Prognostic Value of Concomitant Bronchiectasis in Newly Diagnosed Diffuse Panbronchiolitis Patients on a Maintenance Therapy with Macrolides

Background Factors determining the prognosis of diffuse panbronchiolitis (DPB) remain unclear at present. The objective of this study was to identify the prognostic value of concomitant bronchiectasis in the macrolide treatment efficacy and exacerbation risk in DPB patients. Methods Data of patients initially diagnosed with DPB at the Shanghai Pulmonary Hospital between January 2007 and December 2017 were retrospectively collected and analyzed. The patients were divided into two groups according to the existence of bronchiectasis. Clinical manifestations, laboratory findings, microbiological culture results, as well as exacerbation risks and treatment outcomes, were compared between these two groups. The survival curve and Cox regression analysis models were additionally constructed to further demonstrate the predicting role of bronchiectasis in DPB exacerbation. Results Baseline data revealed more respiratory symptoms, lower body mass index (BMI), and forced expiratory volume in one second (FEV1) as well as increased isolates of Pseudomonas aeruginosa (P. aeruginosa) in DPB subjects with bronchiectasis than those without. Furthermore, bronchiectasis was associated with a lower rate of responsiveness to macrolides and increased exacerbation frequency during follow-up. The survival curve and Cox regression analysis showed that comorbid bronchiectasis was linked to increased time to episode relapse, which remained significant even after controlling for BMI, FEV1, and P. aeruginosa culture results. Conclusion The coexistence of bronchiectasis predicted a poor outcome of maintenance macrolide therapy and an increased exacerbation risk in DPB subjects, possibly through its impacts on nutritional status, pulmonary function, and P. aeruginosa infections.


Introduction
Diffuse panbronchiolitis (DPB) is an idiopathic inflammatory disease, which affects the distal airways and predominantly the transition zone between the respiratory bronchioles and alveoli [1]. It is characterized by progressive suppurative and obstructive airway disease. If left untreated, DPB will progress to bronchiectasis, respiratory failure, and eventually death [2]. e prognosis of DPB was poor before the long-term treatment with erythromycin was introduced. Since then, the 10-year survival rate has improved from 33.2% to 90% [3]. However, there are still cases in which the benefit is minimal and the disease relapses or progresses in spite of causes of macrolide antibiotics treatment. e factors that affect treatment efficacy and relapse are still unknown [4,5].
Bronchiectasis is a chronic lung disease characterized by a vicious cycle of airway infection and inflammation, leading to permanent structural damage to the small airways and sometimes to the surrounding lung parenchyma [6]. e clinical presentation can overlap with other respiratory disorders such as chronic obstructive pulmonary disease (COPD), asthma, obstructive sleep apnea-hypopnea syndrome (OSAHS), and so on, forming overlap syndromes [7][8][9]. Recently, several reports indicated that comorbid bronchiectasis was associated with poorer treatment efficacy and prognosis for various other chronic lung diseases. Bronchiectasis comorbidity also resulted in increased frequency of cases with difficult-to-treat pathogenic bacteria, decreased lung function, and increased risk of death from all causes, among other symptoms [10,11]. DPB has a very wellestablished relationship with bronchiectasis. DPB can occur in conjunction with bronchiectasis or progress to severe bronchiectasis in its advanced stage. Bronchiectasis in conjunction with DPB has been assumed a more severe syndrome in DPB. However, this conclusion was mainly based on clinical observations and expert opinions. Confirmation and detailed data from systematic clinical research are still lacking.
We hypothesized that bronchiectasis would be a negative factor for DPB treatment and exacerbation risk, similar to its role in other chronic lung diseases. erefore, we conducted a retrospective cohort study to identify the prognostic value of bronchiectasis in DPB, with the underlying mechanisms explored.

Patients and Data
Collection. Data of 395 cases with initially suspected DPB from 2007 to 2017, who were then prescribed a macrolide regimen and asked for follow-up on a regular basis, were retrospectively screened and collected from the Shanghai Pulmonary Hospital. Among them, 217 were finally enrolled into the cohort, following the flow chart for selecting the eligible patients that is shown in Figure 1. All patients were well documented with detailed clinical information, including baseline demographic data, arterial blood gas analysis, blood cold hemagglutination (CHA) testing, pulmonary function tests, and microbiological and laboratory results of sputum and bronchial lavage fluid. e radiological findings were also required, with all patients presented typical bilateral diffused tree-in-bud patterns. Lung biopsy was performed in 10 cases, all of which were pathologically consistent with DPB. During the follow-up period, information on treatment efficacy and exacerbation frequency was collected and assessed. Patients were divided into two groups according to the existence of bronchiectasis in chest CT scan images, except for which no meaningful differences were found. e study was approved by the Ethical Committee of Tongji University, Shanghai Pulmonary Hospital. All of the participants signed informed consent for any procedures that were relevant to this study.

