Introduction

Hypersensitivity pneumonitis (HP) is a type of interstitial lung disease (ILD) caused by inhalation of environmental organic antigens in sensitized individuals. The clinical presentation and natural disease course show high heterogeneity, with contributing factors including the intensity and duration of exposure to antigens, nature of antigens, and host factors.[1,2] Thus, how to classify the clinical phenotypes of HP has been challenging.

HP has conventionally been classified into acute, subacute, and chronic subtypes according to the disease duration at the time of onset.[3] However, definitive criteria for distinguishing these three categories, particularly the subacute type, have not been established. Moreover, this classification does not aid in prognostic prediction.[4,5] With an in-depth understanding of ILD, Vasakova et al[4] proposed two main categories in terms of clinical–radiological–histopathological correlation: acute/inflammatory and chronic/fibrotic. This classification highlights the importance of the natural progression of HP. Increasingly, the effect of fibrosis on the prognosis of HP has attracted great attention. Based on updated knowledge, the 2020 American Thoracic Society (ATS)/Japanese Respiratory Society (JRS)/Asociación Latinoamericana del Tórax (ALAT) clinical practice guidelines recommended classifying HP as fibrotic hypersensitivity pneumonitis (FHP) and non-fibrotic hypersensitivity pneumonitis (NFHP) based on the presence/absence of radiological and/or histopathological fibrosis.[6] This classification provides information about prognosis and influences the selection of treatment strategies.

Fibrotic changes revealed by radiological and pathological methods are broadly considered to be independent predictors of adverse outcomes in HP.[7–10] Nevertheless, studies on the development of FHP are still scanty. Epidemiological data suggest that older age and cigarette smoking were independent risk factors for the development of fibrosis.[11–14] Other factors correlated with poor survival, such as inability to identify inciting antigens,[13] concurrent autoimmune features,[15] and hypothyroidism[16] were speculated to facilitate the development of fibrosis. Therefore, understanding the differences between NFHP and FHP may provide clues to the mechanisms of fibrotic progressions in HP. The aim of this study was to systematically summarize differences in clinical characteristics and prognosis between NFHP and FHP and explore factors associated with the development of FHP.

Methods Study population

This prospective cohort study was conducted in China-Japan Friendship Hospital, Beijing, China. Consecutive patients with age ≥18 years diagnosed as HP were enrolled from January 1, 2017 to August 31, 2021. Meanwhile, patients without available high-resolution computed tomography (HRCT) images were excluded from this study. The study was approved by the Institutional Ethics Committee of China-Japan Friendship Hospital, Beijing, China (No. 2017-25-1). Written informed consent was obtained from all patients.

All patients were diagnosed based on multidisciplinary discussion according to the institutional diagnostic procedures of ILD. The multidisciplinary team consisted of at least one respiratory physician, one pathologist, and one radiologist. Typical HP was defined as having antigen exposure history, consistent clinical manifestations, suggestive HRCT appearances (extensive ground-glass opacities [GGOs] and/or centrilobular nodules and/or peribronchovascular fibrosis and/or upper lobe predominant fibrosis and/or air trapping), and bronchoalveolar lavage (BAL) lymphocytosis (defined as the percentage of lymphocyte in BAL fluid ≥20%) or biopsy results which suggested HP. Probable HP referred to cases with less diagnostic confidence of HP, and was defined based on the history of environmental exposure and HRCT appearances without BAL lymphocytosis or suggestive pathological findings.[17] Antigen exposure was detected through detailed clinical history.

Patients were followed up through regular clinical visits every 3 to 6 months or at time intervals determined by clinical requirements, or telephone interview for patients who were unable to come to the clinic, until June 1, 2022. Data on clinical symptoms, laboratory examination, HRCT, and pulmonary function were collected during face-to-face visit. Adverse outcome referred to death or lung transplantation. Follow-up duration was defined as the time from the date of diagnosis to the date of the end of follow-up or death/lung transplantation. Eight (4.0%) patients were lost to follow up.

Clinical subtypes

HRCT and histopathological features were reviewed and patients were stratified into NFHP and FHP subtypes based on whether radiological and/or histopathological fibrosis was predominant according to the 2020 ATS/JRS/ALAT guidelines.[6] Salient features of pulmonary fibrosis on HRCT included irregular reticulation, lung architectural distortion, traction bronchiectasis, and honeycombing. Supplementary Figure 1, https://links.lww.com/CM9/B452 showed the typical radiological features of NFHP and FHP subtypes.

