Prevalence and outcomes of pleuroparenchymal fibroelastosis in patients with non-tuberculous mycobacterial pulmonary disease: a French retrospective cohort study

Abstract

Background

Pleuroparenchymal fibroelastosis (PPFE) is a recently described interstitial lung disease associated with multiple diseases. Japanese studies found an association between PPFE and non-tuberculous mycobacterial pulmonary disease (NTM-PD), but no study has been conducted in Europe. We aimed to assess in France the prevalence of PPFE among patients with NTM-PD, and to look for risk factors of PPFE and mortality.

Method

We conducted a retrospective cohort study in two French hospitals from January 1, 2020 to March 31, 2022. NTM-PD patients were included if they had at least one chest high-resolution computed tomography (HRCT) scan available, to identify the presence or the occurrence of radiological patterns of PPFE based on established criteria. Risk factors for PPFE and five-year mortality were assessed using univariate and multivariate analyses.

Results

A total of 101 patients were included, PPFE was identified in 14 patients (13.9%) and was associated with a lower body mass index (BMI) at diagnosis [odds ratio (OR) 21.4, 95% confidence interval (CI): 3.5–415.3 for BMI <18 kg/m² with PPFE, P<0.001] in multivariate exploratory analysis. Seventy patients had a minimum of two chest HRCT scans spaced by at least 3 years. Among them, 10 had PPFE: 8 at NTM-PD diagnosis, and 1 PPFE developing from pleural caps 9 years later. Five-year mortality could be studied for 71 patients. PPFE was associated with 5-year mortality (OR 6.3, 95% CI: 1.3–29.9, P=0.02) in multivariate exploratory analysis.

Conclusions

PPFE is frequently observed among patients with NTM-PD, and seems to be associated with a lower BMI and the 5-year mortality, supporting its systematic screening in this population.

Highlight box.

Key findings

• Pleuroparenchymal fibroelastosis (PPFE) is a relatively frequent entity found in a non-Asian population with non-tuberculous mycobacterial pulmonary disease (NTM-PD)

• The association of PPFE and NTM-PD tends to have a poor prognosis, associated with both undernutrition and 5-year mortality.

What is known and what is new?

• Three Japanese studies found an association between PPFE, a recently described interstitial lung disease with multiple causes, and NTM-PD, but no study has been conducted in Europe.

• In our study, PPFE was identified in 14 patients out of 101 included (13.9%), which was consistent with Japanese studies results, supporting a physiopathological link between both entities whatever the country nor the ethnicity. Out of 70 patients had two or more chest high-resolution computed tomography (HRCT) scans spaced by at least 3 years, 10 had PPFE at NTM-PD diagnosis, with 3 cases of radiological progression. PPFE was significantly associated with a lower body mass index (BMI) at diagnosis [odds ratio (OR) 21.4, 95% confidence interval (CI): 3.5–415.3 for BMI <18 kg/m² with PPFE, P<0.001], and five-year mortality (OR 6.3, 95% CI: 1.3–29.9, P=0.02) in multivariate exploratory analysis.

What is the implication, and what should change now?

• PPFE seems to be an interesting feature to look for chest HRCT scans in patients with NTM-PD and to include in pronostic scores such as the BACES (body mass index, age, cavity, erythrocyte sedimentation rate, sex) score.

Introduction

Pleuroparenchymal fibroelastosis (PPFE) is a rare fibrosing interstitial lung disease (ILD) predominantly affecting the apex (1), first described in 1992 (2) and now recognized by the American Thoracic Society as a distinct entity (3). PPFE may be idiopathic, associated with mutations in telomere-related genes (4) or secondary to other conditions such as lung or hematopoietic stem cells (HSCs) transplantations (5), thoracic radiotherapy (6), chemotherapy (7) or connective tissue diseases (8,9). These diverse associations have led some authors to propose that PPFE may represent a shared pathological response to various forms of pulmonary parenchymal aggression, regardless its nature (10).

Histopathology remains the gold standard (11) but is rarely used because of its morbidity, and high-resolution computed tomography (HRCT) with characteristic radiological features is usually sufficient for diagnosis (3,10,11).

