1 Introduction

The COVID-19 pandemic has served as a painful reminder of the limited treatments available for acute respiratory distress syndrome (ARDS). Apart from lung protective tidal volumes and prone positioning, no other interventions or therapies have been shown to consistently improve mortality in clinical trials [1, 2]. The complex pathophysiology of ARDS has offered many potential targets for intervention but, similar to sepsis, the various clinical phenotypes are often not discernable in real time, making individualized treatment difficult. Researchers previously described the complex inflammatory, endothelial, and coagulopathic dysfunction that is shared by sepsis and ARDS [3], with Chang further describing sepsis and ARDS as two distinct phenotypes of this shared pathologic cascade. [4]

The overwhelming burden of illness from COVID19 has provided an opportunity to identify different clinical phenotypes and highlights the need to investigate therapies and interventions, both new and old.

Most of these treatments have focused on the pathologic inflammation and coagulopathy demonstrated in sepsis and ARDS. While some of these therapies such as corticosteroids and biologics have shown variable success [5, 6], the morbidity and mortality have remained significant and other treatments continue to be explored.

The emphasis to abate the “cytokine storm” associated with COVID-19 led to renewed interest in previously debunked therapies including therapeutic plasma exchange and high dose corticosteroids [7, 8].

Management of respiratory failure was also called into question, and treatments targeting the respiratory system were again considered. Surfactant, a liquid naturally produced by the lungs, serves to decrease surface tension and maintain lung compliance, and is known to be deficient in ARDS [9]. Similar to steroids, prior outcome data on the efficacy of surfactant replacement in ARDS has been inconsistent in adults with no clinical trial showing benefit.

Additionally, the best approach to “lung protection” was questioned. Small tidal volume, lung protective ventilation has been established as the standard of care for patients with ARDS, and high frequency oscillatory ventilation (HFOV) offers superior benefit in theory based on exceptionally small tidal volumes. While previous studies had previously failed to show benefit [10, 11], our interest in this mode of ventilator support was renewed as the mortality of COVID-19 remained high [12].

The complexity and variability of ARDS, makes a standardized, “one size fits all” treatment protocol unlikely to significantly further reduce mortality beyond ARDSnet ventilation strategy. No two patients are exactly the same, regardless of etiology, and identifying clinical phenotypes and stages of ARDS is important when investigating potential interventions. We postulate that a multimodal approach in the acute inflammatory, recruitable stage of ARDS may be superior to any single intervention alone. We present a case of refractory ARDS that improved following a combination of therapies targeting multiple pathways and discuss the rationale for each therapy, emphasizing the importance of future investigations to improve outcomes in this common, deadly condition.

2 Case Description

We present a case of ARDS secondary to COVID-19 pneumonia that demonstrated significant clinical improvement with a combination of high dose steroids, therapeutic plasma exchange, surfactant replacement, and HFOV support. Written informed consent was obtained from the patient for their anonymized information to be published, and the case report received exemption form the LMC IRB. A 51 year old with a past medical history of diabetes mellitus (DM), Stage III chronic kidney disease (CKD), and bipolar disorder presented with dyspnea and dry cough. He was found to be COVID positive requiring supplemental oxygen and was admitted to the hospital ward, treated with dexamethasone and baricitinib. He quickly deteriorated with worsening hypoxemia and was transferred to the intensive care unit (ICU) where he was intubated, placed in prone position, and started on inhaled nitric oxide (Fig. 1). He developed sepsis with multiple organ dysfunction (MODS), and required vasopressor support and continuous renal replacement therapy (CRRT). Given his continued clinical decline he underwent therapeutic plasma exchange (TPE) on three consecutive days (based on the 2023 ASFA guidelines for sepsis with multiorgan failure, Category III, 2A) [13]. Despite these interventions his respiratory status continued to decline. His case was discussed with a local academic center, but he was determined not to be a candidate for extracorporeal membrane oxygenation (ECMO) based on his prolonged illness and his shock state with multiple organ failure. On day 15 of intubation he was placed on high-frequency oscillatory ventilation (HFOV) and was started on high dose methylprednisolone, empiric antibiotics, and another round of TPE. Additionally, he was treated with direct instillation of surfactant into the tertiary bronchioles bilaterally via flexible bronchoscopy on three consecutive days. He showed a favorable response to these treatments and was transitioned to conventional ventilation on hospital day 20 (Fig. 2). He underwent bedside tracheostomy on hospital day 22, continued to improve, and was de-cannulated on hospital day 29. His renal function recovered to pre-admission baseline, and he no longer required RRT. He was weaned to low flow oxygen and discharged to inpatient rehab. He completed rehab and returned home, now totally independent and on 1L NC.

Fig. 1
figure 1

Admission and post-intubation chest x-rays. A Admission CXR. B Post-intubation CXR

Full size image
Fig. 2
figure 2

X-rays during clinical course. A Pre-HFOV. B Post-HFOV. C Following treatment, on conventional vent

Full size image

3 Discussion

We believe that our patient would not have survived if we had continued to follow only standard protocols for treatment of sepsis and ARDS, and that the application of lesser-used, but physiologically and pharmacologically rational treatments may have contributed to his recovery. With this case as an example, we review the shared pathophysiology of sepsis and ARDS while discussing previously debunked therapies addressing these pathways, including high dose corticosteroids, therapeutic plasma exchange, liquid surfactant, and high flow oscillatory ventilation.