Chest HRCT review and scoring

Two clinicians trained in ILD radiography independently reviewed chest HRCT and reached a consensus on discrepancies. The following features were examined: centrilobular nodules, GGOs, mosaic attenuation, reticulation, honeycombing, traction bronchiectasis, and emphysema. Thoracic abnormalities were assessed semi-quantitatively. In this evaluation system, four representative levels were chosen: arcus aortae, tracheal bifurcation, 2 cm above the diaphragm, and 2 cm below the diaphragm. The extent of GGOs, reticulation, and honeycombing were scored on a 5-point scale (0: no involvement; 1: ≤5%; 2: 6–24%; 3: 25–49%; 4: 50–74%; and 5: ≥75%). Each variable score was calculated as the average of these four levels. Results of the two observers were averaged. Ultimately, fibrosis scores were the sum score of reticulation and honeycombing.

Pulmonary function test

Static lung volumes were measured using plethysmography, and the lung diffusion capacity of carbon monoxide (DLCO) was measured using the single breath-hold method. In the following part, forced vital capacity (FVC) and DLCO were expressed as a percentage of the predicted value and abbreviated as FVC%pred and DLCO%pred, respectively. Normal lung function was defined as the percentage of predicted forced vital capacity (FVC%) ≥80% and forced expiratory volume in 1 s (FEV1)/FVC ratio ≥70%; restrictive ventilation dysfunction was defined as FVC% pred <80% and FEV1/FVC ratio ≥70%[18]; obstructive ventilation dysfunction was defined as FVC% pred ≥80% and FEV1/FVC ratio <70%[19]; mixed ventilation dysfunction was defined as FVC% pred <80% and FEV1/FVC ratio <70%.[20] Diffusion dysfunction was defined as DLCO% pred <80%.[21,22] The degrees of restrictive and diffusion impairment were graded using FVC% pred and DLCO% pred, respectively. Mild restrictive ventilative and diffusive dysfunctions were defined as 60% ≤ FVC% pred < 80% and 60% < DLCO% pred < 80%, respectively; moderate restrictive ventilative and diffusive dysfunctions were defined as 40% ≤ FVC% pred < 60% and 40% ≤ DLCO% pred ≤ 60%, respectively; and severe restrictive ventilative and diffusive dysfunctions were defined as FVC% and DLCO% < 40%, respectively.[18,21,22] And small airway dysfunction was defined according to at least two of the following three indicators of lung function being ≤65% of the predicted value: maximal mid-expiratory flow, forced expiratory flow (FEF) 50%, and FEF 75%.[23]

Autoimmune features

Patients were confirmed as HP with autoimmune features (HPAF) according to a previous study,[15] that is, having a documented diagnosis of autoimmune disease, or having at least one specific connective tissue disease (CTD) symptom and one serological test suggestive of autoimmune disease. Autoimmune diseases included systemic lupus erythematosus, rheumatoid arthritis, scleroderma, Sjögren's disease, or anti-synthetase syndrome. Specific CTD symptoms included Raynaud's phenomenon, arthralgias/joint swelling, morning stiffness, dry mouth/dry eyes, and proximal muscle weakness. Positive autoimmune serologies included an anti-nuclear antibody titer ≥1:320, a rheumatoid factor ≥3 × the upper limit of normal, or positive anti-neutrophil cytoplasmic antibodies, cyclic citrullinated peptide, histidyl-tRNA synthetase (Jo-1), double stranded DNA, anti-ribonucleoprotein (anti-RNP) antibody, DNA topoisomerase 1 (Scl-70), anti-Sjögren's-syndrome-related antigen A autoantibodies, or anti-Sjögren's syndrome B (anti-SSB/La) antibodies.

Treatments

Patients’ treatments were decided by the clinicians. Whether pharmacological intervention was necessary depending on patients’ conditions. Drugs included oral glucocorticoids, antifibrotic drugs, immunosuppressors, acetylcysteine, and inhaled bronchodilators and corticosteroids.

Statistical analysis

Normally distributed continuous data were expressed as mean ± standard deviation, non-normally distributed data as median (Q1, Q3), and categorical data as number and percentages. Patients were divided into NFHP and FHP groups. Clinical data were compared between the two groups using Student's t-test or Mann–Whitney U test for continuous variables, and chi-squared test or Fisher's exact test for categorical variables, as appropriate. Receiver operating characteristic curve was used to establish the optimal cut-off value for BAL eosinophils. Variables (age,[11–14] cigarette smoking,[11–14] concurrent autoimmune features,[15] hypothyroidism,[16] BAL lymphocytes,[12] and BAL eosinophils) included in the univariable analyses were identified based on the published literature. Variables with P value <0.1 in univariable analyses were included in the multivariable logistic regression model. Cumulative survival rate was analyzed using Kaplan–Meier curves and compared using the log-rank test. Statistical analyses were performed using R software (version 3.4.3, http://www.R-project.org) and EmpowerStats software (http://www.empowerstats.com, X&Y Solutions, Inc. Boston, MA, USA). Statistical significance was set at P < 0.05.