The prognosis is worse than for idiopathic pulmonary fibrosis (IPF), with mortality being driven by chronic respiratory failure (12). Identified predictive factors of mortality include older age, male sex, malnutrition, extended radiological involvement and severe functional impact (13). The therapeutic benefit of antifibrotic agents is mixed depending on the studies (14,15), and lung transplantation remains the preferred treatment (16).

Several recent retrospective studies (17-19) all conducted in Japan, have reported a high prevalence of PPFE among patients with non-tuberculous mycobacterial pulmonary disease (NTM-PD), and a worsening of both overall and respiratory prognosis when the two entities co-exist with a hazard ratio for mortality varying from 1.76 to 35.3. A strong association has also been described between malnutrition and the existence of NTM-PD and PPFE, whether occurring independently or concurrently (3,10,11). However, the causal relationship between NTM-PD and PPFE remains unclear, with no data available outside of Japan.

Although the distribution of NTM-PD varies geographically (20), multiple studies have reported a rising global incidence, especially in East Asia (21), and identified main risk factors as malnutrition (22), older age (23) and an underlying respiratory condition (24). Despite multimodal treatment combining prolonged multidrug antibiotic regimens (25) and non-pharmacological interventions (26,27), NTM-PD frequently relapses, and can cause or worsen chronic respiratory failure (28).

The aim of this study is to assess the prevalence of radiological PPFE in a cohort of French patients with NTM-PD, to assess PPFE presence or occurrence during NTM-PD evolution, and to identify potential risk factors for the development of PPFE and for mortality in this population. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1154/rc).

Methods

Study design

This study is a retrospective cohort study, conducted in the pneumology departments of Hôpital Pitié-Salpêtrière and Hôpital Tenon in Paris.

Patients hospitalized or seen in consultation for suspected or confirmed NTM-PD between January 1, 2020 and March 31, 2022 were screened for inclusion. All data were collected until March 31, 2025 or patient’s death if it occurred before this date.

Inclusion criteria were:

1. Age ≥18 years old;

2. Criteria for NTM-PD according to the 2020 official recommendations of the European Respiratory Society, the European Society of Clinical Microbiology and Infectious Diseases, the American Thoracic Society and the Infectious Diseases Society of America (25);

3. At least one chest HRCT scan available from the diagnosis of NTM-PD.

Patients who had active tuberculosis or could not provide a non-objection notice were excluded.

HRCT scan diagnostic criteria for PPFE are based on those of the 2012 study by Reddy et al. (11) and those used in the initial Japanese studies (6,17-19):

1. Definite PPFE: bilateral pleural thickening of at least 5 mm associated with predominant subpleural fibrosis (loss of volume of the upper lobes and bi-apical traction bronchiectasis) in the upper lobes, with less or no involvement of the middle and lower lobes;

2. Inconsistent with PPFE: absence of pleural thickening and subpleural fibrosis;

3. Platythorax and supra-sternal notch are evocative features of PPFE but were not mandatory for the diagnosis.

Apical caps, the main differential diagnosis of an early PPFE, were defined as a unilateral or bilateral apical pleural thickening with a maximal thickness of 5 mm, as in the Japanese studies (17,19).

All HRCT scans were independently reviewed by a blinded expert chest radiologist trained in the detection of PPFE.

Histological analysis of the lung parenchyma was systematically screened, without being required to perform diagnosis of PPFE.

The data collected included:

1. Factors related to the patient: sex, ethnicity, age and body mass index (BMI) at diagnosis of NTM-PD, last available BMI, nature of underlying chronic lung disease, associated comorbidities;

2. Factors related to NTM-PD (species, date of diagnosis, number of recurrences, presence of cavities) and documented bacterial and/or fungal lung co-infections defined according to the criteria of the 2017 European Respiratory Society (ERS) consensus conference (29) that occurred during the study period;

3. Factors related to respiratory disability: pulmonary function tests (PFTs): ratio of forced expiratory volume in one second (FEV1) to forced vital capacity (FVC), FEV1, FVC, corrected carbon monoxide diffusion (cDLCO);

4. BACES (body mass index, age, cavity, erythrocyte sedimentation rate, sex) score calculation performed as previously described by Kim et al. (30) when possible, considering C-reactive protein (CRP) >5 mg/L as a proxy of elevation of the erythrocyte sedimentation rate. The chosen cut-off for an elevated BACES score was >2, given the median value of the BACES score between 2 and 3 in the original derivation and validation cohorts in Kim et al., and the 5-year mortality risk exceeding the threshold of 10% when the BACES score is >2.