Diagnostic Criteria of DPB and Bronchiectasis.
e DPB diagnostic criteria were in accordance with the guidelines proposed by a working group of the Ministry of Health and Welfare of Japan [12]. A chest high-resolution computed tomography (HRCT) scan was used to confirm the diagnosis of bronchiectasis, based on the diagnostic criteria proposed by Naidich et al. [13]. Cases with small bronchiectasis that was only visible in a single pulmonary segment, which also can be detected in healthy people, were excluded. Patients with bronchiectasis caused by congenital diseases such as cystic fibrosis and Marfan syndrome were not found in this study.

Assessment of Treatment Efficacy.
e treatment strategy was a combination of macrolide antibiotics and other measures, including glucocorticoids, antibiotics against infectious pathogens, oxygen therapy, symptomatic and supportive therapy, etc [2,12,14]. e patient's response to macrolide antibiotics treatment was defined as follows [15]: (1) Healed: all symptoms and abnormal body signs disappeared; chest CT, arterial blood gas analysis, and lung function returned to normal. (2) Improved: more than one of the clinical symptoms, body signs, or imaging scan improved after treatment. (3) Unchanged: clinical symptoms and abnormal body signs persisted with no improvement observed in the chest CT. (4) Deteriorating: one or more clinical symptoms, abnormal body signs, or chest CT were aggravated/became worse. (5) Recurrent: healed after treatment, but clinical or radiological regression occurred during follow-up. Conditions (1) or (2) was considered as responsive. Conditions (3), (4), and (5) were considered as unresponsive.
e symptoms and body signs were documented in the medical records. CT images were analyzed and interpreted by two radiologists and a pulmonologist who were blinded to the case. e medical records and radiological reports were evaluated by two pulmonologists to confirm the accuracy of the extracted information.

Definition of DPB
Exacerbations. An exacerbation of DPB was defined, in our study, as the presence for at least 24 h of increased symptoms (including cough, sputum volume and/or consistency, sputum purulence, breathlessness and/or exercise tolerance, fatigue and/or malaise, hemoptysis), with or without an increase in tree-in-bud in CT Diagnosis cannot be confirmed by pathology and/or follow-up n = 29 All patients with initial diagnosis of suspected DPB n = 395 Patients with insufficient data n = 149 Cohort for final analysis n = 217 DPB with bronchiectasis n = 129 DPB without bronchiectasis n = 88 scans, resulting in an unscheduled clinical visit and a subsequent hospital admission.  Tables 1  and 2, through a univariate Cox regression analysis, with factors with a P value of <0.1 entering into the final analysis. P values of <0.05 were considered statistically significant.

Baseline Demographic, Clinical, and Laboratory Findings between DPB Patients with and without Bronchiectasis.
A total of 217 from 395 DPB patients were included into the final analysis. Bronchiectasis was present in 129 patients. e baseline characteristics for all patients are shown in Table 1. DPB patients with bronchiectasis demonstrated significantly lower BMI than those without bronchiectasis (P < 0.001). No significant difference in age, sex, smoking status, personal history, and previous disease history existed between the two groups.
Arterial blood gas analysis showed a lower level of arterial partial pressure of oxygen (P a O 2 ) in DPB subjects with bronchiectasis (P � 0.037). No significant differences in blood CHA test, white blood cell count, neutrophil percentages, C-reactive protein (CRP), or erythrocyte sedimentation rate (ESR) were found between the two groups.
Both groups exhibited damaged pulmonary functions, which were more pronounced in the DPB subjects with bronchiectasis, as reflected by lower levels of forced expiratory volume in one second (FEV 1 , P � 0.040), forced vital capacity (FVC, P � 0.058), and FEV 1 /FVC (P � 0.070), with the former reaching statistical significance. No difference in forced expiratory flow 75% (FEF75) and residual volume (RV) was found.

Follow-Up Outcomes between DPB Patients with and without Bronchiectasis.
erapeutic regimens and efficacy assessments are shown in Table 2. ere were no differences in the selection of macrolide antibiotics and regular use of other auxiliary medicines, including inhaled corticosteroid (ICS), long-acting bronchial agonist (LABA), and longacting muscarinic antagonist (LAMA). DPB patients with bronchiectasis experienced more aggravating events (P � 0.048). e rate of responsiveness (healed plus improved) was significantly lower in DPB patients with bronchiectasis than those without (P < 0.001).