Results Study population

A total of 298 patients diagnosed with HP were identified in the institutional database for ILDs. Ninety-six patients were excluded, including six patients because of admission after lung transplantation, 80 patients because of an indeterminate diagnosis of HP, and ten patients because chest computed tomography (CT) imaging was unavailable. Ultimately, 202 patients were enrolled in this study. There were 152 (75.2%) typical and 50 (24.8%) probable cases of HP. Patients were classified into the NFHP (n = 87) and FHP (n = 115) groups [Figure 1].

Figure 1:

Flowchart for inclusion of patients with HP. CT: Computed tomography; FHP: Fibrotic hypersensitivity pneumonitis; HP: Hypersensitivity pneumonitis; NFHP: Non-fibrotic hypersensitivity pneumonitis.

Comparison of clinical characteristics between NFHP and FHP group

Median age of the population was 60.5 (52.0–67.0) years. The proportion of female was 46.5% (94/202). Seventy-seven (38.1%) patients were smokers, including 31 (15.4%) current-smokers and 46 (22.8%) ex-smokers. Causative agents could be traced in most patients (66.8% [135/202]), but nearly one-third of the patients reported no probable inciting antigens. Avian was the most common source of causative agents, accounting for 50.4% (68/135) of the cases with identified inciting antigen. Regarding symptoms of HP, cough (88.6% [179/202]) was the most frequently reported, followed by dyspnea (82.2% [166/202]), sputum (67.3% [136/202]), and fever (29.7% [60/202]). Nine (4.5%) patients were asymptomatic and hospitalized for pulmonary abnormalities discovered incidentally. Crackles were recorded in half of the study population; however, only 4.0% (8/202) and 10.9% (22/202) patients presented with cyanosis and digital clubbing, respectively. The prevalence of chronic comorbidities is summarized in Table 1. Cardiovascular disorders including hypertension (31.7% [64/202]) and coronary heart disease (13.9% [28/202]), and diabetes (17.8% [36/202]) were relatively common. In addition, 9.4% (19/202) patients were identified as HPAF. And 5.9% (12/202) patients had a documented diagnosis of hypothyroidism. Compared with NFHP patients, FHP patients were older (62.0 [55.0–69.0] years vs. 56.0 [45.5–64.0] years, Z = −3.913, P <0.001), and experienced a longer duration of disease (12.0 [2.0–36.5] months vs. 2.0 [1.0–8.5] months, Z = 4.876, P <0.001); and more FHP patients reported dyspnea (87.0% [100/115] vs. 75.9% [66/87], χ2 = 4.163, P = 0.041) and presented with crackles (64.3% [74/115] vs. 31.0% [27/87], χ2 = 21.987, P <0.001) and digital clubbing (16.5% [19/115] vs. 3.4% [3/87], χ2 = 8.723, P = 0.003). Nevertheless, fever was more prevalent in patients with NFHP than FHP patients (39.7% [34/87] vs. 22.6% [26/115], χ2 = 6.436, P = 0.011).