The primary outcome of the study is the prevalence of PPFE in our cohort with NTM-PD.

Secondary outcomes are the assessment of PPFE at diagnosis or during evolution of NTM-PD, the evaluation of risk factors associated with the presence of PPFE, the evolution of respiratory disability and 5-year mortality from diagnosis of NTM-PD according to the presence or absence of PPFE.

For these purposes, 2 different sub-cohorts were extracted from the main cohort:

1. Sub-cohort A: patients who had 2 or more chest HRCT scans with a span of at least 3 years, to study if PPFE is present at diagnosis or occurs during NTM-PD evolution;

2. Sub-cohort B: patients who had a clinical follow-up equal to or longer than 5 years, or died before this delay, to study factors associated with mortality.

Statistical analysis

Continuous numeric variables are expressed with the median, first and third quartiles. Categorical variables are expressed with absolute number and relative proportion of each category.

Non-parametric statistical tests were used to make no hypothesis about the normality of the distribution of the variables.

The comparison between the groups was performed with the non-parametric Mann-Whitney test for continuous variables and with the Chi-2 test for non continuous variables.

The alpha risk is bilateral and set at 5%. The P value is considered statistically significant if strictly inferior to 0.05.

Exploratory multivariate analyses on predictive factors of PPFE and 5-year mortality were performed with multiple logistic regression model, which take into account the probable short number of events we thought we were going to note in the cohort, in order to keep the robustness of the models.

As it is a retrospective study with a descriptive outcome which is the prevalence of PPFE in our cohort with NTM-PD, we did not calculate a sample size.

All statistical analyses were performed with GraphPad Prism software version 9.0.0.

Ethics

A non-objection notice was collected orally after submission of a written information.

The present study received a favorable ethical and regulatory opinion from the Comité d’Évaluation des Protocoles de Recherche Observationnels (CEPRO) of the Société de Pneumologie de Langue Française (SPLF).

A declaration was made to the Commission Nationale de l’Informatique et des Libertés (CNIL) concerning the data collected. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Results

A total of 236 patients followed for suspected or confirmed NTM-PD between January 1, 2020 and March 31, 2022 were screened for inclusion. Of these, 135 (57.2%) were excluded for the following causes: lack of NTM-PD criteria (n=89), no chest CT scan available (n=44), active tuberculosis (n=2).

A total of 101 patients (42.8%) were included in the final analysis. The cohort was composed of 77.2% women, with a median age of 65 years [interquartile range (IQR), 56–73 years] and mainly of European descent (83.2%). The most prevalent underlying respiratory conditions were bronchiectasis (55.4%), chronic obstructive pulmonary disease (COPD) (11.9%) and post-tuberculosis sequelae (6.9%). There were 48.5% active or ex-smokers, with a median cumulative smoking score of 15 (IQR, 10–36) pack-years. Among them, 70 had a minimum of two chest HRCT scans spaced by at least 3 years (Figure 1, sub-cohort A) and 71 had a clinical follow-up equal to or longer than 5 years (Figure 1, sub-cohort B).

Figure 1.

Study flow chart of the eligible and analyzed population. HRCT, high-resolution computed tomography; NTM, non-tuberculous mycobacterial; PPFE, pleuroparenchymal fibroelastosis.

The median span between the microbiological diagnosis of NTM-PD and the first HRCT scan available was 0.6 months (IQR, 0–10.8 months) for the overall cohort and 1.2 months (IQR, 0–7.2 months) for sub-cohort A. For 68 (67.3%) patients, the first scan was available in a time shorter than or equal to 3 months after the diagnosis of NTM-PD. In the sub-cohort A patients, the medium span between the initial and the last available chest HRCT scans was 5.5 years (IQR, 3.5–7.5 years).

