment options, such as pulmonary endarterectomy (PEA), balloon pulmonary angioplasty (BPA) and medical therapy,1 has expanded. Additionally, advancement in the techniques of PEA and BPA, which are invasive but potentially curative, has led to considerable improvements in patients’ prognosis.1 As a result, the treatment goal of CTEPH has shifted toward relieving pulmonary hypertension (PH), with freedom from symptoms, pulmonary hypertension-specific drugs and supplemental oxygen therapy regarded as markers of complete treatment accomplishment,2 but no appropriate strategy has been established. Microvasculopathy is considered a potential contributing factor to CTEPH, with medical therapy believed to have a direct effect on this damage.3 However, aside from an improvement in cardiac output, the haemodynamic improvement of medical therapy remains limited compared with invasive therapy.4 Additionally, the recurrence of PH after invasive therapy3 and interpatient variability in BPA sessions to achieve haemodynamic improvements has been observed,4 although the cause remains unclear. The diffusing capacity of the lungs for carbon monoxide (DLco), a respiratory function index, decreases for various reasons such as microvascular thrombosis, which reduces blood flow to the microvessels.5 In patients with CTEPH, reduced DLco may reflect pulmonary arterial microvasculopathy,6 and a low DLco is associated with poor outcomes in patients undergoing medical therapy without BPA.7 Single-centre analyses have shown that BPA improves haemodynamics and exercise tolerance regardless of DLco.8 However, a low DLco, a surrogate marker of microvasculopathy, may attenuate improvements in mean pulmonary arterial pressure (mPAP) and pulmonary vascular resistance (PVR), requiring additional BPA sessions.8 Based on these findings, we hypothesise that microvasculopathy is a limiting factor in treating CTEPH. Using nationwide registry data, this study aimed to analyse the effect of DLco on the haemodynamic status, exercise tolerance after invasive therapy and clinical events, such as mortality and worsening PH. Methods The Chronic Thromboembolic Pulmonary Hypertension Anticoagulant (CTEPH AC) registry includes all patients diagnosed at 35 Japanese institutions. We retrospectively screened consecutive registrations (August 2018–July 2023). More information about the CTEPH AC registry is provided in online supplemental appendix 1. Records lacking baseline DLco, follow-up or baseline data collected >12 months before enrolment were excluded. The remaining patients were defined as the event cohort for prognostic analysis. The treatment cohort, a subset of the event cohort, consisted of patients who underwent PEA or BPA during follow-up. The event cohort addressed prognostic implications of baseline DLco, the treatment cohort focused on DLco as an indicator of haemodynamic and functional response to invasive therapy. The clinical outcome of the event cohort relationship between baseline DLco and clinical events was defined as a composite of death, worsening PH, the requirement for invasive rescue therapy owing to deteriorating conditions, lung transplantation and bleeding events. In the treatment cohort, the clinical outcome was the effect of baseline DLco on postprocedural changes in haemodynamic parameters, 6-minute walk distance and B-type natriuretic peptide (BNP) concentration. We constructed a receiver operating characteristic curve (ROC) in the event cohort, using baseline DLco and the composite clinical event as the binary endpoint. The first DLco value that achieves a specificity of >80% was selected to detect high-risk patients while minimising false positives. Kaplan-Meier