Discussion
In this paper, we conducted a relatively large retrospective cohort study, enrolling 217 subjects. To our knowledge, no other study in the literature has surpassed this case number. Our study also captured data from a long followup duration (up to 87 months) to confirm the potential adverse impact of comorbid bronchiectasis in newly diagnosed DPB patients on the efficacy of macrolide maintenance therapy and the exacerbation risk. As expected, our results demonstrated a poorer outcome of macrolide treatment, as well as a shorter period to episode relapse and increased exacerbation frequency since study enrollment in newly diagnosed DPB patients with bronchiectasis compared with their bronchiectasis-free counterparts. Meanwhile, bronchiectasis was also associated with accelerated chronic wasting, more symptoms, poorer lung function, and increased isolates of P. aeruginosa at baseline. According to the multiple Cox regression analysis, comorbid bronchiectasis (in combination with a couple of other clinically significant confounding variables including BMI, FEV 1 , and positive P. aeruginosa culture) was a strong independent factor of macrolide treatment. It could thus be speculated that bronchiectasis exerts a detrimental influence on DPB disease control and prognosis, at least through pathways involving nutritional condition, pulmonary function, and lower airway tract infections.

Canadian Respiratory Journal
P. aeruginosa is a well-documented pathogen that provokes an intense inflammatory response leading to persistent airway inflammation and airway structural damage [16,17]. It has been reported that in the early stage of DPB, Haemophilus influenzae represents the most common colonized microbiolgical agent (44%), followed by P. aeruginosa (22%). Colonization with P. aeruginosa eventually occurs, and the detection rate of P. aeruginosa in DPB can rise to 60%. is accelerates disease progression, causing a wide range of severe opportunistic infections [1]. Chronic colonization by P. aeruginosa is associated with a greater decline in pulmonary function and more frequent exacerbation and hospitalization as well as increased mortality [18,19]. Consistently, in our study, occurrence of P. aeruginosa was more frequently detected in DPB patients with bronchiectasis and was partly responsible for the poor therapeutic outcomes.
Nutritional status is frequently underestimated due to a lack of awareness by health professionals who deal with different chronic respiratory diseases. However, nutritional abnormalities are highly prevalent in this field, leading to important clinical consequences [20]. It has been reported in COPD patients that those with malnutrition present higher degrees of airway obstruction, perception of dyspnea, and CAT scores while having lower exercise capacity [21,22]. As for bronchiectasis, fat-free mass depletion is related to increased inflammatory activity [23]. ere are no published data to our knowledge that refer to the relationship between DPB and malnutrition. erefore, for the first time, our study presented evidence that lower BMI is linked to likelihood of comorbid bronchiectasis and as such has a negative impact in therapeutic outcome. is link justifies a nutritional intervention in the treatment of DPB. e underlying mechanisms are still unknown. However, it may involve impairment of immunological function and increased susceptibility to and severity of community-acquired pneumonia [24][25][26]. Weakness of the respiratory muscles may also play a role.
Pulmonary dysfunction could be a natural consequence of long-term chronic airway inflammation and recurrent infectious episodes. In our manuscript, a decrease in pulmonary function, reflected by a lower FEV 1 value, represented a third confounding risk factor in addition to P. aeruginosa infection and malnutrition in predicting a  e results showed that bronchiectasis remarkably potentiated the risk of exacerbation events. shortened duration of DPB aggravation since maintenance macrolide therapy. is finding was consistent with the established role of FEV 1 in the assessment and prognosis of asthma and COPD [27,28]. Notably, no difference in FEF75 was documented in our reports, and in view of the large and small airway involvement of FEV 1 and FEF75, respectively, our results suggested that coexistence of bronchiectasis affected large rather than small airways of DPB patients.
It should be pointed out that concomitant bronchiectasis remained significant in the final Cox regression model, even after adjusting for all of the abovementioned factors, indicating other pathways that are involved in the relationship between bronchiectasis and DPB but were not identified here. However, it is reasonable to assume that the irreversible airway damage, the persistent inflammatory response, and the resulting recurrent infections (features involved in bronchiectasis), which cannot be reversed even with an effective therapy, may help explain the difficulty in treating and the poor prognosis of DPB in combination with bronchiectasis.
Several issues and limitations should be mentioned here. First, bronchiectasis may be considered as a primary or secondary condition, which could not be distinguished during our data collection due to the retrospective nature of this study. Second, since there was a marginal (but not statistically significant) difference in DPB duration between our patients with and without bronchiectasis, it is thus possible that the poorer therapeutic effects and prognosis found in the bronchiectasis group might only represent the consequence of advanced or end-stage DPB. However, we do not think this is the case. By using a multiple Cox regression analysis to predict treatment responsiveness, DPB duration was finally excluded from the model, but bronchiectasis remained a predictor of treatment outcome. In addition, we cannot determine whether the P. aeruginosa was a colonizing or infecting bacteria. Furthermore, the virulence and resistance of the bacteria cannot be defined either, which will also affect the treatment efficacy and prognosis.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors have no conflicts of interest to declare.

Authors' Contributions
Benyong Xu, Yanhua Mao, and Xiaoyu Wan contributed equally to this work.