Awake VV-ECMO for severe pneumonia caused by Elizabethkingia anophelis: a case report Authors: Wang Guanglin, Kang Xiuwen, Hu Rong Published: 2026 ABSTRACT AbstractBackgroundElizabethkingia anophelis is phenotypically similar to E. meningoseptica and is often misidentified by conventional methods, delaying appropriate therapy. Awake venovenous extracorporeal membrane oxygenation (VV-ECMO) avoids complications of deep sedation and mechanical ventilation, but its role in severe pulmonary infection with rare pathogens remains underexplored.Case presentationWe report a 62-year-old male with chronic hepatitis B who developed type I respiratory failure and septic shock unresponsive to conventional support. VV-ECMO was initiated on January 23, and awake ECMO management was implemented to preserve spontaneous breathing and cough reflex. Serial metagenomic next-generation sequencing (mNGS) of bronchoalveolar lavage fluid revealed influenza A H1N1, Aspergillus fumigatus, and multidrug-resistant bacteria (Detailed mNGS results are provided in Supplementary Table 2). On day 32, sputum culture suggested E. meningoseptica, but subsequent mNGS identified E. anophelis (322,376 reads). The anti-infective regimen was adjusted to minocycline-based combination therapy. Under awake ECMO support, the patient’s infection markers gradually improved, and he was successfully weaned from ECMO on day 38 and from mechanical ventilation thereafter. He was discharged after recovery.ConclusionThis case demonstrates that awake ECMO can serve as an effective respiratory support platform in complex severe pneumonia. When conventional testing reports E. meningoseptica, clinicians should suspect possible E. anophelis infection, and timely mNGS is recommended for accurate species identification. Minocycline-based combination therapy appears promising for E. anophelis infections.Clinical trial Not applicable.Clinical pearlIn critically ill patients with suspected Elizabethkingia infection, do not rely solely on phenotypic identification; use mNGS to distinguish Elizabethkingia anophelis from Elizabethkingia meningoseptica, and consider early minocycline-based therapy.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12879-026-13735-3.Keywords: Elizabethkingia anophelis, Awake extracorporeal membrane oxygenation, Metagenomic next-generation sequencing, Minocycline, Case report Introduction Elizabethkingia anophelis is phenotypically similar to Elizabethkingia meningoseptica ATCC 13,253^T and Elizabethkingia miricola GTC 862^T. It was first isolated from the midgut of Anopheles gambiae mosquitoes in 2011 and named Elizabethkingia anophelis sp. nov [1]. This bacterium is an aerobic, non-motile, non-spore-forming Gram-negative bacillus and is generally considered an environmental bacterium [2]. It exhibits intrinsic resistance to ampicillin, chloramphenicol, kanamycin, streptomycin, and tetracycline [1], and is resistant to most β-lactam agents, but remains susceptible to piperacillin, aminoglycosides, chloramphenicol, trimethoprim-sulfamethoxazole, ciprofloxacin, and others [3]. For newly identified E. anophelis infections, laboratory susceptibility testing is still necessary. This article reports a case of E. anophelis infection in a patient treated with awake VV-ECMO, aiming to enhance clinicians’ ability to diagnose and manage this disease. Its uniqueness and importance lie in that it demonstrates, for the first time, the successful use of awake VV-ECMO as a respiratory support platform in a patient with severe polymicrobial pneumonia secondary to E. anophelis infection, where conventional sputum culture misidentified the pathogen and metagenomic next-generation sequencing (mNGS) corrected the diagnosis, enabling timely minocycline-based therapy. Case presentation A 62-year-old male with chronic hepatitis B was admitted via the emergency department at 10:49 PM on January 4, 2025, with a 7-day history of cough, yellow-white sputum, fever (maximum 39 ℃), chills, and rigors. A chest CT at a local hospital showed a patchy high-density opacity in the right lower lobe suggestive of inflammation (Fig. 1). He had been taking tenofovir disoproxil fumarate 0.3 g daily for hepatitis B e antigen-positive chronic hepatitis B, with regular follow-up showing normal liver function, HBV-DNA, and alpha-fetoprotein. Chest CT plain scan performed at an outside hospital on January 4, 2025, showing a patchy high-density opacity in the right lower lobe, suggestive of inflammation On admission, physical examination revealed temperature 39.0 ℃, pulse 98 beats/min, respiratory rate 20 breaths/min, and blood pressure 122/73 