This was an observational and multicenter study. A total of 21 consecutive treatment naïve patients admitted for the first-time diagnostic right heart catheterization (RHC) at the Regional Centers for PAH in Uppsala or Lund between October 2010 and April 2015 were recruited. Inclusion criteria were clinical diagnosis of PAH and age above 18 years. Exclusion criteria were patients in PH group 2–5. Blood samples were taken and examined for the levels of l-arginine, ADMA, SDMA, l-ornithine, and l-citrulline. Detailed medical histories and physical examinations included age, gender, weight, body surface area (BSA), height, electrocardiography, echocardiography, 6 min walking distance (6MWD), NT-proBNP, and RHC. The 6MWD was carried out in accordance with the American Thoracic Society’s guidelines [24]. The individuals included were screened for high blood pressure, diabetes mellitus, ischemic heart disease, stroke, renal failure, and thyroid disease. For comparison, 27 healthy subjects without drug treatment and 14 patients treated for LVSD were examined for ADMA, SDMA, and l-arginine. All healthy controls were non-smokers and had no symptoms or signs of common cold at the time of recruitment.

The study was approved by the Independent Ethics Committee in Uppsala (Dnr 2010/343) and Lund (Dnr 2010/114, 2011/368, 2011/777), and conducted in accordance with the Helsinki Declaration. All patients gave their informed consent.

All PAH patients underwent routine RHC as part of the initial diagnostic work-up. RHC was performed at the Hemodynamic Laboratory using a fiberoptic thermodilution pulmonary artery catheter, Becton–Dickinson Criticath SP5 107 HTD catheter at Uppsala University hospital, and a Swan-Ganz catheter 7.5 F (2.5 mm), Edward Lifesciences LLC, Irvine, CA, USA at Lund University hospital. The catheter was inserted through the right internal jugular vein into the pulmonary artery. Recorded hemodynamic data were: heart rate, central venous pressure (CVP), systolic pulmonary artery pressure (sPAP), mPAP, mean pulmonary artery wedge pressure (mPAWP), mean right atrial pressure (mRAP), CO, and CI. Mixed venous and arterial oxygen saturation was also measured. Pressures were registered with a Siemens^® system and flow was calculated with the thermodilution technique or with Fick’s principle. Pulmonary vascular resistance (PVR), systemic vascular resistance (SVR), and PVR/SVR were calculated. Systemic blood pressure was measured invasively in Uppsala and non-invasively in Lund. Experienced cardiologists performed all examinations.

Blood samples were obtained from the pulmonary artery or from peripheral veins during diagnostic RHC. All blood samples were collected in EDTA tubes (BD Diagnostics, Burlington, NC, USA) and plasma was retrieved after centrifugation. Plasma was separated and stored at −70 °C in Uppsala and at −80 °C in Lund waiting for biochemical analysis.

