Effects of inhaled nitric oxide on ventilation/perfusion mismatch assessed by electrical impedance tomography in intubated patients with moderate-to-severe ARDS: a prospective physiological study Authors: Yue Yu, Rui Tian, Ling Zhang, Yang Chen, Yunqi Dai, Jiayi Guan, Xiaoli Wang, Xuelian Chen, Jing Wu, Tingting Pan, Hongping Qu, Ruoming Tan Published: 2026 ABSTRACT BackgroundInhaled nitric oxide (iNO) improves oxygenation in acute respiratory distress syndrome (ARDS), but its precise physiological effects on ventilation-perfusion (V/Q) matching and inflammation over 24 h remain to be fully elucidated. To determine the effect of iNO on V/Q matching, gas exchange, and systemic inflammatory response in mechanically ventilated patients with moderate-to-severe ARDS.MethodsWe conducted a prospective, single-center, randomized controlled trial in 40 mechanically ventilated patients with moderate-to-severe ARDS (PaO2/FiO2 ≤ 200 mmHg). Patients were randomized 1:1 to receive either iNO at a clinician-determined dose plus standard mechanical ventilation (iNO group) or standard mechanical ventilation alone (Control group). The primary endpoint was the change in V/Q matching parameters, assessed by electrical impedance tomography (EIT), at 24 hours.ResultsForty patients were randomized (20 per group). At 24 h, the iNO group showed a significantly greater improvement in V/Q matching compared to the control group. The magnitude of reduction in both EIT-measured dead space (between-group difference, P < 0.001) and shunt (between-group difference, P < 0.001 was significantly larger in the iNO group. This physiological improvement was reflected in gas exchange: the median PaO2/FiO2 ratio in the iNO group increased from 114.0 to 240.0 (P < 0.001), an improvement significantly greater than that in the control group (P < 0.001). Similarly, the physiological shunt (Qs/Qt) decreased more substantially in the iNO group (P < 0.001). iNO treatment also resulted in a greater reduction in pulmonary artery pressure (P < 0.001). No significant between-group differences were observed in inflammatory biomarkers or safety parameters.ConclusionsIn patients with moderate-to-severe ARDS, 24-hour administration of inhaled nitric oxide significantly improves oxygenation by enhancing ventilation-perfusion matching. This effect is driven by a marked reduction in both EIT-measured shunt and dead space. This study provides a direct mechanistic rationale for iNO’s physiological effects and supports the potential of using EIT to guide a more personalized therapeutic approach.Trial registrationClinicaltrials.gov; Unique identifier: ChiCTR2500113100; Retrospectively registered; Registration date: 25 November 2025.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12931-026-03744-w. Background Acute respiratory distress syndrome (ARDS) remains a leading cause of mortality in critically ill patients, defined by acute inflammatory lung injury, refractory hypoxemia, and bilateral pulmonary infiltrates [1, 2]. The cornerstone of ARDS pathophysiology is a profound ventilation-perfusion (V/Q) mismatch, driven by an increase in intrapulmonary shunt and alveolar dead space [3]. This disruption in gas exchange is mediated by uncontrolled inflammation, endothelial injury, and microthrombosis, which can also lead to pulmonary hypertension and subsequent right ventricular dysfunction [3, 4]. Inhaled nitric oxide (iNO) is a potent and selective pulmonary vasodilator that theoretically targets this core pathophysiology. By preferentially dilating vessels in well-ventilated lung regions, iNO is intended to redirect pulmonary blood flow, improve V/Q matching, and thereby enhance arterial oxygenation [5]. However, despite this compelling rationale, the clinical role of iNO remains one of the most enduring controversies in critical care. Numerous randomized controlled trials and large meta-analyses have consistently shown that while iNO can transiently improve oxygenation within the first 24 h, it confers no benefit on 28-day mortality and may increase the risk of acute kidney injury [5–8]. Consequently, major clinical practice guidelines recommend against its routine use, reserving it only as a rescue therapy for life-threatening hypoxemia [9]. Furthermore, research indicates that iNO can effectively inhibit neutrophil aggregation and reduce the release of inflammatory cytokines such as IL-6 and IL-8, thereby mitigating damage to the alveolar-capillary membrane [10, 11]. A critical limitation of prior trials, which may explain this disconnect between physiological effect and clinical outcome, is the reliance on global surrogates of gas exchange, such as the PaO2/FiO2 ratio, to assess efficacy [8, 12]. This approach fails to directly measure the mechanistic impact of iNO on V/Q matching and cannot distinguish physiological responders from non-responders, potentially diluting any treatment effect in heterogeneous study populations. The advent of electrical impedance tomography (EIT) provides a non-invasive, radiation-free, bedside tool to overcome this limitation. EIT allows for the real-time visualization and, crucially, the quantification of regional ventilation and perfusion, enabling a direct assessment of shunt, dead space, and overall V/Q mismatch [13]. While recent physiological study has highlighted the potential of EIT to identify iNO responders over short durations [14], no randomized controlled trial has yet utilized this technology to rigorously evaluate the effects of iNO on V/Q matching over a clinically relevant 24-hour period. Therefore, we aimed to determine the impact of iNO on V/Q matching, gas exchange, and systemic inflammation in patients with moderate-to-severe ARDS, using EIT as the primary tool for physiological assessment. Our primary objective was to evaluate the effectiveness of iNO in improving EIT-derived V/Q matching parameters at 24 h, while secondary objectives were