event rates and the log-rank test were used to compare survival estimates after stratification using the DLco cut-off value obtained by the ROC curve. The methods of statistical analysis are provided in online supplemental appendix 2. Results We analysed 1270 patients: 486 in the event and 299 in the treatment cohort (online supplemental figure 1). In the event and treatment cohort, mean age was 66.9±12.4 and 66.0±12.8 years, with women accounting for 71.4% and 69.6%, respectively. Baseline mPAP averaged 34.1±11.5 and 37.0±10.5 mm Hg and %DLco 75.1±19.1 and 75.8±19.2% (table 1). Within the follow-up period, 38 patients were treated with PEA and 274 with BPA. Regarding PH medication, 158 were in the event cohort, and 107 were in the treatment cohort. In the treatment cohort, mPAP, PVR and BNP improved significantly after invasive therapy (all p<0.001). Lower baseline DLco independently predicted attenuated improvement in mPAP (standardised β −0.192, p=0.023), PVR (β −0.132, p=0.013) and CI (β=0.239, p<0.001) (table 2). In the event cohort, over a median of 1334 days, 26 patients (5.3%) reached the composite endpoint. Baseline DLco was the independent indicator (HR 0.971, p=0.005) (table 3). The ROC analysis yielded an area under the curve of 0.65; a cut-off of <59.6 % achieved a specificity of 80.9% and sensitivity of 46.2%. Event rate was 3.6% versus 12.0% for DLCO <59.6 versus ≥59.6% (HR 3.5; 95% CI 1.34 to 9.26; log-rank p<0.001) (online supplemental figure 2). Table 1Baseline characteristicsEvent cohort(N=486)Treatment cohort(N=299)Age, year66.9±12.466.0±12.8Women, %347 (71.4)208 (69.6)Body mass index, kg/m223.5±4.223.7±4.5WHO functional class (I/II/III/IV)54/238/184/1017/134/140/8Median follow-up days, days1334 (702–1445)802 (342–1191)Brain natriuretic peptide, pg/mLn (missing)37.0 (17.0–100.0)425 (61)41.3 (18–134.8)254 (45)6-minute walk distance, mn (missing)373±118418 (68)359±116251 (48)Treatment history of CTEPH before registration Pulmonary endarterectomy, %22 (4.5)7 (2.3) Balloon pulmonary angioplasty, %198 (40.7)77 (25.8) Medical therapy, %245 (50.4)149 (49.8) Supplemental oxygen, %34 (7.0)17 (5.7) Treatment-naïve, %139 (28.6)118 (39.5)Comorbidities Hypercoagulable disorder, %30 (6.2)17 (5.7) Active cancer/history of cancer, %38 (7.8)24 (8.0) Thyroid disease or hormone replacement therapy, %22 (4.5)12 (4.0) Intravenous device, %17 (3.5)11 (3.7) Inflammatory bowel disease, %2 (0.4)1 (0.3) COPD/ILD, %20 (4.1)9 (3.0) Hemiplegia/paraplegia, %4 (0.8)3 (1.0) Use of antipsychotics, %51 (10.5)28 (9.4) History of acute venous thromboembolism, %205 (42.2)124 (41.5) Varicose veins, %23 (4.7)17 (5.7)Haemodynamic status Mean right arterial pressure, mm Hg n (missing)5.4±3.3477 (9)5.6±3.3295 (4) Mean pulmonary arterial pressure, mm Hg n (missing)34.1±11.5486 (0)37.0±10.5299 (0) Mean pulmonary artery wedge pressure, mm Hg n (missing)8.3±3.2485 (1)8.3±3.2299 (0) Cardiac index, L/min/m2 n (missing)2.7±0.7484 (2)2.6±0.7297 (2) Pulmonary vascular resistance, dyn/s/cm−5 n (missing)547±354486 (0)632±357299 (0) PaO2, mm Hg n (missing)65.2±15.2388 (99)64.0±13.3241 (58) SvO2, % n (missing)66.7±7.8453 (33)64.6±7.9275 (24)Respiratory function TLC, % n (missing)102.6±15.3442 (44)103.7±14.9269 (30) FVC, % n (missing)98.7±17.7485 (1)99.3±17.1299 (0) FEV1.0%, % n (missing)86.5±16.7478 (8)86.9±16.2294 (5) %DLco, %75.1±19.175.8±19.2Anticoagulant Warfarin, %158 (32.5)85 (28.4) Direct anticoagulant drug, %318 (65.4)210 (70.2) Dabigatran, %1 (0.2)1 (0.3) Rivaroxaban, %105 (21.6)74 (24.7) Apixaban, %109 (22.4)65 (21.7) Edoxaban, %103 (21.2)70 (23.4) No oral anticoagulant, %10 (2.1)4 (1.3)Pulmonary vasodilators Any pulmonary vasodilator, %245 (50.4)149 (49.8) PDE-5 inhibitors/sGC stimulators, %227 (46.7)136 (45.5) Riociguat, %216 (44.4)132 (44.1) Prostacyclin analogue, PGI2 receptor agonists, %39 (8.0)27 (9.0) Selexipag, %23 (4.7)14 (4.7) Endothelin receptor antagonists, %34 (7.0)18 (6.0)Data are presented as the mean±SD, median (Q1–Q3) or n (%).COPD/ILD, chronic obstructive pulmonary disease/interstitial lung disease; CTEPH, chronic thromboembolic pulmonary hypertension; %DLCO, per cent predicted diffusing capacity of the lungs for carbon monoxide; FEV1.0%, forced expiratory volume % in 1 second; FVC, forced vital capacity; PaO2, partial pressure of arterial oxygen; PDE, phosphodiesterase; PGI2, prostaglandin-I2; sGC, soluble guanylate cyclase; SvO2, mixed venous oxygen saturation; TLC, total lung capacity. Baseline characteristics Data are presented as the mean±SD, median (Q1–Q3) or n (%). COPD/ILD, chronic obstructive pulmonary disease/interstitial lung disease; CTEPH, chronic thromboembolic pulmonary hypertension; %DLCO, per cent predicted diffusing capacity of the lungs for carbon monoxide; FEV1.0%, forced expiratory volume % in 1 second; FVC, forced vital capacity; PaO2, partial pressure of arterial oxygen; PDE, phosphodiesterase; PGI2, prostaglandin-I2; sGC, soluble guanylate cyclase; SvO2, mixed venous oxygen saturation; TLC, total lung capacity. Table 2Comparison of the haemodynamic status, exercise tolerance and BNP concentrations before and after treatment and the effect of DLco on post-treatment parametersComparisonEffect of DLco*Overalln=299BaselineFollow-upChangeP valueStandardised regression coefficient95% CIP valueHaemodynamics Mean RAP, mm Hg n (missing)5.6±3.3295 (4)4.7±3.2204 (95)−1.0(−3.0 to 1.0) (95)<0.001 Mean PAP, mm Hg n (missing)37.0±10.5299 (0)23.5±6.7291 (8)−12.0(−19.0 to −6.0) (8)<0.001−0.192(−0.085 to −0.006)0.023 Cardiac index, L/min/m2 n (missing)2.6±0.7297 (2)2.7±0.7289 (10)0.1(−0.3 to 0.5) (12)<0.0010.239(0.005 to 0.012)<0.001 Mean PAWP, mm Hg n (missing)8.3±3.2299 (0)8.7±3.3204 (95)0.5(−2.0 to 3.0) (95)0.049 PVR, dyn/s/cm−5 n (missing)631.0±357.4299 (0)293.4±150.9291 (8)−242.0(−490.0 to −122.0) (8)<0.001−0.132(−1.896 to −0.224)0.013PaO2, mm Hgn (missing)64.0±13.3241 (58)71.4±18.8151 (148)9.3(2.0 to 16.0) (164)<0.001 SvO2, % n (missing)64.6±7.9275 (24)69.3±6.6266 (33)4.0(0.0 to 8.0) (43)<0.0010.155(0.015 to 0.093)0.006BNP, pg/mLn (missing)41.3(18.0 to 134.8)254 (45)21.2(9.6 to 44.3)256 (43)−11.0(−89.6 to 2.5) (48)<0.001−0.116(−2.685 to −0.011)0.048 6-minute walk distance, m n (missing)359±116251 (48)412±111210 (89)51.0(5.0 to 110.0) (104)<0.001Values are presented as the mean±SD or median (Q1 to Q3).*Based on analysis of covariance adjusted for each baseline value and baseline value of DLco.BNP, B-type natriuretic peptide; DLco, diffusing capacity of the lungs for carbon monoxide; PaO2, partial pressure of arterial oxygen; PAP, pulmonary arterial pressure; PAWP, pulmonary artery wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; SvO2, mixed venous oxygen saturation. Comparison of the haemodynamic status, exercise tolerance and BNP concentrations before and after treatment and the effect of DLco on post-treatment parameters Values are presented as the mean±SD or median (Q1 to Q3). Based on analysis of covariance adjusted for each baseline value and baseline value of DLco. BNP, B-type natriuretic peptide; DLco, diffusing capacity of the lungs for carbon monoxide; PaO2, partial pressure of arterial oxygen; PAP, pulmonary arterial pressure; PAWP, pulmonary artery wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; SvO2, mixed venous oxygen saturation. Table 3Univariate and multivariate analyses for the composite outcomeUnivariate analysisUnadjusted