PPFE epidemiology and risk factors

Definite PPFE was identified in 14 patients, corresponding to a prevalence of 13.9%. (Figure 1).

Among the 10 PPFEs from sub-cohort A: one developed from apical caps in a period of 9 years (Figure 2A,2B), 8 already demonstrated computed tomography (CT) features consistent with PPFE at the time of NTM-PD diagnosis and one had no HRCT scan available in the first year from NTM-PD diagnosis.

Figure 2.

Apical cap (white arrow of A) at 2.5 years from the diagnosis of NTM pulmonary infection, worsening with transformation into subpleural fibrosis typical of PPFE (white arrow of B) at 9.5 years from the diagnosis of NTM pulmonary infection. NTM, non-tuberculous mycobacterial; PPFE, pleuroparenchymal fibroelastosis.

None of the patients in our cohort had histological analysis for the supposed diagnosis of PPFE. Only one had a lobectomy for hemoptysis, with no radiological nor histological criteria for PPFE.

Compared to patients without PPFE, those with PPFE had a significantly lower BMI at diagnosis (50% BMI <18 kg/m² vs. 19.5%, P=0.01), tended to present more often bronchiectasis as an underlying respiratory disease (78.6% vs. 51.7%, P=0.06) and to have a higher BACES score (66.7% of BACES >2 vs. 29.4%, P=0.07) (Table 1).

Table 1. Baseline characteristics of the total population.

Variable	Total (n=101)	Without PPFE (n=87)	With PPFE (n=14)	P value†
Age
At diagnosis, years	65 (56–73)	64 (55–72.5)	70 (63.5–76.8)	0.08
≥70 years at diagnosis	39 (38.6)	32 (36.8)	7 (50)	0.35
Sex (men)	23 (22.8)	21 (24.1)	2 (14.3)	0.42
Ethnicity
Asian	5 (4.95)	5 (5.75)	0 (0)	0.36
Non-Asian	96 (95.1)	82 (94.3)	14 (100)
Including Europeans	84 (83.2)	73 (83.9)	11 (78.6)
BMI
At diagnosis, kg/m²	19.7 (17.6–21)	20.2 (18.2–21.2)	16.6 (15.6–17.6)	<0.001
<18 kg/m² at diagnosis	24 (23.7)	17 (19.5)	7 (50)	0.01
At the end of follow-up, kg/m²	18.8 (16.9–20.6)	16.3 (15.6–17.9)	19.0 (17.7–20.8)	0.007
<18 kg/m² at the end of follow-up	29 (28.7)	22 (25.3)	7 (50)	0.06
BACES score at diagnosis‡	2 (1–3)	3 (2.25–3.75)	2 (1–3)	0.12
Respiratory disease
Bronchiectasis	56 (55.4)	45 (51.7)	11 (78.6)	0.06
Non-bronchiectasis	45 (44.6)	42 (48.3)	3 (21.4)
COPD	12 (11.9)	12 (13.8)	0 (0)
Post-tuberculosis sequelae	7 (6.9)	6 (6.9)	1 (7.1)
ILD	4 (3.9)	3 (3.4)	1 (7.1)
Smoking status
No-smokers	52 (51.5)	45 (51.7)	7 (50)	0.91
Smokers	49 (48.5)	42 (48.3)	7 (50)
Ex-smokers	39 (38.6)	32 (36.8)	7 (50)
Active smokers	10 (9.9)	10 (11.5)	0 (0)
Cumulative smoking (pack-years)	15 (10–36)	18 (10–41.5)	10 (6.75–17.5)
Comorbidities
Connective tissue disease	4 (3.9)	4 (4.6)	0 (0)	0.41
Immunosuppression	8 (7.9)	7 (8.0)	1 (7.1)	0.91
Cancer	25 (24.8)	23 (26.4)	2 (14.3)	0.33
Diabetes	3 (2.9)	3 (3.4)	0 (0)	0.48

Data are presented as n (%) or median (interquartile range). ^†, between patients without PPFE and patients with PPFE; ^‡, available for 57 patients (51 without PPFE and 6 with PPFE). BACES, body mass index, age, cavity, erythrocyte sedimentation rate, sex; BMI, body mass index; COPD, chronic obstructive pulmonary disease; ILD, interstitial lung disease; PPFE, pleuroparenchymal fibroelastosis.