to assess its impact on the inflammatory response and its overall safety. Methods Study design and oversight This prospective, single-center, randomized controlled trial was conducted in the intensive care unit (ICU) of Ruijin Hospital, Jiaotong University. The study protocol was approved by the hospital’s Institutional Ethics Committee (no.2024-LLS-33), and the trial was registered in the Chinese Clinical Trial Registry(ChiCTR2500113100). Written informed consent was obtained from all patients or their legally authorized representatives prior to enrollment, in accordance with the Declaration of Helsinki. Patient population We enrolled adult patients (≥ 18 years) who were receiving invasive mechanical ventilation and met the GLOBAL definition for moderate-to-severe ARDS [15]. Exclusion criteria were: pregnancy; severe hemodynamic instability requiring escalating doses of vasopressors (defined as a ≥ 20% increase in norepinephrine-equivalent dose within the preceding 6 h, or initiation of a second vasopressor to maintain MAP ≥ 65 mmHg); known contraindications to iNO therapy (e.g., methemoglobin reductase deficiency); or conditions precluding the placement of an EIT belt (e.g., extensive thoracic dressings or trauma). Randomization and blinding Using a computer-generated sequence, patients were randomized in a 1:1 ratio to either the iNO group or the control group. Allocation was concealed using sealed, opaque envelopes prepared by an independent research pharmacist. Due to the nature of the intervention, blinding of the clinical care team and investigators to group allocation was not feasible. Study interventions and standardized care Upon enrollment, baseline demographic and clinical data were recorded, including age, sex, weight, height, and severity of illness scores (APACHE II and SOFA). Recorded variables additionally included body mass index (BMI), ARDS etiology (classified as pulmonary vs. extrapulmonary), duration of respiratory symptoms before intubation, duration of mechanical ventilation before ARDS diagnosis, total mechanical ventilation duration, ICU length of stay, sedation and analgesia regimens (drugs), use of neuromuscular blocking agents (proportion, duration), vasopressor support (type, dose range), corticosteroid therapy, transfusion requirements, and 24-hour fluid balance (see Additional file1 Tables S4). The GCS component of APACHE II and SOFA was scored prior to sedation initiation. All patients were managed with a standardized lung-protective ventilation strategy and supportive care protocol. Mechanical Ventilation: All patients were ventilated in a pressure-controlled mode to achieve a tidal volume of 6 mL/kg of predicted body weight. The respiratory rate was adjusted to maintain an arterial pH between 7.30 and 7.45. PEEP and FiO₂ were titrated by the clinical team to maintain SpO₂ within a target range of 90%–96%, with the upper bound serving as a ceiling to avoid hyperoxia; FiO₂ was actively reduced whenever SpO₂ exceeded 96% [16–19].General Supportive Care: Sedation, analgesia, use of neuromuscular blocking agents, fluid management, and decisions regarding prone positioning were managed according to institutional protocols, applied uniformly to both groups.Control Group: Patients received the standardized mechanical ventilation and supportive care strategy alone.iNO Group: In addition to standard care, patients received iNO administered continuously into the inspiratory limb of the ventilator circuit via the eNOaire Nitric Oxide Generator and Delivery System (Novlead Biotechnology Co., Ltd., China). The iNO dose was set at a fixed 20 ppm throughout the 24-hour study period without titration, based on dose-response data from prior studies showing near-maximal oxygenation benefit at 10–20 ppm with acceptable safety [20]. Safety was monitored by measuring methemoglobin levels at baseline, 4 h after initiation, and daily thereafter, and by continuous monitoring of inspired nitrogen dioxide (NO2) levels, with a target of < 3 ppm. Mechanical Ventilation: All patients were ventilated in a pressure-controlled mode to achieve a tidal volume of 6 mL/kg of predicted body weight. The respiratory rate was adjusted to maintain an arterial pH between 7.30 and 7.45. PEEP and FiO₂ were titrated by the clinical team to maintain SpO₂ within a target range of 90%–96%, with the upper bound serving as a ceiling to avoid hyperoxia; FiO₂ was actively reduced whenever SpO₂ exceeded 96% [16–19]. Mechanical Ventilation: All patients were ventilated in a pressure-controlled mode to achieve a tidal volume of 6 mL/kg of predicted body weight. The respiratory rate was adjusted to maintain an arterial pH between 7.30 and 7.45. PEEP and FiO₂ were titrated by the clinical team to maintain SpO₂ within a target range of 90%–96%, with the upper bound serving as a ceiling to avoid hyperoxia; FiO₂ was actively reduced whenever SpO₂ exceeded 96% [16–19]. General Supportive Care: Sedation, analgesia, use of neuromuscular blocking agents, fluid management, and decisions regarding prone positioning were managed according to institutional protocols, applied uniformly to both groups. General Supportive Care: Sedation, analgesia, use of neuromuscular blocking agents, fluid management, and decisions regarding prone positioning were managed according to institutional protocols, applied uniformly to both groups. Control Group: Patients received the standardized mechanical ventilation and supportive care strategy alone. Control Group: Patients received the standardized mechanical ventilation and supportive care strategy alone. iNO Group: In addition to standard care, patients received iNO administered continuously into the inspiratory limb of the ventilator circuit via the eNOaire Nitric Oxide Generator and Delivery System (Novlead Biotechnology Co., Ltd., China). The iNO dose was set at a fixed 20 ppm throughout the 24-hour study period without titration, based