mmHg. He was alert but lethargic. Coarse breath sounds and moist crackles were audible in the right lower lung. Initial laboratory tests showed white blood cells 16.33 × 109/L (neutrophils 78.2%), C-reactive protein 99.16 mg/L, and positive hepatitis B serology (HBeAg 1.27 IU/mL, HBsAg > 250 IU/mL, HBcAb 44.15 PEIU/mL). Electrocardiogram revealed sinus rhythm, right atrial enlargement, and left axis deviation. After admission to the Department of Respiratory Medicine, the patient received biapenem for anti-infection along with supportive care, but his condition progressed rapidly. His temperature rose to 39.1 ℃. Chest radiographs on January 9 and 10 showed bilateral pneumonia predominantly involving the right lung with rapid progression (Fig. 2). Bronchoscopy revealed hyperemia of the right middle lobe mucosa, scattered milky-white protrusions, and purulent secretions. The first bronchoalveolar lavage fluid metagenomic next-generation sequencing (mNGS) returned positive for influenza A virus H1N1 (816,653 reads) with suspected oseltamivir resistance, and baloxavir marboxil was administered. Chest X-ray (posteroanterior view) obtained at our hospital on January 10, 2025, revealing bilateral pneumonia predominantly involving the right lung with rapid progression mNGS procedure The mNGS testing was performed by Hangzhou Di’an Medical Laboratory. A targeted next-generation sequencing (tNGS) approach was used, based on multiplex PCR amplification and high-throughput sequencing on the MGI sequencing platform. A molecular internal standard was introduced to allow semi-quantitative analysis of pathogen loads. The assay had a detection limit of 100–500 copies/mL. The workflow included nucleic acid extraction, library preparation, sequencing, and bioinformatic analysis against a comprehensive database covering bacteria, fungi, viruses, parasites, and resistance genes. Detailed detection ranges are provided in the supplementary appendix of the test report. On January 13, the patient suddenly developed wheezing. Despite high-flow nasal cannula (60 L/min, FiO2 80%), his oxygen saturation remained 85%, and arterial blood gas showed pH 7.54, PaCO2 30 mmHg, PaO2 45 mmHg, lactate 1.8 mmol/L, and oxygen saturation 86%. Type I respiratory failure was diagnosed. Endotracheal intubation and mechanical ventilation were started at 1:38 PM, and the patient was transferred to the intensive care unit (ICU). In the ICU, he received sedation, analgesia, mechanical ventilation, and empirical nemonoxacin malate. On January 17, the second bronchoalveolar lavage mNGS revealed influenza A H1N1 (849,980 reads) and Aspergillus fumigatus complex (19 reads). Baloxavir marboxil was combined with peramivir. On January 20, fungal biomarkers (1→3)-β-D-glucan 119.48 pg/mL and galactomannan 4.266 µg/L confirmed pulmonary aspergillosis, and voriconazole was added. Despite these measures, his infection markers deteriorated: white blood cells peaked at 34.16 × 109/L, C-reactive protein increased markedly, and procalcitonin peaked at 29.67 ng/mL. Repeated chest X-rays and CT scans showed no significant improvement. On January 20, septic shock was diagnosed. On January 22, a tracheostomy was performed and invasive mechanical ventilation continued (assist/control mode (A/C) mode, tidal volume 450 mL, FiO2 65%, respiratory rate 23 breaths/min, PEEP 3 cmH2O). At that time, the patient’s oxygenation index was only 62 mmHg, meeting the criteria for severe acute respiratory distress syndrome (ARDS). However, because the patient had rheumatic heart disease (severe mitral stenosis), higher positive end-expiratory pressure (PEEP) could reduce preload and worsen right heart function; therefore, a lower PEEP (3 cmH₂O) was initially used for lung-protective ventilation. During subsequent ECMO support, PEEP was gradually titrated to 10 cmH₂O based on respiratory mechanics monitoring (recorded on January 24), combined with prone positioning. Despite lung-protective ventilation and prone positioning, oxygenation remained difficult to maintain, with an oxygenation index of 62 mmHg. Although prone positioning had been applied daily for > 16 h since January 14, and a lung-protective strategy (tidal volume 6 mL/kg ideal body weight, plateau pressure < 30 cmH₂O) was used, the patient’s oxygenation index persistently remained below 80 mmHg. On January 23, arterial blood gas showed PaO₂ 65 mmHg with FiO₂ 55% (oxygenation index 118 mmHg), which later dropped to 62 mmHg. Concurrently, the patient developed septic shock requiring norepinephrine. Moreover, due to rheumatic heart disease (severe mitral stenosis), the patient experienced