No statistically significant differences were observed between patients with PPFE and without PPFE in terms of sex (respectively 14.3% vs. 24.1% of men, P=0.42), age (respectively 50% ≥70 years vs. 36.8%, P=0.35), smoking status (50% vs. 51.7% of non-smokers, P=0.91), or comorbidities: connective tissue disease (respectively 0% vs. 4.6%, P=0.41), immunosuppression (respectively 7.1% vs. 8.0% P=0.91), active or remitting cancer (respectively 14.3% vs. 26.4%, P=0.33), diabetes (respectively 0% vs. 3.4%, P=0.48) (Table 1).

In multivariate analysis, BMI <18 kg/m² at diagnosis remained independently associated with the presence of PPFE [odds ratio (OR) 21.4, 95% confidence interval (CI): 3.5–415.3, P<0.001] (Table 2).

Table 2. Univariate and multivariate analysis of risk factors for PPFE and for 5-year mortality.

Variables	Univariate				Multivariate
	OR	95% CI	P value		OR	95% CI	P value
Univariate and multivariate analysis of risk factors for PPFE
BMI <18 kg/m² at diagnosis	4.1	1.3–12.8	0.01		21.4	3.5–415.3	<0.001
Bronchiectasis	3.4	0.9–11.9	0.06		2.2	0.6–10.4	0.22
BACES score >2	4.8	1.0–26.5	0.07		4.8	0.8–37.4	0.08
Univariate and multivariate analysis of risk factors for 5-year mortality
PPFE	6.3	1.6–31.9	0.01		6.3	1.3–29.9	0.02
Age ≥70 years at diagnosis	5.8	1.4–21.3	0.001		5.8	1.5–28.6	0.01
Male sex	4.9	1.2–18.1	0.02		5.4	1.3–22.6	0.02
BMI <18 kg/m² at diagnosis	3.9	1.1–14.0	0.03		2.5	0.6–10.2	0.19

BACES, body mass index, age, cavity, erythrocyte sedimentation rate, sex; BMI, body mass index; CI, confidence interval; OR, odds ratio; PPFE, pleuroparenchymal fibroelastosis.

Microbiology

Patients in the cohort were mostly infected by Mycobacterium avium complex (MAC) (75.2% of the overall population). Other NTM species were less frequently represented: 11.9% Mycobacterium abscessus complex (MABSC), 10.9% Mycobacterium xenopi and 5.9% Mycobacterium kansasii (Table 3).

Table 3. Characteristics of NTM pulmonary infection and respiratory superinfections.

Variable	Total (n=101)	Without PPFE (n=87)	With PPFE (n=14)	P value†
NTM species (alone or co-infections)
MAC	76 (75.2)	63 (72.4)	13 (92.9)	0.26
Non-MAC	34 (33.7)	31 (35.6)	3 (21.4)
MABSC	12 (11.9)	12 (13.8)	0 (0)
Xenopi	11 (10.9)	11 (12.6)	0 (0)
Kansasii	6 (5.9)	5 (5.7)	1 (7.1)
Parenchymal cavities	43 (42.6)	35 (40.2)	8 (57.1)	0.24
Including with location at the apex	21 (20.8)	21 (24.1)	0 (0)	0.04
NTM infection treatment failure	55 (77.5) (n=71)	49 (69.0) (n=62)	6 (84.5) (n=9)	0.41
Median number of recurrences of NTM infections	1	1	1	0.55
Bacterial infections (alone or co-infections)
None	51 (50.5)	40 (46.0)	11 (78.6)	0.02
At least one	50 (49.5)	47 (54.0)	3 (21.4)
Pseudomonas aeruginosa	15 (14.9)	14 (16.1)	1 (7.1)
Haemophilus influenzae	19 (18.8)	18 (20.7)	1 (7.1)
Staphylococcus aureus	13 (12.9)	10 (11.5)	3 (21.4)
Other	30 (29.7)	29 (33.3)	1 (7.1)
Fungal infections
None	89 (88.1)	75 (86.2)	14 (100)	0.14
At least one	12 (11.9)	12 (13.8)	0 (0)
Aspergillus fumigatus	11 (10.9)	11 (12.6)	0 (0)

Data are presented as n (%). ^†, between patients without PPFE and patients with PPFE. MABSC, Mycobacterium abscessus complex; MAC, Mycobacterium avium complex; NTM, non-tuberculous mycobacterial; PPFE, pleuroparenchymal fibroelastosis.