on dose-response data from prior studies showing near-maximal oxygenation benefit at 10–20 ppm with acceptable safety [20]. Safety was monitored by measuring methemoglobin levels at baseline, 4 h after initiation, and daily thereafter, and by continuous monitoring of inspired nitrogen dioxide (NO2) levels, with a target of < 3 ppm. iNO Group: In addition to standard care, patients received iNO administered continuously into the inspiratory limb of the ventilator circuit via the eNOaire Nitric Oxide Generator and Delivery System (Novlead Biotechnology Co., Ltd., China). The iNO dose was set at a fixed 20 ppm throughout the 24-hour study period without titration, based on dose-response data from prior studies showing near-maximal oxygenation benefit at 10–20 ppm with acceptable safety [20]. Safety was monitored by measuring methemoglobin levels at baseline, 4 h after initiation, and daily thereafter, and by continuous monitoring of inspired nitrogen dioxide (NO2) levels, with a target of < 3 ppm. Measurements and data collection Physiological data were collected at two predefined time points: T0 (baseline, immediately before randomization) and T1 (24 h post-randomization). The following data were collected at each time point: (1) respiratory parameters (ventilator settings, tidal volume, peak and plateau pressures, respiratory system static compliance [Crs = VT/(Pplat − PEEP)], and airway resistance); (2) arterial blood gas analysis (pH, PaO2, PaCO2, PaO2/FiO2, alveolar-arterial oxygen gradient [PA-aO2],, physiological shunt fraction [Qs/Qt], and physiological dead space fraction [VD/VT]); (3) hemodynamic parameters (systemic blood pressures, heart rate, and pulmonary artery pressure); (4) EIT-derived V/Q parameters; and (5) laboratory markers (inflammatory biomarkers and organ function indices). Electrical Impedance Tomography (EIT) EIT measurements were performed using the PulmoVista 500 system (Dräger Medical). An electrode belt was placed around the patient’s thorax at the fourth or fifth intercostal space. Regional ventilation distribution was derived from tidal impedance changes (the difference between end-inspiratory and end-expiratory impedance at each pixel). Regional perfusion was assessed using a first-pass indicator dilution technique as described by Spinelli et al. [21]: a 10 mL bolus of 5% hypertonic NaCl—chosen for its high electrical conductivity and validated safety profile—was rapidly injected via a central venous catheter with the tip in the superior vena cava during a 20-second end-inspiratory hold (to eliminate ventilation-related impedance changes). The resulting impedance-time dilution curves were used to quantify regional perfusion distribution. By spatially co-registering the ventilation and perfusion maps on a pixel-by-pixel basis, each lung unit was classified as: matched (ventilated and perfused), dead space (ventilated but not perfused), or shunt (perfused but not ventilated). The EIT image was divided into four equal horizontal regions of interest (ROIs) along the ventral-to-dorsal axis: ROI 1 (most ventral/non-dependent), ROI 2 (ventral-intermediate), ROI 3 (dorsal-intermediate), and ROI 4 (most dorsal/dependent). A single perfusion measurement was performed at each time point. EIT and arterial blood gas measurements were obtained simultaneously. Study endpoints The primary endpoint was the change in the percentage of EIT-derived V/Q mismatch (sum of shunt and dead space units) from baseline to 24 h. Secondary endpoints included: (1) changes in gas exchange (PaO2/FiO2 ratio) and pulmonary hemodynamics; (2) changes in plasma concentrations of inflammatory biomarkers (IL-6, TNF-α); and (3) safety, assessed by the incidence of methemoglobinemia (> 5%), significant hypotension, and acute kidney injury. Statistical analysis As this was a mechanistic study evaluating a novel physiological endpoint for which no prior data existed to inform a formal power calculation, we planned to enroll 40 patients (20 per group). This sample size is consistent with previous physiological studies in ARDS and was deemed adequate to detect meaningful between-group differences in EIT-derived parameters [21, 22]. Continuous variables were expressed as medians with interquartile ranges (IQR) and categorical variables as frequencies and percentages. Between-group comparisons of continuous variables were performed using the Mann-Whitney U test. Within-group changes from baseline were assessed using the Wilcoxon signed-rank test. Categorical variables were compared using the chi-squared test or Fisher’s exact test. Repeated measures ANOVA with post-hoc Bonferroni correction was used to analyze the effect of time points on the variables. Mauchly’s test was conducted for sphericity, and the Greenhouse–Geisser correction was applied when sphericity was violated. The Pearson regression coefficient was utilized to evaluate the correlation between continuous variables. A two-tailed P value < 0.05 was considered statistically significant. All analyses were performed using SPSS (Version 29.0, IBM Corp) and GraphPad Prism (Version 10.0, GraphPad Software).In all tables, P1 denotes the within-group comparison from T0 to T1 in the iNO group (Wilcoxon signed-rank test); P2 denotes the within-group comparison from T0 to T1 in the control group (Wilcoxon signed-rank test); P3 denotes the between-group comparison of the change from baseline (T1 − T0) between the two groups (Mann-Whitney U test). Results Patient characteristics A total of 40 patients with moderate-to-severe ARDS were included. The median age was 67 years, and 60% were male. Baseline demographics, severity scores, and etiologies of ARDS were comparable between groups, with no significant differences in clinical characteristics (Table 1). ARDS etiologies were predominantly pulmonary in origin (pneumonia being the most common cause), with a smaller proportion of extrapulmonary causes (non-pulmonary sepsis, post-surgical). No patients with isolated acute brain injury or polytrauma were included. A detailed etiology breakdown is provided in Table 1. The details of prone position, sedations and NMBA are shown in additional file 1 (Table S5 and Table S6). Table 1Baseline characteristic of the enrolled patients between groupsOverall (n = 40)iNO (n = 20)Control (n = 20) P Males, n (%)24 (58)9 (45)14 (70)0.110Age, yr67 (55–75)65 (52–76)69 (58–72)0.465Height, cm166 (159–172)165 (159–170)169 (160–174)0.158Weight, kg62 (50–68)61 (50–70)62 (53–67)0.725BMI21.5 (19.4–25.2)22.1 (18.3–25.9)21.5 (19.8–25.2)0.799ARDS etiology Pneumonia, n(%)18 (45)9 (45)9 (45)1.000 Non-pulmonary sepsis, n(%)15 (38)8 (40)7 (35)0.744 Post-surgical, n(%)7 (18)3 (15)4 (20)0.677 APACHE II23 (17–26)22 (18–26)25 (17–27)0.364 SOFA10 (8–12)11 (8–13)9 (8–11)0.404 PA, mmHg38 (30–49)46 (38–55)39 (36–49)0.188 Respiratory parameters Pi, cmH2O16 (15–18)16 (15–18)16 (14–18)0.336 RR, bpm20 (20–22)20 (20–22)20 (20–22)0.930 PEEP, cmH2O10 (8–12)10 (9–12)10 (8–10)0.151 FiO2, %60 (50–74)70 (50–80)50 (50–68)0.112 VT, ml383 (350–420)375 (346–408)420 (350–434)0.170 Ppeak, cmH2O30 (27–33)32 (27–34)29 (27–33)0.342 Pplat, cmH2O25 (21–28)26 (22–28)24 (20.0-28.8)0.307 Crs, ml/cmH2O40.0 (35.0-43.8)40.5 (35.3–42.8)39.5 (33.5–45.5)0.860 Rrs, cmH2O/L/s10.0 (9.0–11.0)10.0 (9.0-11.8)10.0 (9.0–11.0)0.793 BG parameters P/F144.5 (101.3–167.0)114.0 (99.3-163.5)147.5 (132.0-170.3)0.221 PA−aO2320.5 (227.0-413.0)375.8 (265.9–429.0)284.3 (177.3-325.4)0.032 Qs/Qt, %21.9 (16.5–26.5)24.3 (16.4–27.1)20.2 (16.7–24.6)0.512 VD/VT, %29.5 (22.2–32.4)28.1 (21.1–33.0)29.6 (24.7–32.0)0.841 EIT data Deadspace, %28.3 (23.4–34.9)29.0 (22.5–34.9)28.3 (25.2–34.9)0.947 Shunt, %9.5 (6.9–12.8)8.7 (4.8–14.7)9.8 (7.4–12.7)0.414 Matched region, %61.3 (51.8–66.4)59.4 (52.8–66.5)62.0 (51.4–66.4)1.000Data are expressed as N (%) or median (interquartile range) unless otherwise specifiedDefinition of abbreviations: iNO inhaled nitric oxide, ARDS Acute Respiratory Distress Syndrome, BMI Body Mass Index, APACHE Acute Physiology Assessment and Chronic Health Evaluation, SOFA Sequential Organ Failure Assessment, Pi inspiratory pressure, RR respiratory rate, PEEP positive end-expiratory pressure, FiO2 inspired fraction of O2 ratio, VT tidal volume, Ppeak peak pressure, Pplat plateau pressure, Crs respiratory system compliance, Rrs respiratory system resistance, PA pulmonary arterial pressure, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomographyP value by chi-squared test or Mann-Whitney U test, as appropriate Table 1Baseline characteristic of the enrolled patients between groupsOverall (n = 40)iNO (n = 20)Control (n = 20) P Males, n (%)24 (58)9 (45)14 (70)0.110Age, yr67 (55–75)65 (52–76)69 (58–72)0.465Height, cm166 (159–172)165 (159–170)169 (160–174)0.158Weight, kg62 (50–68)61 (50–70)62 (53–67)0.725BMI21.5 (19.4–25.2)22.1 (18.3–25.9)21.5 (19.8–25.2)0.799ARDS etiology Pneumonia, n(%)18 (45)9 (45)9 (45)1.000 Non-pulmonary sepsis, n(%)15 (38)8 (40)7 (35)0.744 Post-surgical, n(%)7 (18)3 (15)4 (20)0.677 APACHE II23 (17–26)22 (18–26)25 (17–27)0.364 SOFA10 (8–12)11 (8–13)9 (8–11)0.404 PA, mmHg38 (30–49)46 (38–55)39 (36–49)0.188 Respiratory parameters Pi, cmH2O16 (15–18)16 (15–18)16 (14–18)0.336 RR, bpm20 (20–22)20 (20–22)20 (20–22)0.930 PEEP, cmH2O10 (8–12)10 (9–12)10 (8–10)0.151 FiO2, %60 (50–74)70 (50–80)50 (50–68)0.112 VT, ml383 (350–420)375 (346–408)420 (350–434)0.170 Ppeak, cmH2O30 (27–33)32 (27–34)29 (27–33)0.342 Pplat, cmH2O25 (21–28)26 (22–28)24 (20.0-28.8)0.307 Crs, ml/cmH2O40.0 (35.0-43.8)40.5 (35.3–42.8)39.5 (33.5–45.5)0.860 Rrs, cmH2O/L/s10.0 (9.0–11.0)10.0 (9.0-11.8)10.0 (9.0–11.0)0.793 BG parameters P/F144.5 (101.3–167.0)114.0 (99.3-163.5)147.5 (132.0-170.3)0.221 PA−aO2320.5 (227.0-413.0)375.8 (265.9–429.0)284.3 (177.3-325.4)0.032 Qs/Qt, %21.9 (16.5–26.5)24.3 (16.4–27.1)20.2 (16.7–24.6)0.512 VD/VT, %29.5 (22.2–32.4)28.1 (21.1–33.0)29.6 (24.7–32.0)0.841 EIT data Deadspace, %28.3 (23.4–34.9)29.0 (22.5–34.9)28.3 (25.2–34.9)0.947 Shunt, %9.5 (6.9–12.8)8.7 (4.8–14.7)9.8 (7.4–12.7)0.414 Matched region, %61.3 (51.8–66.4)59.4 (52.8–66.5)62.0 (51.4–66.4)1.000Data are expressed as N (%) or median (interquartile range) unless otherwise specifiedDefinition of abbreviations: iNO inhaled nitric oxide, ARDS Acute Respiratory Distress Syndrome, BMI Body Mass Index, APACHE Acute Physiology Assessment and Chronic Health Evaluation, SOFA Sequential Organ Failure Assessment, Pi inspiratory pressure, RR respiratory rate, PEEP positive end-expiratory pressure, FiO2 inspired fraction of O2 ratio, VT tidal volume, Ppeak peak pressure, Pplat plateau pressure, Crs respiratory system compliance, Rrs respiratory system resistance, PA pulmonary arterial pressure, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomographyP value by chi-squared test or Mann-Whitney U test, as appropriate Baseline characteristic of the enrolled patients between groups Data are expressed as N (%) or median (interquartile range) unless otherwise specified Definition of abbreviations: iNO inhaled nitric oxide, ARDS Acute Respiratory Distress Syndrome, BMI Body Mass Index, APACHE Acute Physiology Assessment and Chronic Health Evaluation, SOFA Sequential Organ Failure Assessment, Pi inspiratory pressure, RR respiratory rate, PEEP positive end-expiratory pressure, FiO2 inspired fraction of O2 ratio, VT tidal volume, Ppeak peak pressure, Pplat plateau pressure, Crs respiratory system compliance, Rrs respiratory system resistance, PA pulmonary arterial pressure, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomography P value by chi-squared test or Mann-Whitney U test, as appropriate Blood gas measurements At baseline, blood gas parameters were generally similar between groups, except for a lower PA–aO₂ in the control group (Table 1). From T0 to T1, patients receiving iNO exhibited marked improvements in gas exchange: P/F ratio increased significantly, while