Interstitial pneumonia with autoimmune features: an additional risk factor for ARDS?

We retrospectively reviewed the electronic files of patients with a diagnosis of ARDS (according to the Berlin definition criteria) [14] admitted to the tertiary level ICUs of San Gerardo Hospital in Monza and of Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico in Milan (Italy) from May 2012 to October 2016. Based on the presence or absence of a recognized risk factor for ARDS among those listed in the Berlin definition criteria [14], patients were subdivided into two groups: (1) patients with a known risk factor; (2) patients without a known risk factor. The latter group was further subdivided into two subgroups according to the presence of diagnostic criteria for IPAF [13]: a. patients with IPAF (study group); b. patients without both a recognized risk factor and not fulfilling the criteria for IPAF.

The following data at ICU admission were recorded: demographics (sex, age), comorbidities, severity scores (SAPSII and SOFA), risk factors for ARDS, PaO₂/FiO₂ (P/F) ratio, intrapulmonary shunt (Qs/Qt), ventilator and respiratory mechanics parameters [positive end-expiratory pressure (PEEP), tidal volume (Vt), plateau pressure (P _plat), driving pressure (ΔP), respiratory rate (RR), static respiratory system compliance (C _rs)].

P _plat was measured during an end-inspiratory pause of at least 2 s ; ΔP was defined as P _plat − PEEP_tot; static C _rs was calculated as Vt/ΔP.

The following outcome data were collected: use of prone positioning, need for extracorporeal membrane oxygenation (ECMO) support, duration of invasive mechanical ventilation (IMV) and ECMO (if applicable), ICU length of stay (ICU-LOS) and ICU mortality.

IPAF was diagnosed according to the criteria described by the joint ATS–ERS task force (Table S1) [13].

In all patients with IPAF (study group) and with ARDS without known risk factors, computed tomography (CT) of the chest and bronchoalveolar lavage (BAL) were performed within 2 days from ICU admission, and their findings were reviewed according to the ATS Guidelines [15, 16]. Timing, dose and schedule of administration of steroid therapy and other eventual immunosuppressive treatments were recorded. Episodes of ventilator-associated pneumonia (VAP) occurring in IPAF patients were analyzed. Finally, in patients with IPAF we obtained long-term follow-up data on lung morphology (by CT scan), respiratory function (by spirometry) and immunological status.

Seven patients (four males and three females) with ARF due to a newly diagnosed ILD meeting the diagnostic criteria for IPAF were identified. Median age was 61 years (IQR 55–64); median SAPS II and SOFA scores at admission were 32 (32–36) and 11 (11–12.5), respectively. Five patients (71%) survived: Their median ICU stay was 54 (23–78) days and they were all discharged alive from the hospital. Two patients died 72 and 23 days after ICU admission, respectively (Table 1).

All patients were admitted to the hospital for acute dyspnoea and hypoxemia. Extrathoracic symptoms were few and nonspecific (fatigue, weight loss, fever), and none had skin lesions or arthritis. No patient had previous history of ILD and/or CTD, and detailed medication and occupational history were negative.

According to the oxygenation criteria established by the Berlin definition [14], three patients had moderate and four had severe acute respiratory distress syndrome (ARDS); however, in none of them a “known clinical insult” [14] could be identified as the cause of ARDS. At admission, median P/F ratio was 120 mmHg (range 80–163) and median respiratory system compliance 26 ml/cmH₂O. All patients required intubation and invasive mechanical ventilation (MV). At first day of MV, median positive end-expiratory pressure (PEEP) was 12 cmH₂O, plateau airway pressure 26 cmH₂O and tidal volume 380 mL (Table 1). The time course of selected parameters during the ICU stay is depicted in Additional file 1: Figure S1. The following rescue therapies were applied: recruitment maneuvers (all patients), inhaled nitric oxide (three patients) and prone positioning (four patients). The median duration of MV was 49 days (21–81) and six subjects (85%) were tracheostomized. Five patients (71%) required an extracorporeal respiratory support with veno-venous extracorporeal membrane oxygenation (ECMO) in four cases and with low-flow extracorporeal CO₂ removal (ECCO₂R) with a dedicated device (Prolung™, Estor) in one case. Three of these patients survived and the median ECMO duration was 53 days (17–63).

All patients needed vasopressor support during the ICU stay and two presented with shock at admission. One patient required continuous renal replacement therapy.

Bacterial, viral or fungal infections were excluded in all patients by means of a complete microbiological workup, consisting of cultural, serological and molecular biology tests on blood, urine and BAL samples.

Other potentially reversible causes, such as pneumothorax, pulmonary embolism, left heart failure, acute eosinophilic pneumonia and hypersensitivity pneumonia, were ruled out.

Findings of BAL samples and CT scans are detailed in Table 2. Briefly, cytomorphological analysis of BAL showed mixed alveolitis with significant lymphocytosis (>30%) in three cases and significant neutrophilia (>50%) in two other patients, without signs of viral inclusion or diffuse alveolar hemorrhage.