hemodynamic fluctuations during prone positioning, limiting its duration. After multidisciplinary discussion, the patient met Extracorporeal Life Support Organization (ELSO) criteria for VV-ECMO (oxygenation index < 80 mmHg despite conventional therapy). At 5:00 PM on January 23, venovenous extracorporeal membrane oxygenation (VV-ECMO) was initiated (centrifugal pump speed 4095 rpm, blood flow 4.3 L/min, membrane oxygenator FiO2 90%, sweep gas 3 L/min, water bath 36.8 ℃, heparin 2.8 mg/h). Under stable oxygenation and hemodynamics, the depth of sedation was gradually reduced to allow spontaneous breathing and cooperation, maintaining a Richmond Agitation-Sedation Scale score between − 1 and 0 (awake ECMO). On January 24, the third bronchoalveolar lavage mNGS detected influenza A H1N1 (oseltamivir/peramivir resistant), Acinetobacter baumannii (carbapenem/penicillin resistant), Klebsiella pneumoniae, human herpesvirus type 7, and Aspergillus fumigatus complex (see Supplementary Table 2 for reads and subsequent results). The anti-infective regimen was adjusted to meropenem plus polymyxin, continuing voriconazole. On January 26, blood culture grew Klebsiella pneumoniae, and cefoperazone-sulbactam was added. The patient’s infection markers then improved significantly: C-reactive protein decreased from 191.28 mg/L to 40.49 mg/L, and procalcitonin from 29.67 ng/mL to 0.15 ng/mL. Repeat chest CT on January 27 showed marked improvement (Fig. 3). Chest CT of the patient on January 27, 2025, showing marked improvement compared to previous imaging. (A) Lung window, (B) Mediastinal window On January 30, sputum culture suggested Elizabethkingia meningoseptica. On February 5, the fourth bronchoalveolar lavage mNGS identified Elizabethkingia anophelis, Klebsiella pneumoniae, Candida glabrata, Acinetobacter baumannii, and Chryseobacterium indologenes (see Supplementary Table 2 for reads). Sulbactam and meropenem were discontinued, and the regimen was adjusted to polymyxin, piperacillin-tazobactam, minocycline, and voriconazole based on published antimicrobial susceptibility data [8]. However, isolate-specific susceptibility testing for the identified E. anophelis strain was not available, as the pathogen was not isolated in pure culture after mNGS diagnosis. Chest X-ray on February 9 showed improvement (Fig. 4). With awake ECMO support, the patient’s infection markers fluctuated but trended downward, and he was successfully weaned from ECMO on February 10.The patient tolerated ECMO and the prolonged anti-infective regimen without major adverse events, and no device-related complications occurred. Chest X-ray (DR) performed at our hospital on February 9, 2025, demonstrating gradual resolution of pulmonary infiltrates following awake ECMO and adjusted antimicrobial therapy After ECMO weaning, the patient entered bedside rehabilitation while continuing mechanical ventilation. On February 15, the fifth bronchoalveolar lavage mNGS revealed Elizabethkingia anophelis, Acinetobacter baumannii, Klebsiella pneumoniae (carbapenem/penicillin resistant), human herpesvirus type 5, rhinovirus A, and Candida glabrata (see Supplementary Table 2 for reads). C-reactive protein rose to 93.94 mg/L and procalcitonin increased. Meropenem, levofloxacin, and sulbactam were added to the regimen. On February 22, the patient developed intermittent fever (maximum 38.3 °C); piperacillin-tazobactam and minocycline were replaced with eravacycline. Thereafter, his temperature normalized and infection markers declined. Respiratory support was gradually reduced from A/C to synchronized intermittent mandatory ventilation (SIMV) and then to pressure support ventilation (PSV). The tracheostomy tube was capped, replaced with a metal cannula on March 6, and removed on March 10 when the patient was afebrile, eating independently, and had stable vital signs (heart rate 80 beats/min, respiratory rate 16 breaths/min, blood pressure 156/72 mmHg, oxygen saturation 99%). Repeat mNGS on February 26 showed decreasing loads of Elizabethkingia anophelis and Acinetobacter baumannii (see Supplementary Table 2). On March 7, the seventh mNGS showed further reduction of Elizabethkingia anophelis and Acinetobacter baumannii (see Supplementary Table 2). All intravenous antibiotics were discontinued on March 7. A chest CT on March 10 demonstrated marked resolution of pulmonary infection (Fig. 5). The patient was transferred back to the Respiratory Medicine department on March 11 and was discharged after recovery on March 21.At follow-up one month after discharge, the patient remained afebrile, had regained full functional independence, and