PA–aO₂, Qs/Qt, and VD/VT decreased (all P < 0.001, Table 2). In the control group, P/F ratio increased significantly, while Qs/Qt and VD/VT declined modestly, but no significant changes were observed in PA-aO2. Between-group comparisons confirmed that the magnitude of improvement in P/F, PA–aO₂, Qs/Qt, and VD/VT was greater with iNO therapy (Fig. 1). Table 2Comparison of blood gas parameters and electrical impedance tomography parameters between the two groupsiNO (n=20)Control (n=20) P 1 P 2 P 3 T0T1T0T1BG parameters P/F114.0 (99.3-163.5)240.0 (188.8-288.5)147.5 (132.0-170.3)169.5 (155.0-190.8)<0.0010.019<0.001 PA-aO2, mmHg375.8 (265.9-429.0)195.2 (176.4-281.9)284.3 (177.3-325.4)287.4 (220.0-325.3)<0.0010.881<0.001 Qs/Qt, %24.3 (16.4-27.1)11.3 (9.1-17.4)20.2 (16.7-24.6)19.3 (13.5-21.6)<0.0010.0160.001 VD/VT, %28.1 (21.1-33.0)19.5 (17.4-22.2)29.6 (24.7-32.0)26.5 (20.9-28.1)<0.001<0.0010.033EIT parameters Global Deadspace, %29.0 (22.5-34.9)11.9 (9.3-16.3)28.3 (25.2-34.9)24.9 (22.2-30.8)<0.001<0.001<0.001 Shunt, %8.7 (4.8-14.7)4.2 (2.2-7.0)9.8 (7.4-12.7)8.6 (6.5-11.2)<0.001<0.001<0.001 Matched region, %59.4 (52.8-66.5)82.6 (78.1-86.1)62.0 (51.4-66.4)66.5 (57.3-70.4)<0.001<0.001<0.001 ROI 1 Deadspace, %36.2 (29.8-45.3)15.2 (12.0-18.5)39.3 (28.5-60.1)38.2 (27.6-58.3)<0.001<0.001<0.001 Shunt, %17.9 (7.2-26.1)6.5 (3.0-11.6)13.2 (7.6-24.9)12.8 (7.4-24.2)<0.001<0.001<0.001 Matched region, %41.6 (32.7-55.7)78.3 (66.9-84.6)35.6 (24.0-61.6)37.5 (26.3-62.7)<0.001<0.001<0.001 ROI 2 Deadspace, %31.6 (19.4-38.5)13.1 (7.8-18.1)32.5 (20.9-38.2)31.5 (20.2-37.0)<0.001<0.001<0.001 Shunt, %3.3 (1.7-17.9)1.4 (0.6-8.6)3.1 (0.1-6.0)3.0 (0.1-5.8)<0.001<0.001<0.001 Matched region, %59.9 (55.8-67.1)83.8 (78.4-89.6)63.9 (60.4-75.2)64.9 (61.6-75.9)<0.001<0.001<0.001 ROI 3 Deadspace, %18.3 (7.8-28.1)17.4 (7.4-26.7)15.6 (12.1-31.8)15.2 (11.7-30.8)<0.001<0.0010.091 Shunt, %8.1 (2.6-12.0)7.7 (2.4-11.4)9.4 (4.7-22.0)9.1 (4.6-21.3)<0.001<0.0010.698 Matched region, %70.2 (61.9-77.3)71.7 (63.8-78.4)66.4 (58.2-73.8)67.4 (59.4-74.6)<0.001<0.0010.033 ROI 4 Deadspace, %38.7 (0.0-76.7)36.7 (0.0-72.9)30.6 (9.8-80.6)29.6 (9.5-78.2)<0.001<0.0010.277 Shunt, %2.7 (0.0-12.8)2.5 (0.0-11.9)34.6 (11.9-59.2)33.5 (11.6-57.4)0.002<0.0010.102 Matched region, %40.0 (13.4-70.8)43.0 (17.8-72.4)22.2 (0.0-42.6)24.6 (3.0-44.3)<0.001<0.0010.060Data are expressed as median (interquartile range) unless otherwise specifiedDefinition of abbreviations: iNO inhaled nitric oxide, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomography, ROI region of interest, T0 before treatment, T1 1 day after treatmentP value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriateT0 vs T1 in iNO group, P1T0 vs T1 in control group, P2Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Table 2Comparison of blood gas parameters and electrical impedance tomography parameters between the two groupsiNO (n=20)Control (n=20) P 1 P 2 P 3 T0T1T0T1BG parameters P/F114.0 (99.3-163.5)240.0 (188.8-288.5)147.5 (132.0-170.3)169.5 (155.0-190.8)<0.0010.019<0.001 PA-aO2, mmHg375.8 (265.9-429.0)195.2 (176.4-281.9)284.3 (177.3-325.4)287.4 (220.0-325.3)<0.0010.881<0.001 Qs/Qt, %24.3 (16.4-27.1)11.3 (9.1-17.4)20.2 (16.7-24.6)19.3 (13.5-21.6)<0.0010.0160.001 VD/VT, %28.1 (21.1-33.0)19.5 (17.4-22.2)29.6 (24.7-32.0)26.5 (20.9-28.1)<0.001<0.0010.033EIT parameters Global Deadspace, %29.0 (22.5-34.9)11.9 (9.3-16.3)28.3 (25.2-34.9)24.9 (22.2-30.8)<0.001<0.001<0.001 Shunt, %8.7 (4.8-14.7)4.2 (2.2-7.0)9.8 (7.4-12.7)8.6 (6.5-11.2)<0.001<0.001<0.001 Matched region, %59.4 (52.8-66.5)82.6 (78.1-86.1)62.0 (51.4-66.4)66.5 (57.3-70.4)<0.001<0.001<0.001 ROI 1 Deadspace, %36.2 (29.8-45.3)15.2 (12.0-18.5)39.3 (28.5-60.1)38.2 (27.6-58.3)<0.001<0.001<0.001 Shunt, %17.9 (7.2-26.1)6.5 (3.0-11.6)13.2 (7.6-24.9)12.8 (7.4-24.2)<0.001<0.001<0.001 Matched region, %41.6 (32.7-55.7)78.3 (66.9-84.6)35.6 (24.0-61.6)37.5 (26.3-62.7)<0.001<0.001<0.001 ROI 2 Deadspace, %31.6 (19.4-38.5)13.1 (7.8-18.1)32.5 (20.9-38.2)31.5 (20.2-37.0)<0.001<0.001<0.001 Shunt, %3.3 (1.7-17.9)1.4 (0.6-8.6)3.1 (0.1-6.0)3.0 (0.1-5.8)<0.001<0.001<0.001 Matched region, %59.9 (55.8-67.1)83.8 (78.4-89.6)63.9 (60.4-75.2)64.9 (61.6-75.9)<0.001<0.001<0.001 ROI 3 Deadspace, %18.3 (7.8-28.1)17.4 (7.4-26.7)15.6 (12.1-31.8)15.2 (11.7-30.8)<0.001<0.0010.091 Shunt, %8.1 (2.6-12.0)7.7 (2.4-11.4)9.4 (4.7-22.0)9.1 (4.6-21.3)<0.001<0.0010.698 Matched region, %70.2 (61.9-77.3)71.7 (63.8-78.4)66.4 (58.2-73.8)67.4 (59.4-74.6)<0.001<0.0010.033 ROI 4 Deadspace, %38.7 (0.0-76.7)36.7 (0.0-72.9)30.6 (9.8-80.6)29.6 (9.5-78.2)<0.001<0.0010.277 Shunt, %2.7 (0.0-12.8)2.5 (0.0-11.9)34.6 (11.9-59.2)33.5 (11.6-57.4)0.002<0.0010.102 Matched region, %40.0 (13.4-70.8)43.0 (17.8-72.4)22.2 (0.0-42.6)24.6 (3.0-44.3)<0.001<0.0010.060Data are expressed as median (interquartile range) unless otherwise specifiedDefinition of abbreviations: iNO inhaled nitric oxide, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomography, ROI region of interest, T0 before treatment, T1 1 day after treatmentP value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriateT0 vs T1 in iNO group, P1T0 vs T1 in control group, P2Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Comparison of blood gas parameters and electrical impedance tomography parameters between the two groups Data are expressed as median (interquartile range) unless otherwise specified Definition of abbreviations: iNO inhaled nitric oxide, BG blood gas, P/F PaO2/FiO2, EIT electrical impedance tomography, ROI region of interest, T0 before treatment, T1 1 day after treatment P value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriate T0 vs T1 in iNO group, P1 T0 vs T1 in control group, P2 Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Fig. 1Changes in blood gas parameters before (T0) and 24h after (T1) treatment. Comparisons of venous admixture (Qs/Qt), physiological dead space (VD/VT), oxygenation index (P/F), and alveolar-arterial oxygen gradient (PA−aO2) between the iNO