All patients were treated by a first-line antibiotherapy according to the 2020 official recommendations (25). For the patients of sub-cohort B, NTM-PD recurred in 77.5% of cases, with no significant difference between patients with PPFE and without PPFE in terms of NTM relapse rate (respectively 69.0% vs. 84.5%, P=0.41) or median number of recurrences (1 per patient in both groups, P=0.55) (Table 3).

Patients with PPFE developed fewer respiratory bacterial superinfections than those without PPFE (respectively 21.4% vs. 54%, P=0.02) and no more fungal infections (respectively 0% vs. 13.8%, P=0.14) (Table 3).

Radiology

PPFE involvement was generally limited in extent, and less pronounced than radiological lesions related to NTM-PD.

Progression of PPFE was observed in 3 cases of sub-cohort A while the lesions remained stable in the others. Figure 3A-3F illustrates the case with the longest radiological follow-up (18 years between the initial and the final CT scans) and for who the worsening of PPFE lesions, both pleural thickening and upper lobe fibrosis with traction bronchectasis and volume loss, were the most visible.

Figure 3.

(A-C) Axial (A), coronal (B) and sagittal (C) images in a patient with PPFE at the diagnosis of NTM pulmonary infection. Pleural and subpleural fibrosis predominantly involving the upper lobes and the fissure is visible (white arrows) and associated with mild volume loss, subtle architectural distortion and a deep suprasternal notch (asterisk). These lesions typical of PPFE co-exist with middle lobe bronchiectasis (black arrow), a typical feature of NTM disease. (D-F) Axial (D), coronal (E) and sagittal (F) images of the same patient with PPFE at 18 years follow-up of NTM pulmonary infection PPFE. This CT scan demonstrates progression of upper lobe fibrosis (white arrows), persistent middle lobe bronchiectasis (black arrows), and newly developed bronchiolar nodules and areas of consolidation. A right upper lobe consolidation abutting the major fissure (asterisk) suggests disease progression involving both PPFE and NTM infection, with a right-sided predominance, a pattern that may be seen in PPFE. CT, computed tomography; NTM, non-tuberculous mycobacterial; PPFE, pleuroparenchymal fibroelastosis.

NTM-PD parenchymal cavities were present in 42.6% of the total population, with no significant difference between patients with PPFE and without PPFE (respectively 40.2% vs. 57.1%, P=0.24). None of the patients with PPFE had an apical cavity or an apical fibro-cavitatry lesion, that could mislead to a false positive (Table 3).

Among patients without PPFE, 44 (43.6% of the overall cohort) demonstrated apical caps (Figure 4A,4B).

Figure 4.

Axial (A) and coronal (B) CT images show a typical apical cap (white arrows), without atypical features and no evidence of PPFE. Associated lower lobe bronchiectasis with marked bronchial wall thickening (black arrow) is consistent with nontuberculous mycobacterial infection. CT, computed tomography; PPFE, pleuroparenchymal fibroelastosis.

Respiratory function

PFTs were available for 67 patients (including 6 with PPFE) at the start of follow-up, and for 30 patients (including 1 with PPFE) during follow-up. There were no statistical differences between groups for the first PFTs available (Table 4).

Table 4. PFTs evaluation during follow-up.