showed no evidence of infection recurrence. Chest CT of the patient on March 10, 2025, showing marked resolution of pulmonary infection. (A, B) Lung window views The timeline of the patient’s clinical course is presented in Fig. 6. Timeline of the patient’s clinical course from admission to discharge, highlighting key diagnostic and therapeutic events including initiation of awake VV-ECMO, identification of Elizabethkingia anophelis, and weaning from ECMO The detailed serial laboratory values of CRP, WBC, and PCT throughout the hospitalization are presented in Supplementary Table 1. Discussion At the time of admission, this patient was initially considered to have lung abscess with pneumonia and chronic hepatitis B. Treatment with biapenem was ineffective. The first bronchoalveolar lavage fluid NGS indicated infection with influenza A virus H1N1, and antiviral therapy was switched to baloxavir marboxil. However, the patient’s condition continued to deteriorate, with sudden onset of type I respiratory failure requiring endotracheal intubation and mechanical ventilation, followed by transfer to the ICU. The second bronchoalveolar lavage fluid NGS showed an increased viral load of influenza A virus along with the emergence of Aspergillus fumigatus complex infection, prompting adjustment to combination antiviral and antifungal therapy with baloxavir marboxil, peramivir, and voriconazole. Nevertheless, the patient’s infection markers remained persistently elevated, and despite lung-protective ventilation, prone positioning, and other measures, oxygenation remained difficult to maintain. VV-ECMO support was initiated on January 23, and “awake ECMO” management was implemented. Subsequently, the patient developed a secondary infection with Elizabethkingia anophelis. After extended anti-infective therapy and comprehensive treatment, the patient was successfully weaned from ECMO and mechanical ventilation, and ultimately discharged. Elizabethkingia meningoseptica has long been recognized as an important pathogen causing neonatal meningitis and sepsis, and has become a significant healthcare-associated pathogen [4]. Notably, prior to the first isolation and identification of Elizabethkingia anophelis in 2011, it was often misidentified as E. meningoseptica due to phenotypic similarity. Subsequent retrospective studies based on stored strains confirmed that many previously reported E. meningoseptica infections were actually caused by E. anophelis. Currently, clinical microbiology laboratories widely employ matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) for rapid identification of bacteria and fungi, with simultaneous detection of certain resistance markers [5]. However, limitations remain in the differentiation of species within the genus Elizabethkingia using molecular methods, including 16 S rRNA sequencing and MALDI-TOF MS [6]. This is largely because some 16 S rRNA sequences annotated as E. meningoseptica in the GenBank database are actually derived from E. anophelis; such historical misannotations may interfere with accurate interpretation of sequencing results. Once the database is supplemented with reference mass spectra for E. anophelis, MALDI-TOF MS can effectively distinguish this species from E. meningoseptica, but it remains challenging to differentiate it from E. miricola. Overall, with the support of an expanded database, MALDI-TOF MS remains the preferred method for rapid diagnosis of E. anophelis infection [7]. From a clinical perspective, Elizabethkingia infections show distinct population characteristics. A study involving 92 patients with Elizabethkingia infection showed that 69.6% were male, with a median age of 61 years; 78.3% had underlying chronic diseases, with malignancies being the most common comorbidity (43.5%), followed by hypertension (28.3%) and diabetes mellitus (26.1%). Up to 80.4% of patients received inappropriate initial empirical therapy, with an overall mortality rate of 27.2% [8]. In addition, this study showed that the susceptibility rate of E. anophelis to minocycline was as high as 100%, to levofloxacin was 58.3%, and susceptibility to other commonly used antibiotics was low [8]. Based on these epidemiological, diagnostic, and antimicrobial susceptibility data, it is recommended that in clinical practice, when sputum culture reports E. meningoseptica, clinicians should be alert to the possibility of actual E. anophelis infection. Before further species identification is completed, empirical use of minocycline alone or in combination with levofloxacin may be considered for anti-infective therapy. Furthermore, a domestic