and control groups. Data are presented as individual patient trajectories with bars representing median and interquartile range. ****P < 0.0001; *P < 0.05; NS, not significant Changes in blood gas parameters before (T0) and 24h after (T1) treatment. Comparisons of venous admixture (Qs/Qt), physiological dead space (VD/VT), oxygenation index (P/F), and alveolar-arterial oxygen gradient (PA−aO2) between the iNO and control groups. Data are presented as individual patient trajectories with bars representing median and interquartile range. ****P < 0.0001; *P < 0.05; NS, not significant EIT measurements At baseline, global EIT-derived indices of deadspace, shunt, and matched ventilation–perfusion regions were comparable between groups, except for a lower shunt fraction in ROI 4 in the iNO group (see Additional file 1; Table S1). From T0 to T1, patients treated with iNO demonstrated a significant reduction in deadspace fraction and shunt fraction, accompanied by a notable increase in matched regions (all P < 0.01, Table 2). Importantly, these improvements were not limited to global indices: regional analysis across the four ROIs revealed consistent reductions in deadspace and shunt, along with parallel increases in matched regions (Fig. 2). The changes were most pronounced in ventral non-dependent regions (ROI 1 and ROI 2), where baseline dead space was highest. Fig. 2Comparison of EIT parameters between the iNO and control groups. The figure displays EIT-derived parameters evaluated before treatment (T0) and 24 hours after treatment (T1). Comparisons of dead space, shunt, and matched region are shown for the global lung and four gravitational regions of interest (ROI 1–4). A–E Comparison of dead space percentage in the global lung (A), ROI 1 (B), ROI 2 (C), ROI 3 (D), and ROI 4 (E). F–J Comparison of shunt percentage in the global lung (F), ROI 1 (G), ROI 2 (H), ROI 3 (I), and ROI 4 (J). K–O Comparison of matched region percentage in the global lung (K), ROI 1 (L), ROI 2 (M), ROI 3 (N), and ROI 4 (O). Data are presented as median with interquartile range. Light-colored bars represent data at T0 (before treatment), and dark-colored bars represent data at T1 (24h after treatment). Statistical significance is indicated by brackets. iNO = inhaled nitric oxide; ROI = region of interest; NS = not significant. *P < 0.05; ****P < 0.0001 Comparison of EIT parameters between the iNO and control groups. The figure displays EIT-derived parameters evaluated before treatment (T0) and 24 hours after treatment (T1). Comparisons of dead space, shunt, and matched region are shown for the global lung and four gravitational regions of interest (ROI 1–4). A–E Comparison of dead space percentage in the global lung (A), ROI 1 (B), ROI 2 (C), ROI 3 (D), and ROI 4 (E). F–J Comparison of shunt percentage in the global lung (F), ROI 1 (G), ROI 2 (H), ROI 3 (I), and ROI 4 (J). K–O Comparison of matched region percentage in the global lung (K), ROI 1 (L), ROI 2 (M), ROI 3 (N), and ROI 4 (O). Data are presented as median with interquartile range. Light-colored bars represent data at T0 (before treatment), and dark-colored bars represent data at T1 (24h after treatment). Statistical significance is indicated by brackets. iNO = inhaled nitric oxide; ROI = region of interest; NS = not significant. *P < 0.05; ****P < 0.0001 Similarly, patients in the control group showed a modest but significant decline in deadspace and shunt fractions with an increase in matched regions (all P < 0.001, Table 2). Regional trends in the control group were directionally similar to those observed with iNO, but the overall magnitude of improvement was smaller. Between-group comparisons demonstrated that iNO led to significantly greater reductions in deadspace and shunt fractions and larger increases in matched regions than mechanical ventilation alone (all P < 0.001; Table 2). However, when stratified by individual ROIs, the differences between groups were less evident (Fig. 2), suggesting that iNO exerted a regionally preferential effect, with the greatest impact on ventral regions where baseline dead space predominated, rather than a uniform global effect on V/Q matching. Notably, both groups showed directionally similar improvements that were more evident in ventral regions, but the magnitude was significantly greater in the iNO group. Hemodynamic parameters Baseline hemodynamic variables were comparable between groups except for a lower MAP in the iNO group (see Additional file 1; Table S1). From T0 to T1, systemic hemodynamics remained stable in both groups, whereas pulmonary artery pressure decreased significantly. The reduction was more pronounced in patients receiving iNO compared with mechanical ventilation alone (P = 0.001, Table 3). Table 3Comparison of hemodynamic parameters between the two groupsiNO (n = 20)Control (n = 20) P 1 P 2 P 3 T0T1T0T1SBP, mmHg123 (107–138)131 (114–138)135 (120–144)124 (120–136)0.4780.1500.101DBP, mmHg66 (53–75)67 (57–72)73 (65–86)72 (62–76)0.9130.4590.163MBP, mmHg83 (77–94)83 (78–94)92 (84–103)90 (82–94)0.6410.2040.203HR, bpm85 (77–101)80 (69–88)86 (74–100)81 (67–93)0.0230.0380.786PA, mmHg43 (34–51)30 (21–36)35 (30–45)35 (30–40)< 0.0010.0010.001Data are expressed as median (interquartile range) unless otherwise specifiedFor PA in the control group, although the median values at T0 and T1 are both 35 mmHg, the Wilcoxon signed-rank test evaluates paired within-subject differences rather than group-level medians. The significant p-value (p = 0.001) reflects asymmetric changes in individual patients despite identical group mediansDefinition of abbreviations: iNO inhaled nitric oxide, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure, HR heart rate, PA pulmonary arterial pressure, T0 before treatment, T1 1 day after treatmentP value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriateT0 vs. T1 in iNO group, P1T0 vs. T1 in control