Plasma levels of ADMA, SDMA, l-arginine, l-ornithine, and l-citrulline were analyzed with liquid chromatography–tandem mass spectrometry (LC–MS/MS) at the Swedish National Veterinary Institute in Uppsala, Sweden. Sample pretreatment was as follows. To 100 µL of plasma, 50 µL of water and 50 µL of the internal standard solution containing ²H₇-ADMA, ²H₆-SDMA, ¹³C₆-arginine, 2H6-ornithine, and ²H₇-citrulline were added followed by addition of 400 µL of acetonitrile/trifluoroacetic acid/propionic acid (1000/0.25/10 v/v/v). The samples were vortex-mixed for 5 min followed by centrifugation for 5 min at 10,000g. The supernatants were transferred vials for analysis with ultra-high-performance liquid chromatography–tandem mass spectrometry (UHPLC–MS/MS). A Waters Acquity UPLC system was coupled to a Quattro Ultima Pt tandem quadrupole mass spectrometer with an electrospray interface operating in the positive mode (Waters Corporation, Milford, MA, USA). The column was an Indra Almtakt Amino Acid (length 100 mm, I.D. 2.0 mm, particle size 3 µm) kept at 25 °C. The mobile phase consisted of (A) 100 mM ammonium formate in water and (B) 0.1% formic acid in acetonitrile. The injection volume was 10 µL. The elution was carried out as follows: isocratic at 80% A for 1 min, increase to 95% A during 1 min, constant at 95% A for 5.5 min, and decrease to 80% A during 0.1 min, constant at 80% A for 2.4 min. The flow-rate was 200 µL/min. The five analytes were quantified simultaneously in the same chromatographic run using a positive capillary voltage of 0.50 kV and a cone voltage of 40 V. The desolvation and source block temperatures were 350 and 120 °C, respectively, and the cone and desolvation gas flows were 120 and 924 L/h, respectively. The quantifications were performed in the selected reaction monitoring (SRM) mode with the collision cell filled with argon gas at a pressure of 1.9 × 10⁻³ mBar. The mass transitions used in SRM were m/z 203 → 46 for ADMA (collision energy 18 eV), m/z 210 → 77 for [²H₇]-ADMA (collision energy 23 eV), m/z 203 → 172 for SDMA (collision energy 16 eV), m/z 209 → 116 for [²H₆]-SDMA (19 eV), m/z 175 → 70 for arginine (collision energy 18 eV), m/z 181 → 74 for [¹³C₆]-arginine (collision energy 18 eV), m/z 133 → 70 for ornithine (collision energy 13 eV), m/z 139 → 76 for [¹³H₆]-ornithine (collision energy 14 eV), m/z 176 → 113 for citrulline (collision energy 18 eV), and m/z 183 → 120 for [¹³C₇]-citrulline (collision energy 16 eV). The dwell time was 0.010 s. Stock solutions of ADMA, SDMA, arginine, ornithine, citrulline, and the internal standards were prepared in water at approximately 0.1–0.3 mg/mL. The reference standards and the internal standards were obtained from Toronto Research Chemicals (North York, ON, Canada). To check for matrix effects and to compensate for endogenous levels of the analytes in the spiked plasma, calibration samples were constructed for all three analytes in both water and in control plasma. The calibration curves were constructed using the chromatographic peak area ratio (analyte/internal standard) as a function of analyte concentration. The calibration functions were calculated by linear curve fit using a weighting factor of 1/x ² for all three analytes. The calibration range for ADMA was 0.090–3.4 µM and the precision was in the range of 5.3–7.3%. The calibration range for SDMA was 0.38–3.0 µM and the precision was in the range of 5.8–9.3%. The calibration range for arginine was 4.5–150 µM and the precision was in the range of 3.5–6.2%. The calibration range for ornithine was 4.5–151 µM and the precision was in the range of 4.1–5.2%. The calibration range for citrulline was 4.4–150 µM and the precision was in the range of 2.5–4.8%.

Clinical characteristics, biochemical indicators, and hemodynamic parameters were summarized by median (interquartile range) and investigated biomarkers by median (IQR) or median (range). Non-parametric tests were used due to non-normal data distributions and small sample size. Differences between two groups were evaluated with the Mann–Whitney U test. For testing differences between three groups (patients with PAH, LVSD and healthy subjects), the Kruskal–Wallis test was used. In cases where results of the Kruskal–Wallis test were significant, Mann–Whitney U tests were performed to compare the PAH patients with LVSD patients and healthy subjects, separately. Spearman’s test was used to evaluate the association between hemodynamic variables and methylarginines. Multiple linear regression analysis was used to evaluate associations between l-arginine and 6MWD. p values <0.05 were considered statistically significant. Statistical analyses were performed with SAS 9.3 (SAS Institute Inc, Cary, NC, USA) and IBM SPSS Statistics package v. 24.