case report described a patient with severe pneumonia co-infected with E. anophelis and Acinetobacter baumannii admitted during the COVID-19 pandemic, who was successfully treated with a minocycline-based regimen (combined with cefoperazone-sulbactam) and discharged [9]. This case further reinforces the central role of minocycline in the treatment of E. anophelis infections and demonstrates the value of mNGS technology for rapid and accurate diagnosis in mixed infections. Therefore, if conditions permit, mNGS or real-time multiplex PCR testing [10] should be performed as soon as possible to clarify the species and provide a basis for precision therapy. Regarding E. anophelis infection in the ICU, the literature reports that it tends to occur in immunocompromised patients, those with indwelling central venous catheters, and those with prolonged hospital stays. It can cause pneumonia, sepsis, and meningitis, with mortality rates for E. meningoseptica bacteremia exceeding 50% [11]. This organism is well adapted to the intensive care environment, capable of forming biofilms and surviving for extended periods in moist environments, including tap water [12]. A potential source is the outlet of water faucets equipped with aerators, with transmission occurring via contaminated hands. Removing aerators can effectively eliminate E. anophelis from the water source; if removal is not feasible, the use of alcohol-based hand sanitizers can eliminate bacteria on healthcare workers’ hands, which is of great importance for infection control [11]. In this case, when E. anophelis was detected by NGS, antimicrobial susceptibility results were not provided. Minocycline was selected based on previously reported antimicrobial susceptibility data for E. anophelis [8], and subsequent fifth to seventh NGS results showed a persistent decline in the bacterial load, indicating that the anti-infective therapy was effective. Of note, in recent years, with the emergence of the COVID-19 pandemic, the role of Elizabethkingia species as secondary or co-infecting pathogens following viral infections has gradually become evident. Reyes-Barros et al. [13] reported seven cases of Elizabethkingia infection in Chile, four of which occurred in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia, marking the first international report of co-infection of this bacterium with COVID-19. The study suggested that although this may be related to the concentration of multiple traditional risk factors in critically ill COVID-19 patients, SARS-CoV-2 infection itself may constitute a new, independent risk factor that disrupts the respiratory mucosal barrier and local immunity, creating conditions for secondary E. anophelis infection. The physiological basis of awake ECMO is that, in spontaneously breathing patients, diaphragmatic movement preferentially occurs in the dorsal regions where compliance is higher, directing ventilation to gravity-dependent areas and achieving optimal ventilation-perfusion matching. In contrast, during sedation and neuromuscular blockade, the diaphragm moves passively, and ventilation is preferentially directed to non-dependent areas, exacerbating ventilation-perfusion mismatch. Thomas Langer et al. [14] systematically elucidated this physiological mechanism in their review and noted that the use of awake VV-ECMO in patients with respiratory failure can effectively avoid the various side effects associated with deep sedation, endotracheal intubation, and mechanical ventilation, including ventilator-associated pneumonia, diaphragmatic dysfunction, ICU-acquired weakness, and increased risk of delirium [14]. To objectively assess the patient’s potential for recovery before initiating ECMO, we calculated several validated scores using data from the day of ECMO cannulation (January 23). The APACHE II score was 27, reflecting critical illness but with most derangements potentially reversible. The PRESERVE score was 4 (after confirming that plateau pressure remained below 30 cmH₂O throughout), predicting a 6-month survival probability of approximately 50–60% for VV-ECMO. The RESP score was 0, corresponding to an estimated in-hospital survival of 55–60%. These scores, together with the patient’s age (62 years), absence of irreversible end-stage organ disease, and lack of severe immunocompromise, indicated an acceptable chance of recovery. Therefore, the multidisciplinary team proceeded with ECMO as a bridge to recovery rather than as a last-resort salvage measure. Detailed score components are provided in the supplementary materials. Based on accumulating clinical evidence in recent years, the