Impact of lymphocyte differential count > 15% in BALF on the mortality of patients with acute exacerbation of chronic fibrosing idiopathic interstitial pneumonia

Acute exacerbation (AE) in patients with idiopathic pulmonary fibrosis (IPF), characterised by clinically progressive respiratory failure and worsening pulmonary fibrosis, was first reported in 1993 and has been recognised as a relatively common and highly critical event in the clinical course of IPF [1, 2]. AE has also been described in other cases of fibrosing interstitial pneumonia [3‐6], suggesting the potential risk of AE in chronic fibrosing idiopathic interstitial pneumonia (CFIIP).

The diagnostic consensus criteria for AE were suggested by Collard et al. in 2007 [7] and include a lack of evidence of pulmonary infection by endotracheal aspirate or bronchoalveolar lavage (BAL). However, there have been no studies of the specificity of the cellular analysis of BAL fluid (BALF) and appropriate treatment for AE has not been established, and so the mortality rate of AE remains high [8].

Although corticosteroids were weakly recommended in the American Thoracic Society/European Respiratory Society/Japanese Respiratory Society/Latin American Thoracic Association (ATS/ERS/JRS/ALAT) guidelines for patients with AE of IPF [9], no controlled trials on the effectiveness of corticosteroids have been conducted. According to the ATS/ERS guidelines, BAL cell differential counts > 15% lymphocytes represent a lymphocytic cellular pattern such as cryptogenic organizing pneumonia [10], which has been shown to have a good recovery with corticosteroid treatment [11]. So patients with BAL cell differential counts > 15% lymphocytes may be good responders to corticosteroids. However, the effect of corticosteroids and other immunosuppressants in AE of CFIIP with > 15% lymphocytes in BALF has never been evaluated.

In this study, we retrospectively analysed patients with AE of CFIIP who underwent bronchoscopy with BAL in order to evaluate the impact of the lymphocyte differential count > 15% in BALF on outcomes.

In this study, we demonstrated that the lymphocyte differential count > 15% in BALF was associated with favourable outcomes in patients with AE of CFIIP. To the best our knowledge, this is the first study to demonstrate the impact of the lymphocyte differential count > 15% in BALF on the 90-day mortality and OS in patients with AE of CFIIP.

The pathogenesis of AE of CFIIP remains unclear, but the pathological findings of AE of CFIIP were reported as DAD or organizing pneumonia (OP), without other evidence of the organizing phase of DAD [1, 4, 7, 11, 15]. The ATS guidelines indicate that the presence of > 15% lymphocytes in BALF represents a lymphocytic cellular pattern such as OP, nonspecific interstitial pneumonia, etc. [10]. Some of these conditions may respond to corticosteroids. In this study, corticosteroids and other immunosuppressants were more effective in the group with a lymphocyte differential count > 15% than in the group with a lymphocyte differential count ≤ 15%. These results suggest that the 15% lymphocytes in BALF in AE of CFIIP patients is a helpful factor for identifying the therapeutic response and prognosis in these patients. BAL cell differential counts with > 3% neutrophils represent a neutrophilic cellular pattern such as DAD [10] and, in previous reports, patients with DAD were more resistant to corticosteroid and other immunosuppressants and showed a worse prognosis than patients with OP [1, 4, 16, 17]. We analysed association the degree of neutrophil differential count in BALF and OS. The area under the receiver operating characteristic curve for the neutrophil differential count in BALF for predicting the OS in AE of CFIIP was 0.78 [95% CI: 0.59–0.92] (Additional file 1: Figure S1). The univariate analysis identified neutrophil differential count in BALF was associated with OS (HR 1.02, 95% CI 1.01–1.04, p < 0.01). In this study, however, 35 patients (94.6%) showed > 3% neutrophils in BALF and a > 3% neutrophil differential count in BALF was not useful in predicting prognosis.

In the group with a lymphocyte differential count ≤ 15%, the OS was significantly inferior to that of the group with a lymphocyte differential count > 15%. Surprisingly, in the group with a lymphocyte differential count ≤ 15%, all five patients that survived the AE died within 1 year from admission. Among these five patients, three died from a recurrence of AE and one died from CFIIP (Table 4). This suggests that in the group with a lymphocyte differential count ≤ 15%, corticosteroid and other immunosuppressants could suppress the appearance of new infiltrations, but could not stop the progression from initial ground-glass opacities to fibrosis. As a result of the progression of fibrosis, respiratory function considerably worsened and AE recurred more frequently. Several previous reports suggested that recombinant human soluble thrombomodulin, pirfenidone, and polymyxin B-immobilized fiber columns may improve survival in AE. The good practice for the management and treatment of AE based on current knowledge, evidence, and available treatment was clinician-led consensus [18]. These treatments may represent additional therapeutic options to improve the prognosis of patients with AE [19‐21].

In the present study, we analysed the prognosis of AE in all CFIIP patients including both UIP and not UIP patients. Usui et al. reported that the computed tomography classification (UIP pattern and not UIP pattern) had no significant effect on survival in AE patients [22], and we considered it valid to place all patients in the CFIIP category. However, although many reports have described the prognosis of AE in IPF patients, few reports have thoroughly investigated the prognosis of AE in not UIP patients [1‐3, 7]. Therefore, we classified all patients into the two groups (UIP and not UIP) and analysed the 90-day mortality and OS. Using a cut point of 15% lymphocytes in BALF appeared to be a helpful factor in identifying patients with AE-CFIP who may respond to immunosuppressive therapy, and this appeared to be the case for patients with or without apparent UIP. These findings suggest that >15% lymphocytes in the BALF differential cell count may be a stronger prognostic indicator than whether the underlying histopathologic lesion is UIP or a non-UIP form of CFIIP.

In this study, most of the patients underwent BAL on the first day and received treatment at an early phase. The average total volume of retrieved BALF was 43.3% of the total instilled volume, and this amount was adequate to estimate the BALF cell differential counts [10]. BAL is a safe and well-tolerated procedure [23], and no patient in this study had any serious complications, such as severe worsening of AE. Video-assisted thoracoscopic surgery was too risky to perform during AE. BAL was not only helpful distinguishing AE from infection [10], but it was also useful and safe for evaluating the impact of the lymphocyte differential count in BALF on the mortality of patients with AE of CFIIP.

There are several limitations to the present study. The first is its retrospective design. The second is the fact that it included a relatively small number of patients with AE of CFIIP. The third is that the diagnosis of CFIIP and the classifications were somewhat dependent on the HRCT scanning. Especially, the diagnosis of nonspecific interstitial pneumonia was not adequately accurate because most of the patients were difficult to perform surgical lung biopsies. So, we classified CFIIP into UIP and not UIP and showed “UIP/not UIP” in Table 1. Needless to say, we excluded, in reference to the 2013 ATS guideline, known entities associated with the development of pulmonary fibrosis such as collagen vascular disease and chronic hypersensitivity pneumonitis. The fourth is that we have not compared between BAL in stable CFIIP and in AE of CFIIP. BAL differential counts in stable CFIIP may influence that in AE of CFIIP. However, BAL was performed at the new abnormal infiltration and may represent the acute phase of AE of CFIIP. Finally, because some patients were treated with corticosteroids or other immunosuppressants for CFIIP before AE, the BALF cell differential counts may have been affected. Prospective studies with larger patient cohorts are required to overcome these limitations.

In conclusion, our findings indicate that the lymphocyte differential count > 15% in BALF may be associated with good outcomes in patients with AE of CFIIP. Further studies are needed to confirm our findings, to elucidate the mechanism of AE of CFIIP, and to develop the optimal treatment strategies for AE of CFIIP.