Patient characteristics and biochemical data in patients with PAH and LVSD and in healthy subjects are presented in Table 1. The PAH group consisted of 13 women (62%) and eight men (38%) with a median age of 73 (45–85) years. Four of these patients had PAH associated with connective tissue disease and 17 had idiopathic PAH. Three of the PAH patients had a history of ischemic heart disease. None of these patients had signs of left-sided heart disease during the study (PAWP = 12, 6 and 8 mmHg). Patients with PH due to left heart disease have PAWP >15 mmHg according to ESC guidelines for the diagnoses and treatment of pulmonary hypertension [1]. Hemodynamic parameters in PAH patients are shown in Table 2. The LVSD group consisted of eight women (57%) and six men (43%) with a median age of 67 (48–82) years. The majority of the LVSD patients had dilated cardiomyopathy of non-ischemic orient with a median ejection fraction of 33 (15–35) %. Based on echocardiography, no right ventricular failure was present in the patients with LVSD except in one patient. This patient underwent RHC and had an mPAP of 31, mPAWP of 16, and CO of 5.8, and was diagnosed with PH due to left heart disease. The patient is included in the result.

The plasma ADMA, SDMA, and l-arginine levels in PAH, LVSD, and healthy subjects are shown in Table 3. In PAH, median ADMA was 0.50 (0.34–0.91) μM, SDMA 0.83 (0.47–2.43) μM, l-arginine 55.1 (30.0–116.0) μM, l-citrulline 23.5 (10.6–45.1) μM, l-ornithine 83.2 (33.4–126.7) μM, l-arginine/ADMA ratio 102.2 (59.3–230.2), l-arginine/l-ornithine ratio 0.72 (0.35–1.36), and l-arginine/(l-ornithine + l-citrulline) ratio (GABR) 0.51 (0.28–1.08).

In LVSD, median ADMA was 0.56 (0.50–0.74) μM, SDMA 0.74 (0.50–1.08) μM, l-arginine 81.8 (43.8–113.8) μM, and l-arginine/ADMA ratio 140.4 (81.2–214.8). The patient with PH with left heart disease had the lowest l-arginine levels (43.8 μM) in the LVSD group (Fig. 2). In healthy subjects, median ADMA was 0.36 (0.23–0.44) μM, SDMA 0.42 (0.32–0.59) μM, l-arginine 85.8 (58.2–132.8) μM, and l-arginine/ADMA ratio 237.8 (176.5–365.7).

ADMA and SDMA levels were significantly higher, and l-arginine levels and l-arginine/ADMA ratio were significantly lower in PAH compared to healthy controls (p < 0.001). Furthermore, patients with PAH had significantly lower l-arginine levels than patients with LVSD (p = 0.0012) (Fig. 2).

A Spearman’s correlation to determine the relationship between the clinical determinants, and l-arginine and dimethylarginines was performed. In PAH, l-arginine correlated positively with 6MWD (r _s = 0.58, p = 0.006) (Fig. 3) and correlated inversely with creatinine clearance (eGFR) (r _s = −0.52, p = 0.015). l-Arginine/ADMA ratio correlated inversely with WHO functional class (r _s = −0.46, p = 0.043) and eGFR (r _s  = −0.49, p = 0.023). l-Ornithine correlated positively with 6MWD (r _s = 0.54, p = 0.012) and l-citrulline correlated inversely with eGFR (r _s = −0.44, p = 0.044). No significant correlations for cardiovascular risk factors or other clinical determinants were found, and no significant correlations between NT-proBNP and the above clinical determinants in PAH patients were found.

In LVSD, ADMA correlated inversely with 6MWD (r _s = −0.77, p = 0.004). SDMA correlated positively with WHO functional class (r _s = 0.54, p = 0.048). l-Arginine and l-arginine/ADMA ratio correlated inversely with eGFR (r _s = −0.69, p = 0.010) and (r _s = −0.56, p = 0.046), respectively. As in PAH patients, no significant correlations were found between NT-proBNP and clinical determinants were found.

A Spearman’s correlation to determine the relationship between the hemodynamics and the plasma ADMA, SDMA, and l-arginine levels in PAH was performed. l-arginine/l-ornithine ratio and l-arginine/(ADMA + l-ornithine) ratio correlated inversely with mean aorta pressure (r _s = −0.56, p = 0.016) and (r _s = −0.67, p = 0.002), respectively. No other statistically significant relationships with haemodynamic parameters were seen for the metabolic markers. For NT-proBNP, there was a positive correlation with sPAP (r _s = 0.53, p = 0.024) and PVR/SVR (r _s = 0.50, p = 0.047). No other statistically significant relationship for NT-proBNP was seen.