role of ECMO in the management of severe ARDS has been clearly recognized in clinical guidelines. Seitaro Fujishima [15], citing multiple guidelines in his review, pointed out that for ARDS patients who still exhibit severe hypoxemia (e.g., PaO2/FiO2 < 80 mmHg) or hypercapnia (pH < 7.25) after optimization of conventional treatment (including prone positioning), VV-ECMO has been weakly recommended as a routine treatment modality. However, the initiation of ECMO is not based solely on gas exchange parameters; it also requires comprehensive assessment of the reversibility of the underlying disease, age, degree of organ dysfunction, and long-term prognosis. The decision-making algorithm proposed by E. Caroline Bullen et al. [16] emphasizes that ECMO should serve as a platform for achieving lung-protective ventilation, and the key to success lies in selecting patients with potential for recovery. They also pointed out that when ECMO is unable to facilitate patient recovery or transplantation, the patient’s awake state may present complex ethical and emotional challenges when it comes to ECMO withdrawal [16]. The use of a low PEEP (3 cmH₂O) in this patient during the early phase of ARDS was not a deviation from standard care but a deliberate, physiology-driven adjustment. Although the ARDSnet low-tidal-volume strategy recommends higher PEEP for moderate-to-severe ARDS, this patient had severe mitral stenosis from rheumatic heart disease, resulting in preload-dependent cardiac output. Higher intrathoracic pressure from increased PEEP could further impair right ventricular filling and reduce left ventricular preload, potentially precipitating cardiogenic shock. Therefore, initial lung-protective ventilation was attempted with a low PEEP (3–5 cmH₂O) while maintaining a driving pressure < 15 cmH₂O. After VV-ECMO was established and hemodynamics stabilized, PEEP was titrated daily based on respiratory system compliance and driving pressure, reaching 10 cmH₂O on January 24, in conjunction with prone positioning (daily > 16 h). This case illustrates that PEEP selection in ARDS must be individualized, and that severe valvular heart disease may warrant a lower initial PEEP, with subsequent escalation only after circulatory support is secured. In the present case, ECMO was used not merely as a last-resort salvage therapy but as an active treatment platform. Early implementation of awake ECMO allowed the patient to maintain spontaneous breathing while awake, avoiding complications such as ventilator-associated pneumonia and acquired muscle weakness associated with deep sedation and neuromuscular blockade, preserving the cough reflex and airway self-cleaning capacity, and laying the foundation for prolonged treatment. This finding precisely illustrates the limitations of conventional testing methods in distinguishing between species within the genus Elizabethkingia. Notably, mNGS correctly identified E. anophelis at the species level, overcoming the phenotypic ambiguity that commonly leads to misidentification by conventional methods. Awake VV-ECMO refers to the application of extracorporeal support in non-intubated patients who remain conscious and breathe spontaneously [14, 17]. This strategy avoids deep sedation and invasive mechanical ventilation, thereby reducing the risks of ventilator-induced lung injury, ventilator-associated pneumonia, diaphragm dysfunction, and ICU-acquired weakness [14, 17, 18]. Additional advantages include preserved cough reflex, improved ventilation-perfusion matching due to active diaphragmatic movement, and the possibility of early mobilization and physiotherapy [17, 19]. Awake ECMO has been used in several clinical scenarios: as a bridge to lung transplantation in patients with end-stage respiratory disease, in ARDS patients at high risk of barotrauma (e.g., those with Macklin effect on CT), in COPD exacerbations failing noninvasive ventilation, and in patients who refuse endotracheal intubation [17, 18, 20]. However, awake ECMO carries specific risks, including patient self-inflicted lung injury (P-SILI) from excessive respiratory effort, delirium, accidental cannula displacement or self-removal, bleeding, and circuit thrombosis [17, 19, 21]. Therefore, careful patient selection, continuous monitoring of respiratory effort (e.g., respiratory rate, use of accessory muscles), and a multidisciplinary approach are essential for safe implementation [17, 18]. During the awake ECMO period (January 23 to February 10), the patient was maintained with a Richmond Agitation-Sedation Scale (RASS) score of -1 to 0 to balance comfort and cooperation. Under