At CT scans, consolidations were present in all cases and mainly found in subpleural caudal areas, often associated with perilobular opacities (Fig. 2a; Additional file 1: Figure S2). Ground-glass opacities, in some cases with subpleural sparing, were also quite common but more diffuse and predominant at follow-up scans (Fig. 2b; Additional file 1: Figure S2). In four patients, mediastinal or hilar lymphadenopathies (with a diameter greater than 10 mm) were observed, while pleural effusion was present in only one case. Of the five patients who survived, two had limited or widespread organizing pneumonia (OP) consolidations; one had typical features of AIP/DAD with ground-glass opacities, patchy consolidations and fibrotic changes; one had extensive ground-glass opacities in lower lobes with some cysts suggestive of lymphocytic interstitial pneumonia (LIP) but with atypical extensive superimposition of “crazy-paving” pattern; the last one had and OP/nonspecific interstitial pneumonia (NSIP) overlap pattern. Both patients who died had an OP pattern at presentation, but in one of them the initial OP pattern rapidly progressed to acute interstitial pneumonia/diffuse alveolar damage (AIP/DAD) pattern and he subsequently developed striking extensive fibrotic parenchymal involvement. Follow-up scans were obtained in the first four surviving patients after a median interval from hospital discharge of 20 months. Patients with OP pattern at presentation developed ground-glass opacities both in new parenchymal territories and in previously consolidated areas. Three patients developed mild to moderate signs of fibrosis (traction bronchiectasis and mild parenchymal distortion), but no patient developed honeycombing. Representative images from baseline and follow-up CT scans of the first five patients are presented in Additional file 1: Figure S2.

A complete autoantibody panel was obtained, consisting of antinuclear antibodies (ANA), anti-SSA/Ro52, anti-SSB/LA, anti-ribonucleoprotein, anti-Scl70, anti-Smith, anti-Jo1, anti-centromere, anti-double-stranded DNA, anti-neutrophil cytoplasmic antibody (ANCA), cyclic anti-citrullinated peptide and rheumatoid factor (RF). All patients had anti-SSA/Ro52; one patient had high titer ANA and anti-centromere autoantibodies and one was positive for anti-Jo1. Since no other causes of ILD were identified, extrathoracic symptoms were nonspecific and pulmonary manifestations were by far the most important, all patients met the criteria for IPAF.

All patients received immunosuppressive therapy while in ICU, after a median interval of 10 days from admission. Five of them received high doses of methylprednisolone (1000 mg/day for 3 days and then 1 mg/kg/day), followed by cyclophosphamide (10–15 mg/kg every 2–3 weeks); they also received iv immunoglobulin infusion, and two were treated with cyclosporine for inadequate response to cyclophosphamide. The remaining two patients were treated with methylprednisolone (1 mg/kg/day) without loading dose.

Five patients (71%) developed a ventilator-associated pneumonia, after a median time interval of 19 (10–31) days from admission: Pneumonia was caused by multidrug-resistant Gram-negative bacteria in three cases and by Aspergillus fumigatus in one case.

The median follow-up of the first four surviving patients was 21 months. At the last visit, forced vital capacity range was 77–112% and diffusing lung capacity for carbon monoxide 39–95% of predicted and no patient requires oxygen supplementation. Immunological follow-up examinations showed that anti-SSA/Ro52 became negative in three out of four patients. Maintenance immunosuppressive therapy with low-dose prednisone, azathioprine and mycophenolate is ongoing in three patients.

All IPAF patients, except from the two who died, gave their written informed consent to publication. Due to the retrospective nature of the study and since written consent was obtained from the study patients, ethics committee approval was waived according to local regulations.

Table 3 shows the comparison of baseline characteristics and clinical outcome among the three patient groups. We did not find significant differences in regard to demographics, severity of hypoxemia (P/F ratio and Qs/Qt) and ventilator and respiratory mechanics parameters. IPAF patients tended to have a higher median SOFA score compared to both control groups (11 vs 8 vs 8.5; p = 0.068) and a higher frequency of use of prone positioning (86 vs 53 vs 43%; p = 0.197) and ECMO (71 vs 41 vs 25%; p = 0.176). ICU survival of IPAF patients was exactly equal to that of patients with a known risk factor (71%), while it was markedly higher than that of the patients without risk factors (38%); however, due to the small number of patients, this difference was not statistically significant. Compared to the other subgroups, IPAF patients had significantly longer median ICU-LOS (54 vs 19 vs 15.5 days; p = 0.0045) and duration of IMV (49 vs 17 vs 15.5 days; p = 0.031) and ECMO (53 vs 9.5 vs 28; p = 0.006).

In all the eight control patients without risk factors, microbiological and immunological screening resulted negative. Revision of chest CT scans and BAL performed in these patients did not allow the identification of a typical radiologic or cytomorphologic pattern, similarly to what observed IPAF patients. Of note, in all these patients a course of corticosteroid therapy was performed, mainly for “nonresolving ARDS.”