A structured diagnostic algorithm for patients with ARDS

Patients admitted to the intensive care unit (ICU) with acute respiratory failure frequently fulfil the criteria for acute respiratory distress syndrome (ARDS) [1]. The diagnosis is based on radiological, physiological, and clinical criteria described in the ‘Berlin definition’ (Table 1) [2]. Yet establishing the diagnosis of ARDS has limited treatment consequences in and of itself, as the available evidence-based interventions are mainly related to minimizing iatrogenic damage (e.g., ventilator-induced lung injury [VILI] and fluid overload) rather than the use of specific treat-ments. Whereas the intervention options for the syndrome itself are limited, adequate and timely treatment of the causal underlying condition has a major impact on the improvement of outcomes for patients with ARDS [3].

The classical description of ARDS relies on the histological finding of diffuse alveolar damage secondary to another condition (one of the clinical risk factors described in Table 1) [4]. Diffuse alveolar damage is an untreatable finding and must be distinguished from a large number of diseases that also meet the ARDS syndrome definition but are treatable [5]. Table 2 provides an overview of the differential diagnoses that must be taken into account in patients suspected of having ARDS.

It should be possible to establish a definitive causal diagnosis within 7 days after onset in the vast majority of patients with ARDS. Yet, the often chaotic nature of clinical reality can lead to a delayed and haphazard search for underlying causes, especially in patients with multiple important problems.

This narrative review aims to provide a structured approach to the diagnosis of underlying conditions in patients who fulfil the ARDS criteria according to the Berlin definition, in order to enable underlying causes to be rapidly and adequately treated. The diagnostic steps are described point by point in three phases and are summarized in a flowchart (Fig. 1). We also provide a summary of the most important uncertainties relevant to clinicians managing patients with ARDS.

With the steps and timeframe described here, we have intended to strike a balance between early and vigorous diagnostic investigations where needed and a more parsimonious approach where appropriate. Nevertheless, the authors’ experience is one rooted in academic medicine in a resource-rich environment. The details of our approach should therefore be adapted to local resources and possibilities. The most important aspect of the here described approach is not the number of laboratory investigations or imaging modalities, but rather the structuredness and timeliness of the diagnostic evaluation.

The diagnosis of ARDS is largely based on hypoxic respiratory failure and the detection of pulmonary edema, from which hydrostatic cardiogenic pulmonary edema must be excluded [6]. It is therefore essential that false positive results are excluded as much as possible by means of non-invasive imaging. Ultrasound of the lungs is superior to chest X-ray for detecting and ruling out pleural effusion [7, 8]. The chest X-ray can give the illusion that the lung is consolidated when there is a large effusion and this can result in a false positive diagnosis, but the treatment differs. Transthoracic ultrasound of the heart can facilitate the diagnosis of acute heart failure. In general, acute heart failure is sufficient explanation for pulmonary edema and treatment should be focused on decongestion. However, a complicating factor is that cardiomyopathy due to a hyperinflammatory state can be associated with ARDS (and thus non-cardiogenic pulmonary edema) [9]. Usually, these are patients with sepsis, with a high a priori risk of ARDS who should also be diagnosed with ARDS.

Extra-pulmonary and pulmonary risk factors for ARDS need to be identified as soon as possible. When an evident extra-pulmonary cause for ARDS, such as septic shock, is present, timely treatment of the underlying cause determines the patient’s prognosis [15]. Patients with community-acquired pneumonia and ARDS are indis- tinguishable from patients in the ICU with community-acquired pneumonia without ARDS in terms of epidemiological data, microbiological results, and outcome, and likely require the same treatment [16]. As opportunistic infections can present with specific radiological patterns that cannot be appreciated on chest X-ray, a chest computed tomography (CT) scan should be performed in patients with increased a priori risk for such infections (see practical steps) [17]. If no risk factor can be iden- tified, there is an increased probability of an underlying systemic disease or drug related cause [5]. Further information for this should be obtained through history, physical examination, autoimmune serology and chest CT. The most important serological tests are: extractable nuclear antigens (ENA), anti-nuclear antibodies (ANA), anti-neutrophil cytoplasmic antibodies (ANCA), myositis blot, anti-cyclic citrulline peptide antibody (aCCP) and rheumatoid factor (RF). When hemoptysis and/or acute renal insufficiency with (microscopic) hematuria are present, anti- glomerular basal membrane (aGBM) levels should also be obtained [18].

A considerable proportion of patients will improve during the first phase of evalua- tion and in other patients it will be possible to establish a definitive causal diagnosis [25]. If after 2 days no causative agent of pneumonia has been demonstrated, and if the clinical condition of the patient does not improve, it is important to reconsider the diagnosis. To not miss any mimicking conditions, the same approach is followed as in patients without a risk factor at presentation including but not limited to a detailed history, physical examination, and autoimmune serology.

Performing a chest CT scan can help distinguish between opportunistic infections and interstitial lung diseases [17, 18, 26]. A bronchoscopy with lavage is a reliable method to take microbiological cultures from the lower respiratory tract [18], espe- cially if PCR analyses are performed for viruses and opportunistic pathogens such as Pneumocystis and Aspergillus. Bronchoscopy can result in loss of pressure from the ventilation system and thus to collapse of previously opened parts of the lung. However, multiple studies suggest that bronchoscopy with lavage is safe in intubated patients with ARDS provided that it follows the prevailing guidelines [27‐29].

If ARDS persists for more than 5 days after diagnosis it is considered as ‘non-resolving’ and the risk of fibrosis formation is considerable if the cause for ARDS remains untreated [3, 30]. There are several studies that show that reactivation of HSV and CMV is frequent in non-resolving ARDS [31‐33]. No randomized trials have been conducted on whether or not to treat HSV reactivations, but observational studies suggest independent excess mortality in the HSV and CMV group. Due to the lack of other treatable conditions in this patient group, it is therefore advisable to treat reactivation [18].

There is conflicting evidence from several randomized trials of corticosteroid treatment for non-resolving ARDS. There appears to be a positive effect on ventilation duration and possibly on mortality if treatment is started early in the non-resolving phase, e.g., before day 14 [34‐36]. Apart from hyperglycemias, relatively few adverse reactions have been described [37].

In patients in whom ARDS persists and in whom the underlying cause has not been confirmed despite all the above steps, a lung biopsy should be considered. The risks of an open lung biopsy (< 10% serious complications, < 1% lethal complications [38]) must be weighed against the substantial burden of ongoing ICU treatment without a clear diagnosis. Open lung biopsy provided a specific diagnosis in about 75% of cases and led to an adjustment in medication in about one in three patients [38]. In addition, findings other than diffuse alveolar damage are associated with change in therapy and better outcome [32, 39‐42]. Prolonged ventilation in the context of non-resolving ARDS without a diagnosis has a very poor prognosis and can result in unnecessarily prolonged ICU treatment with all the adverse consequences that this entails. All in all, this results in the recommendation to consider an open lung biopsy in all patients with non-resolving ARDS without a proven risk factor [18]. Traditionally, a surgical open lung biopsy is obtained by thoracoscopy, but there are new developments in the field of bronchoscopically obtained cryobiopsies that can be considered as an alternative [43].

Establishing the underlying cause of ARDS is of great importance as adequate treat- ment of this cause improves outcome. The proposed structured diagnostic algorithm helps clinicians to systematically evaluate patients with ARDS and to decrease time to diagnosis and thereby start of adequate treatment.