Variable	Total	Without PPFE	With PPFE	P value†
PFTs at diagnosis	n=67	n=61	n=6
FEV1/FVC (%)	69 (57–76)	68 (57–75)	74 (62.3–80.5)	0.33
FEV1 (% of the predicted value)	81 (58.5–94)	81 (59–96)	83 (63.5–86.8)	0.81
FVC (% of the predicted value)	96 (83.5–110.5)	96 (84–111)	89.5 (73.3–97.5)	0.25
DLCOc (% of the predicted value)‡	65.5 (46.5–82)	66 (44.8–82.3)	59.5 (51.8–68.3)	0.74
PFTs at the end of follow-up	n=30	n=29	n=1
FEV1/FVC (%)	63.3 (55.3–74.3)	63 (55–72)	80
FEV1 (% of the predicted value)	74 (50.3–83.8)	72 (50–84)	83
FVC (% of the predicted value)	85 (74–109.5)	85 (74–110)	89
DLCOc (% of the predicted value)§	55 (45.8–62.8)	54 (54.5–63.5)	58

Data are presented as median (interquartile range). ^†, between patients without PPFE and patients with PPFE; ^‡, data available for 48 patients (44 without PPFE and 4 with PPFE); ^§, data available for 20 patients (19 without PPFE and 1 with PPFE). CI, confidence interval; DLCOc, corrected diffusing capacity of the lung for carbon monoxide; FEV1, forced expired volume in one second; FVC, forced vital capacity; PFT, pulmonary function tests; PPFE, pleuroparenchymal fibroelastosis.

Mortality risk factors

At 5 years from the diagnosis of NTM-PD, 11 deaths were recorded (15.5% of sub-cohort B), including 4 patients with PPFE.

The median age at death was 76.5 years (IQR, 70.8–77.3 years), and the median duration of NTM-PD prior to death was 2.5 years (IQR, 1.2–3.7 years). Causes of death were respiratory failure (n=5) general alteration (n=2), other or unknown (n=4).

Five-year mortality was significantly higher in patients with PPFE (respectively 44.4% vs. 11.3%, P=0.01), with a BMI <18 kg/m² at diagnosis (respectively 54.6% vs. 23.3%, P=0.03), with an age ≥70 years at diagnosis (respectively 72.7% vs. 31.7%, P=0.01), and in male patients (respectively 45.5% vs. 15.3%, P=0.02). There were no statistical differences between groups for BACES score (P=0.13) and presence of cavities (P=0.09).

In multivariate analysis (Table 2), the presence of PPFE (OR 6.3, 95% CI: 1.3–29.9; P=0.02), the age ≥70 years at diagnosis (OR 5.8, 95% CI: 1.5–28.7; P=0.01) and the male sex (OR 5.3, 95% CI: 1.3–22.6; P=0.02) remained independently associated with 5-year mortality, whereas BMI <18 kg/m² did not reach statistical significance (OR 2.5, 95% CI: 0.6–10.2; P=0.19).

Discussion

In our French cohort of patients with NTM-PD, the prevalence of PPFE is 13.9%, which is consistent with the results of cross-sectional Japanese studies (11.9% in Aono et al. and 26.3% in Yamamoto et al.). To our knowledge, this is the only study that evaluates PPFE in patients monitored for NTM-PD over a prolonged period in a predominantly European population. These findings support the notion that the association between NTM-PD and PPFE is not restricted to specific geographic regions or ethnic groups.

The longitudinal design of the study provided insights into the temporal relationship between PPFE and NTM-PD. Thus, 8 patients with PPFE out of the 10 among our sub-cohort A, as well as 3 out of the 4 PPFEs with only one HRCT available, already exhibited radiological features of PPFE at the time of NTM-PD diagnosis. These findings do not support the hypothesis that PPFE develops consequently to NTM-PD, subject to the possible diagnostic delay for NTM-PD.

Alternative pathophysiological hypotheses could be considered: PPFE may predispose individuals to NTM-PD, or both conditions may arise from a shared underlying susceptibility (malnutrition, genetic predisposition).

Furthermore, patients with PPFE had as many recurrences of NTM-PD and fewer bacterial and/or fungal infections, as those without PPFE, suggesting that PPFE does not seem to be a risk factor for respiratory superinfections.

In our study, bronchiectasis are surprisingly the underlying lung condition of NTM-PD in 11 out of 14 cases in patients with PPFE. These do not appear to be traction bronchiectasis related to PPFE, as they were not predominantly located in the upper lobes. Bronchiectasies are a classic risk factor for NTM-PD, but are hardly described in association with PPFE, excepted in cases of bronchiolitis obliterans secondary to lung or allogeneic HSC transplantation (31). In the Japanese studies, this statistical association between bronchiectasis and PPFE prevalence was not observed, although Yamamoto et al. (17) found a significant link between bronchiectasis and PPFE severities.