group, P2Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Table 3Comparison of hemodynamic parameters between the two groupsiNO (n = 20)Control (n = 20) P 1 P 2 P 3 T0T1T0T1SBP, mmHg123 (107–138)131 (114–138)135 (120–144)124 (120–136)0.4780.1500.101DBP, mmHg66 (53–75)67 (57–72)73 (65–86)72 (62–76)0.9130.4590.163MBP, mmHg83 (77–94)83 (78–94)92 (84–103)90 (82–94)0.6410.2040.203HR, bpm85 (77–101)80 (69–88)86 (74–100)81 (67–93)0.0230.0380.786PA, mmHg43 (34–51)30 (21–36)35 (30–45)35 (30–40)< 0.0010.0010.001Data are expressed as median (interquartile range) unless otherwise specifiedFor PA in the control group, although the median values at T0 and T1 are both 35 mmHg, the Wilcoxon signed-rank test evaluates paired within-subject differences rather than group-level medians. The significant p-value (p = 0.001) reflects asymmetric changes in individual patients despite identical group mediansDefinition of abbreviations: iNO inhaled nitric oxide, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure, HR heart rate, PA pulmonary arterial pressure, T0 before treatment, T1 1 day after treatmentP value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriateT0 vs. T1 in iNO group, P1T0 vs. T1 in control group, P2Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Comparison of hemodynamic parameters between the two groups Data are expressed as median (interquartile range) unless otherwise specified For PA in the control group, although the median values at T0 and T1 are both 35 mmHg, the Wilcoxon signed-rank test evaluates paired within-subject differences rather than group-level medians. The significant p-value (p = 0.001) reflects asymmetric changes in individual patients despite identical group medians Definition of abbreviations: iNO inhaled nitric oxide, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure, HR heart rate, PA pulmonary arterial pressure, T0 before treatment, T1 1 day after treatment P value by Mann-Whitney U test or Wilcoxon signed rank test, as appropriate T0 vs. T1 in iNO group, P1 T0 vs. T1 in control group, P2 Comparison of the change from baseline (T1−T0) between the iNO and control groups, P3 Inflammatory biomarkers Inflammatory biomarkers including WBC, CRP, PCT, IL-6, IL-8, IL-10, TNF-α, and IFN-γ showed no significant temporal changes and did not differ between groups (see Additional file 1; Figure S2, Table S2). Adverse events No significant changes in safety-related laboratory parameters—including platelet count, renal function, muscle injury markers, and methemoglobin—were observed during the study period (see Additional file 1; Figure S3, Table S3). Discussion In this prospective, randomized controlled trial in patients with moderate-to-severe ARDS, we demonstrated that a 24-hour administration of inhaled nitric oxide (iNO) significantly improves ventilation-perfusion (V/Q) matching. Using electrical impedance tomography (EIT), we provide the first evidence from a randomized setting that this physiological improvement is driven by a marked reduction in both intrapulmonary shunt and alveolar dead space. These direct physiological benefits were mirrored by substantial improvements in gas exchange and pulmonary hemodynamics, without a detectable effect on systemic inflammatory biomarkers or an increased incidence of adverse events during the study period. The choice of a 24-hour assessment period is clinically significant for several reasons. First, it extends beyond the short-term measurements (minutes to hours) of prior EIT studies [8, 20], demonstrating that iNO’s effect on V/Q matching is sustained rather than merely transient. Second, meta-analyses have consistently identified the first 24 h as the window of maximum oxygenation benefit from iNO, making this the most clinically relevant period for physiological assessment [23]. Third, a 24-hour endpoint provides a practical decision point: if EIT demonstrates improved V/Q matching at 24 h, this could inform clinical decisions about continuing therapy. Our findings provide a crucial mechanistic link that helps to explain the known physiological effects of iNO. Intrapulmonary shunting and increased dead space are primary pathophysiological features of ARDS [3]. By directly visualizing regional lung function, we showed that iNO effectively targets both lesions. This is consistent with its known mechanism of activating the soluble guanylate cyclase (sGC)-cyclic guanosine monophosphate (cGMP) pathway, which preferentially dilates pulmonary vessels in well-ventilated lung regions [24–26]. This action redistributes pulmonary blood flow away from non-ventilated (shunt) areas and reverses the excessive pathological vasoconstriction in ventilated regions, thereby optimizing V/Q matching [27]. Prior work has shown that iNO can reduce the physiological shunt (Qs/Qt) by 10–30% [20, 28–30], and a notable RCT demonstrated a reduction from 36% to 25% [29]. Our study complements these findings by providing direct, regional evidence of this phenomenon and highlights that the shunt-reducing effect is robust at 24 h, though its long-term durability may be influenced by dose and tolerance, as suggested by Gerlach et al. [20]. While the physiological effect of iNO on oxygenation is well-established, its clinical value remains highly controversial. Consistent with previous studies, we confirmed that the profound improvements in V/Q matching translated into a significant short-term increase in the PaO2/FiO2 ratio [8, 23, 31]. However, this improvement in oxygenation has not consistently led to benefits in hard clinical endpoints. Large-scale randomized controlled trials and subsequent meta-analyses have decisively shown that routine iNO use in unselected ARDS patients does not reduce mortality, shorten the duration of mechanical ventilation, or decrease the need for ECMO [7, 8, 27, 32]. Our study, by confirming a purely physiological benefit, reinforces the conclusion that improving this single parameter