multivariate analysisAdjusted multivariate analysisHR (95% CI)P valueHR (95% CI)P valueHR (95% CI)P valueAge0.995 (0.965 to 1.026)0.7340.573 (0.544 to 1.002)0.992Gender1.155 (0.502 to 2.656)0.7351.279 (0.544 to 3.007)0.573BMI0.965 (0.873 to 1.067)0.4880.635 (0.883 to 1.079)0.635WHO-FC1.895 (1.063 to 3.377)0.030BNP1.002 (1.001 to 1.004)<0.0011.002 (1.000 to 1.003)0.0111.002 (1.000 to 1.003)0.0246MWD0.995 (0.992 to 0.998)<0.001Mean RAP1.090 (0.981 to 1.211)0.110Mean PAP1.035 (1.005 to 1.066)0.0211.024 (0.991 to 1.059)0.1601.025 (0.991 to 1.061)0.150CI1.106 (0.655 to 1.866)0.707PAWP1.002 (0.891 to 1.127)0.968PVR1.001 (1.000 to 1.001)0.232SvO20.943 (0.904 to 0.983)0.006TLC0.987 (0.961 to 1.014)0.330FVC0.977 (0.958 to 0.997)0.027FEV1.0%0.985 (0.962 to 1.009)0.230%DLco0.970 (0.949 to 0.992)0.0090.971 (0.951 to 0.992)0.0070.971 (0.950 to 0.991)0.005BMI, body mass index; BNP, B type natriuretic peptide; CI, cardiac index; %DLco, per cent predicted diffusing capacity of the lungs for carbon monoxide; FEV1.0%, forced expiratory volume % in 1 s; FVC, forced vital capacity; 6MWD, 6-minute walk distance; PAP, pulmonary artery pressure; PAWP, pulmonary artery wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; SvO2, mixed venous oxygen saturation; TLC, total lung capacity; WHO-FC, WHO functional class. Univariate and multivariate analyses for the composite outcome BMI, body mass index; BNP, B type natriuretic peptide; CI, cardiac index; %DLco, per cent predicted diffusing capacity of the lungs for carbon monoxide; FEV1.0%, forced expiratory volume % in 1 s; FVC, forced vital capacity; 6MWD, 6-minute walk distance; PAP, pulmonary artery pressure; PAWP, pulmonary artery wedge pressure; PVR, pulmonary vascular resistance; RAP, right atrial pressure; SvO2, mixed venous oxygen saturation; TLC, total lung capacity; WHO-FC, WHO functional class. Discussion Baseline DLco was inversely related to postinterventional haemodynamic gains and independently linked to the composite clinical endpoint. Lower baseline DLco showed smaller reductions in mPAP and PVR and a lower increase in cardiac index, and patients with baseline DLco below 59.6% experienced more adverse events during follow-up. The 59.6% threshold was selected as the first value achieving ≥80% specificity; while it separated survival curves, its prognostic utility should be considered exploratory and require validation in external cohorts. DLco is inexpensive and widely available, and—together with haemodynamics and imaging—may help refine surveillance intensity or guide adjunctive therapy. However, it should complement rather than replace established risk parameters. Our observations also suggest that residual microvasculopathy can blunt the efficacy of PEA or BPA: invasive procedures relieve obstruction prior to microvessel, yet diffusion capacity did not improve post-treatment, implying persistent microvessel disease.9 The role of microvasculopathy in the treatment of CTEPH is becoming recognised,10 while not fully understood. Our data suggest that low DLCO may exhibit distal disease that limits haemodynamic improvement after PEA or BPA, implying the need to individualise therapy and follow-up around diffusion capacity. Further research into targeting microvasculopathy may help improve outcomes in this high-risk group of patients with CTEPH. Limitations of this study include the following: retrospective design, single-country registry and nearly 60% case exclusion limit generalisability. Baseline data collected up to 12 months pre-enrolment could mask the treatment effect during the follow-up period. DLco/alveolar volume and haemoglobin-adjusted DLco were unavailable. Supplementary material