Therefore, the presence of bronchiectasis at the time of PPFE diagnosis should prompt evaluation for concomitant or future NTM-PD.

None of the cases of PPFE in our cohort were identified for any classically described risk factors (connective tissue diseases, history of HSC or lung transplantation, combination with IPF). Two out of 14 cases of PPFE had a history of cancer in complete remission, but had not received chemotherapy or thoracic radiotherapy.

None of the patients with PPFE in our cohort received specific treatments for PPFE.

The radiological diagnosis of PPFE was explicitly mentioned in chest CT scan reports in only 4 (28.6%) of patients with PPFE, before our central expert radiological revision with strict criteria, highlighting that PPFE remains still largely underdiagnosed.

This pleads for training and a general raising of awareness among radiologists for its detection. Conversely, some apical lesions could be misinterpreted as PPFE features such as pleural calcifications, post-radiotherapy sequelae or a continuity between a parenchymal cavity with an apical cap.

Recent small series (32) or case reports (33) have established a link between apical cap, historically described in NTM-PD (34), and risk for secondary development of PPFE. Apical caps and PPFE share the same underlying conditions as malnutrition (34) and have histological lesions similarities (33), which could lead to errors.

In our cohort, the number of pleural caps is high (43.6% of patients), all present at diagnosis of NTM-PD but only one case progressed to PPFE in 9 years of follow-up. Therefore, our study associates neither NTM-PD as a risk factor for onset of pleural cap, nor the pleural cap as a risk factor for onset of PPFE.

Another point of interest is the significative link between PPFE and mortality for NTM-PD patients, both in our study and Japanese ones. Whatever variables integrated in our multivariate analysis models, PPFE remained independently associated with mortality, subject to the small sample size of our cohort and events (PPFE and deaths).

Incorporating PPFE into prognostic scores for NTM-PD could enhance risk stratification and help identify patients at higher risk of poor outcomes.

Moreover, our cohort and the Japanese studies highlight an association between malnutrition and increased mortality, in both NTM-PD and PPFE, supporting the hypothesis of a particular phenotype (female sex, older age, malnutrition). Multiple pathophysiological mechanisms may contribute to this relationship (22,35,36): hypercatabolism, abnormally amplified pro-inflammatory response by adipocytes, intestinal malabsorption, dysmorphophobia.

Several limitations of this study should be acknowledged.

First, the sample size is relatively small and the retrospective nature of the data collection led to missing clinical and radiological variables. This may have impacted the robustness of the analyses and explains why certain results identified in univariate analyses and in the Japanese studies did not reach statistical significance in multivariate models, limiting external validity.

Then, even if the chest radiologist involved in our study was trained to identify PPFE, the single blind review is less robust than a double or triple independent review, given that PPFE detection can be highly operator-dependent. Moreover, the lack of histological evidence of PPFE diagnosis has to be consider as a limitation of our study. But, in another hand, in daily clinical practice, lung biopsy is often considered as dangerous in the PPFE context with a very high risk of difficult-to-treat pneumothorax, and therefore is rarely performed.

Among the 71 patients with a long follow-up, fewer than 50% of them had a regular PFTs evaluation. This highlights a gap in the functional monitoring of NTM-PD despite its probable major prognostic relevance.

Conclusions

In this French cohort of patients with NTM-PD, PPFE was identified in 13.9% of cases, with the majority of PPFE lesions already present at the time of NTM diagnosis and only one case developing from apical caps during follow-up. PPFE was significantly associated with lower BMI and emerged as an independent predictor of 5-year mortality in our study, subject to the small sample size and number of events of our cohort. These findings support the relevance of systematical research of PPFE on chest HRCT scans, and its incorporation into the prognostic evaluation of NTM-PD. Prospective trials, with a larger population and extended follow-up periods, are warranted to confirm the findings of our study.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1154/dss

jtd-18-08-845-dss.pdf ^{(68.6KB, pdf)}

DOI: 10.21037/jtd-2026-1154