is likely insufficient to alter the overall course of a complex, multi-system syndrome like ARDS. Furthermore, we found no significant effect of iNO on systemic inflammation. This observation aligns with previous large clinical trials, such as the analysis by Taylor et al. of 385 patients, which found no difference in plasma or bronchoalveolar lavage fluid inflammatory markers (e.g., IL-6, TNF-α) between iNO and control groups [8, 20]. This clinical reality stands in contrast to promising pre-clinical data from in-vitro and animal models, which suggest iNO may limit neutrophil adhesion, reduce pro-inflammatory cytokine production, and decrease reactive oxygen species by inhibiting NADPH oxidase [24, 25, 33–36]. The disconnect between the pre-clinical anti-inflammatory potential and the lack of effect in human ARDS trials suggests that the systemic inflammatory cascade in patients is either too overwhelming to be modulated by inhaled therapy or operates through pathways resistant to NO-mediated effects. This study also contributes important safety data. We observed no significant adverse events, including methemoglobinemia, hemodynamic compromise, or renal dysfunction. This is noteworthy given the known risks associated with iNO. A meta-analysis by Karam et al. found that iNO significantly increased the risk of acute kidney injury (AKI) [7], and other studies have noted potential for NO to inhibit platelet aggregation or form toxic NO₂ [37–42]. While our study was not powered to definitively assess safety, the absence of these signals in our closely monitored cohort provides some reassurance for its use as a short-term physiological intervention. The true value of our findings may therefore lie not in advocating for the routine use of iNO, but in paving the way for a more personalized approach to its application. The central challenge in iNO therapy has been the inability to distinguish responders from non-responders at the bedside. EIT has the potential to solve this problem. Our study demonstrates that EIT can be used to directly visualize and quantify a physiological response to iNO. This opens the door to using EIT as a bedside tool to screen for “physiological responders.” Future clinical trials could employ an enrichment strategy, whereby only patients who demonstrate a positive V/Q response to a test dose of iNO are randomized to continue therapy. Such a trial, by focusing on a population with a demonstrated mechanistic benefit, would be far better positioned to finally determine if improving V/Q matching can lead to improved clinical outcomes. Our regional analysis revealed that iNO’s predominant effect was on ventral non-dependent regions, where dead space was highest at baseline. This finding is physiologically consistent with iNO’s mechanism: as a selective pulmonary vasodilator, iNO preferentially dilates vessels in well-ventilated lung regions. In supine ARDS patients, ventral regions are relatively over-ventilated compared to perfusion, creating dead space. iNO redirects blood flow to these regions, converting dead space into matched V/Q units. In contrast, dorsal dependent regions are often consolidated or atelectatic, limiting iNO delivery and hence its vasodilatory effect. This regional selectivity may partly explain why iNO has not improved mortality in clinical trials: the predominant physiological benefit occurs in relatively well-ventilated regions rather than in the most pathological dorsal zones that harbor the greatest burden of shunt and consolidation. Future EIT-guided strategies could potentially identify patients in whom iNO also improves dorsal V/Q matching, representing a more meaningful therapeutic target. Despite its strengths, including the randomized design and novel primary endpoint, our study has several limitations. First, as a single-center study with a small sample size, its external validity is limited. Second, EIT measures a two-dimensional cross-section of the thorax and may not capture the full three-dimensional heterogeneity of V/Q mismatch throughout the entire lung. Third, we did not evaluate the effects of iNO beyond 24 h, so we cannot comment on the durability of the observed physiological benefits; moreover, dose-response studies suggest potential tolerance development with prolonged iNO exposure. Fourth, although we report detailed sedation, NMBA, vasopressor, corticosteroid, and fluid balance data (Additional file 1), these co-interventions may have influenced perfusion distribution. Fifth, as our cohort consisted of intubated patients, these results may not be generalizable to non-intubated patients with respiratory failure. Finally, our oxygenation target range of SpO₂ 90%–96% is consistent with contemporary conservative-oxygen practice but is slightly more permissive at the upper bound than thresholds proposed in some recent trials. Although FiO₂ was actively down-titrated to avoid SpO₂ exceeding 96%, we cannot fully exclude that residual oxygen exposure influenced the absolute values of EIT-derived parameters. Because the protocol was applied symmetrically across both arms, this is unlikely to explain the observed between-group differences. Conclusions In conclusion, among intubated patients with moderate-to-severe ARDS, a 24-hour administration of inhaled nitric oxide significantly improves oxygenation by enhancing ventilation-perfusion matching, an effect driven by reductions in both intrapulmonary shunt and dead space. While confirming the physiological rationale for iNO, our study simultaneously highlights its lack of systemic anti-inflammatory effect. This work provides robust, mechanistically-focused evidence that supports a paradigm shift away from using iNO in unselected populations and toward an EIT-guided, personalized strategy to select patients most likely to benefit from its physiological effects. Supplementary Information Supplementary Material 1. Supplementary Material 1.