the supervision of ICU nurses and physiotherapists, he received daily passive and active range-of-motion exercises of the limbs, as well as respiratory muscle training including directed coughing and deep breathing exercises. However, the patient had markedly elevated D-dimer levels (up to 24,738 ng/mL on January 30) and lower-limb ultrasonography revealed bilateral gastrocnemius deep vein thrombosis. Consequently, bed rest with lower-limb immobilization and therapeutic anticoagulation (heparin followed by argatroban) was prescribed to prevent thromboembolic complications. Standing or ambulation was therefore not attempted during the ECMO run. After ECMO weaning and stabilization of coagulation parameters, the patient gradually resumed bed-based exercises and sitting, and subsequently achieved assisted standing and short-distance walking before discharge. The outcomes of our patient compare favorably with those reported in recent literature. In a scoping review by Sklienka et al. encompassing 64 papers and 95 patients with COVID-19-related ARDS managed with awake ECMO, the success rate (defined as discharge without intubation) was 36.4% among cohort studies, with an overall mortality of 33.0% [18]. Our patient was successfully weaned from ECMO and mechanical ventilation and discharged alive, representing a successful outcome. Notably, in that review, patients who ultimately required intubation during awake ECMO had a significantly higher mortality (49.1%) compared with those who remained non-intubated (8.6%) [18]. Our patient did not require intubation after ECMO initiation, which aligns with the better prognosis group. Furthermore, Paternoster et al. reported a series of awake ECMO patients with ARDS at high risk for barotrauma, achieving a survival rate of 71.4% and no progression of barotrauma [17]. Our patient had a PEEP of 3 cmH₂O and an oxygenation index of 62 mmHg before ECMO, and no barotrauma occurred during the awake ECMO run. In a case report by Zewude et al., an immunocompromised patient developed E. anophelis ventilator-associated pneumonia and bacteremia from contaminated holy water; despite targeted antibiotics, the patient eventually died [22]. Our patient survived after minocycline-based therapy, highlighting the importance of accurate species identification and timely adjustment of antibiotics. Maraki et al. reviewed pediatric E. anophelis infections and found a case fatality rate of 33.3%; our adult patient’s recovery adds to the evidence that minocycline-susceptible strains can be treated successfully [23]. Unlike the case reported by Larkin et al., where E. anophelis was initially misidentified as E. meningoseptica by MALDI-TOF and corrected by whole-genome sequencing, our case used mNGS to achieve accurate speciation and guide therapy, resulting in a favorable outcome [24]. To our knowledge, this is the first report describing the combination of awake VV-ECMO and dynamic mNGS-guided therapy in a patient with severe polymicrobial pneumonia secondary to Elizabethkingia anophelis infection. The key novel aspects are: (1) successful application of awake ECMO as a bridging platform for prolonged anti-infective therapy in a patient with this rare multidrug-resistant pathogen; (2) demonstration that mNGS can correct species misidentification caused by conventional culture (which reported E. meningoseptica), enabling timely minocycline-based therapy; and (3) serial mNGS monitoring to track pathogen load dynamics and guide antibiotic adjustments. These features distinguish this case from previous reports of E. anophelis infection or awake ECMO alone. To quantitatively assess the patient’s illness severity and respond to the reviewer’s request, we calculated sequential organ failure assessment (SOFA) scores at key time points (Supplementary Table 3). The SOFA score increased from 0 on admission (T1) to 7 on January 13 (T2) and to 13 on January 23 just before ECMO initiation (T3). This acute rise of ≥ 2 points, together with confirmed infection, fulfills the Sepsis-3 criteria for sepsis. On January 23, the patient had a lactate level of 3.1 mmol/L (normal < 2 mmol/L) and required norepinephrine at 0.61 µg/kg/min to maintain mean arterial pressure ≥ 65 mmHg, meeting the definition of septic shock. All detected pathogens – influenza A H1N1, Aspergillus fumigatus, Acinetobacter baumannii, Klebsiella pneumoniae, and Elizabethkingia anophelis – were temporally associated with worsening clinical status (fever, declining oxygenation, rising inflammatory markers). Furthermore, targeted antibiotic adjustment (particularly minocycline for E. anophelis) led to