The definition of ARDS is based solely on the acuity of illness, chest imaging, and the degree of hypoxemia while not factoring in the etiology, except that it is not cardiac in nature [14]. Classically ARDS has been defined by three relatively discrete stages—exudative, proliferative, and fibrotic. Treatments have focused on preventing progression to the later stages, but clinical trials in ARDS do not define the stage of ARDS, rather the severity of hypoxemia based on PaO2/FiO2 ratio (ratio of partial pressure of arterial oxygen to fraction of inspired oxygen). The mortality of ARDS has remained largely unchanged since the landmark ARMA trial in 2000, despite ARDS remaining one of the most common conditions necessitating ICU admission. Low tidal volume, “lung-protective” ventilation strategies have proven most beneficial, and application of these this strategy has led to the reemergence of previously debunked interventions including prone ventilation and ECMO. Apart from abating the pathologic cascade at the alveolar level, as described by Fujimara [15], none of these interventions directly address the pathophysiology of ARDS. While not often a point of emphasis, it is extremely important to note that different clinical phenotypes exist and that various treatments and interventions may be beneficial in some cases but not in others. The reader is referred to the paper by Chang which reviews the complex inflammatory, endothelial, and coagulopathic dysfunction that is shared by sepsis and ARDS [7], with Chang further describing sepsis and ARDS as two distinct phenotypes of this shared pathologic cascade [3]. Additionally, clinicians and researchers have begun to identify distinct phenotypes in children with sepsis [16], and clinical trials are being designed to investigate treatments for each of these phenotypes in specific and different manners. As it relates to ARDS, much debate has taken place about the pathophysiology as it relates to COVID-19. At least two distinct phenotypes have been described, the L-type with relatively preserved lung compliance (sometimes described as “happy hypoxia”) and the H-type, which presents like “classic” ARDS with poor lung compliance [17], it is unlikely that these phenotypes are unique to COVID-19, and is also unlikely that these are the only two phenotypes of ARDS. Further work to identify phenotypes is necessary in order to identify potential individualized treatments.

Our patient failed standard COVID directed therapies and ARDS guidelines, developing hemodynamic instability and progressing to multiple failure. We were unable to adequately oxygenate or ventilate with ARDSnet settings despite paralysis, prone ventilation, and inhaled nitric oxide. His case was discussed with a local ECMO referral center, but for the reasons noted above he was not felt to be a candidate for ECMO support. While severely hypoxemic, he was able to improve oxygenation with high PEEP despite extremely poor compliance. Given this response, we maintained guarded optimism that he remained in a potentially reversible phase of illness and adjusted treatment.

He was transitioned to HFOV in order to maintain adequate mean airway pressures for gas exchange while limiting injurious distending alveolar volumes. Despite falling out of clinical use, this modality of support physiologically makes sense, and our group has had favorable outcomes with its use during the pandemic [12]. His ability to ventilate (with improving CO2 levels), oxygenate at lower FiO2 levels, and his improved aeration on chest x-ray all offered encouragement (Fig. 2, Table 1).

Table 1 Arterial blood gases and ventilator settings
Full size table

In addition, steroids were transitioned to weight-based methylprednisolone derived from prior studies by Meduri in early, and/or persistent ARDS [18, 19]. While both these studies were limited, both showed improvements in certain predefined outcomes (mortality and/or ventilator free days) if initiated within the first 14 days of illness. Our patient’s response to HFOV and his inflammatory profile suggested a possible benefit, despite being just outside the two week window reported by Meduri.

Pathologically, patients with ARDS are noted to have a reduction in surfactant content and activity [9], similar to preterm infants with respiratory distress syndrome (RDS) who have shown favorable responses to exogenous surfactant replacement [20]. While trials in adults with ARDS have failed to show a similar clinical response, in hindsight these trials have been filled with flaws, most notably the volume of replacement and the means of administration. Piva et al. published results of a case–control, retrospective pilot study on surfactant therapy for COVID-19 related ARDS which demonstrated feasibility, safety, and possible clinical benefit [21]. Given our patient’s condition, we chose to instill at total of 720 mg of liquid surfactant directly into third generation bronchi, a procedure adapted from Piva, et al.

Finally, with progression of both ARDS and sepsis with MODS, therapeutic plasma exchange was initiated in attempt to abate the shared pathology. Our group has had favorable outcomes in sepsis driven by COVID-19 and non-COVID infection 22, 23, particularly sepsis driven by pneumonia and our patient met the clinical profile and criteria suggesting a possible favorable response.

With this combination of interventions and treatments, our patient experienced a clinical benefit that would not have been expected based on prior studies of ARDS. The timeline and static ventilator measurements would have suggested progression to a non-recoverable fibrotic lung disease. While it is impossible to determine which therapy, intervention, or combination (if any) led to his recovery, this case serves as example of multimodal approach to treatment while emphasizing the need to identify specific, clinically distinct phenotypes that are likely be responsive to specific interventions rather that “one size fits all” protocols.

4 Conclusion

Acute respiratory distress syndrome represents a syndrome of lung failure, but does not define the exact pathophysiology of each individual patient. Similar to sepsis, multiple phases and distinct phenotypes likely exist. Research is necessary to help identify these distinct phenotypes so that clinical trials can be designed to investigate interventions and treatments specific to the different phenotypes. It is likely that prior “failed” treatments might play a role in select patients, as demonstrated in our case report.