sustained reduction of pathogen loads on serial mNGS, improvement of CRP/PCT, and eventual clinical recovery. Therefore, these organisms were clinically significant pathogens rather than incidental colonizers or contaminants. The key to successful management in this case lay in awake ECMO support and the treatment of Elizabethkingia anophelis infection. Awake ECMO enabled the patient to tolerate a complex anti-infective regimen while facilitating adjunctive therapies such as prone positioning and early rehabilitation. Meanwhile, continuous dynamic mNGS monitoring promptly revealed the presence of multidrug-resistant and rare pathogens, with Elizabethkingia anophelis as a central component, providing a basis for multiple precise adjustments to the anti-infective regimen. The combination of these two strategies offers a replicable clinical pathway for managing such complex critically ill cases. However, this case also exposed several shortcomings. First, after the initial sputum culture reported Elizabethkingia meningoseptica, no immediate confirmatory testing was performed. Had subsequent mNGS not corrected the identification, the misdiagnosis of E. anophelis would have directly led to a delay in initiating minocycline therapy. Second, after mNGS identified E. anophelis, no isolate-specific antimicrobial susceptibility testing was performed because the organism was not isolated in pure culture. Consequently, the anti-infective regimen (minocycline-based combination) was selected based on published susceptibility data [8] rather than on patient-specific testing. This highlights the urgent need to establish standardized methods for isolating and testing E. anophelis to guide precision therapy. Third, after the patient developed E. anophelis infection following ICU admission, environmental source tracing was not initiated. In the future, a mechanism for prompt investigation upon infection detection should be established, with immediate screening of potential reservoirs such as water faucet aerators. Conclusion This case demonstrates that awake ECMO can serve as an effective respiratory support platform in complex severe pneumonia, facilitating prolonged anti-infective therapy and early rehabilitation. When conventional testing reports Elizabethkingia meningoseptica, clinicians should suspect possible Elizabethkingia anophelis infection, and timely mNGS is recommended for accurate species identification. If further testing to confirm the species is not possible, empirical anti-infective therapy with minocycline alone or in combination with levofloxacin may be considered. If the response is poor, prompt use of mNGS or real-time multiplex PCR should be pursued to identify the specific pathogen and guide therapy. In parallel, strengthening environmental infection control measures - such as screening for colonization sources like water faucet aerators - is essential to interrupt nosocomial transmission. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Author contributions WG contributed to data collection, literature review, and manuscript writing. KX and HR contributed to patient management, critical revision of the manuscript, and final approval of the version to be published. All authors read and approved the final manuscript. Funding This study received no specific funding. Data availability All data generated or analysed during this study are included in this published article. The datasets are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This retrospective case report did not require ethical approval according to the regulations of The Affiliated Lianyungang Hospital of Xuzhou Medical University, and was conducted in accordance with the Declaration of Helsinki, as it involved no experimental interventions and only routine clinical data. Consent to participate Not applicable. This retrospective case report involved no experimental interventions and only routine clinical data. Consent for publication Written informed consent for publication of this case report and any accompanying images was obtained from the patient. A copy of the signed consent form is available for review by the Editor-in-Chief of this journal. Patient perspective The patient did not provide a personal perspective on his treatment experience. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Kang Xiuwen, Email: mafei8922@163.com. Hu Rong, Email: huronglygyy2018@163.com. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All data generated or analysed during this study are included in this published article. The datasets are available from the corresponding author on reasonable request.