Table 1 - Comparison of baseline characteristics between NFHP and FHP group. Characteristics Total HP (n = 202) NFHP (n = 87) FHP (n = 115) Statistics P values Age (years) 60.5 (52.0–67.0) 56.0 (45.5–64.0) 62.0 (55.0–69.0) –3.913∗ <0.001 Female 94 (46.5) 34 (39.1) 60 (52.2) 3.413† 0.065 Smoking status 2.674† 0.263  Non-smokers 125 (61.9) 58 (66.7) 67 (58.2)  Current-smokers 31 (15.3) 14 (16.1) 17 (14.8)  Ex-smokers 46 (22.8) 15 (17.2) 31 (27.0) Identified inciting antigen 135 (66.8) 59 (67.8) 76 (66.1) 0.067† 0.796 Sources of causative agents‡  Avians 68 (50.4) 26 (44.1) 42 (55.3) 1.665† 0.197  Mold 29 (21.5) 15 (25.4) 14 (18.4) 0.966† 0.326  Animal (except for avians) 29 (21.5) 15 (25.4) 14 (18.4) 0.966† 0.326  Plastic 3 (2.2) 0 (0) 3 (4.0) – 0.256§  Others 17 (12.6) 5 (8.5) 12 (15.8) 1.615† 0.204  Hot hub 5 (3.7) 1 (1.7) 4 (5.3) – 0.386§ Duration of disease (months) 5.0 (1.3–24.0) 2.0 (1.0–8.5) 12.0 (2.0–36.5) 4.876∗ <0.001 Symptoms  Fever 60 (29.7) 34 (39.1) 26 (22.6) 6.436† 0.011  Cough 179 (88.6) 74 (85.1) 105 (91.3) 1.916† 0.166  Sputum 136 (67.3) 55 (63.2) 81 (70.4) 1.173† 0.279  Dyspnea 166 (82.2) 66 (75.9) 100 (87.0) 4.163† 0.041 Clinical signs  Cyanosis 8 (4.0) 4 (4.6) 4 (3.5) – 0.728§  Crackles 101 (50.0) 27 (31.0) 74 (64.3) 21.987† <0.001  Digital clubbing 22 (10.9) 3 (3.4) 19 (16.5) – 0.003§ Comorbidities  COPD 6 (3.0) 3 (3.4) 3 (2.6) – 1.000§  Asthma 3 (1.5) 0 (0) 3 (2.6) – 0.261§  Hypertension 64 (31.7) 23 (26.4) 41 (35.7) 1.943† 0.163  CHD 28 (13.9) 8 (9.2) 20 (17.4) 2.786† 0.095  Diabetes 36 (17.8) 10 (11.5) 26 (22.6) 4.178† 0.041  Tumor 10 (5.0) 7 (8.0) 3 (2.6) – 0.104§  Hypothyroidism 12 (5.9) 6 (6.9) 6 (5.2) 0.250† 0.617  HAPF 19 (9.4) 9 (10.3) 10 (8.7) 0.158† 0.691 Tumor markers  CEA (ng/mL) 3.6 (2.1–5.1) 2.7 (1.6–4.1) 4.3 (2.5–6.1) –4.586∗ <0.001  CA125 (IU/mL) 19.2 (12.2–33.9) 14.9 (9.6–23.7) 23.3 (13.9–38.9) –3.823∗ <0.001  CA153 (IU/mL) 24.0 (13.0–50.2) 15.3 (9.2–30.8) 35.4 (17.1–56.0) –4.571∗ <0.001  Pro-GRP (pt/mL) 38.2 (29.2–50.6) 35.1 (27.4–42.4) 42.1 (32.1–54.2) –3.135∗ 0.002  SCC (μg/L) 0.8 (0.5–1.1) 0.7 (0.5–0.9) 0.8 (0.5–1.2) –2.199∗ 0.028  CYFRA21-1 (ng/mL) 3.9 (2.7–5.2) 3.7 (2.4–4.4) 4.2 (2.9–5.5) –2.182∗ 0.029  CA199 (IU/mL) 12.1 (7.9–22.3) 11.2 (7.2–18.3) 13.9 (9.0–23.8) –1.828∗ 0.068  NSE (ng/mL) 13.0 (11.0–16.0) 13.0 (10.0–15.0) 13.0 (11.0–16.0) –1.466∗ 0.143  CA724 (U/mL) 2.5 (1.5–4.8) 2.4 (1.5–3.6) 2.6 (1.6–6.1) –1.196∗ 0.232  AFP (ng/mL) 2.7 (2.1–3.5) 2.8 (2.2–3.9) 2.7 (2.1–3.5) –1.021∗ 0.307 BAL fluid cellular analysis  Lymphocytes (%) 39.0 (14.8–56.5) 44.3 (25.1–59.9) 31.5 (12.0–53.0) –2.880∗ 0.004  Eosinophils (%) 2.5 (1.0–6.0) 1.5 (0.5–5.0) 3.0 (1.0–6.5) –2.264∗ 0.024  Neutrophils (%) 25.0 (9.9–45.0) 19.3 (9.4–40.6) 26.5 (10.8–48.6) –1.499∗ 0.134  Macrophages (%) 21.9 (12.0–36.9) 21.0 (10.4–35.5) 22.8 (13.0–38.8) –0.912∗ 0.362 BAL lymphocytosis|| 131 (70.1) 67 (81.7) 64 (61.0) 9.454† 0.002 Pulmonary function¶ – <0.001§  Normal 107 (56.3) 58 (72.5) 49 (44.6)  Obstructive dysfunction 12 (6.3) 4 (5.0) 8 (7.3)  Restrictive dysfunction 67 (35.3) 15 (18.8) 52 (47.3)  Mixed dysfunction 4 (2.1) 3 (3.8) 1 (0.9) Diffusion dysfunction¶ 144 (79.6) 53 (68.0) 91 (88.4) 11.359∗ <0.001  FEV1 predicted (%) 83.7 ± 21.0 89.1 ± 20.8 79.6 ± 20.4 3.089∗∗ 0.002  FVC predicted (%) 85.8 ± 22.4 93.2 ± 21.2 80.3 ± 21.7 4.068∗∗ <0.001  TLC predicted (%) 75.6 ± 17.2 81.7 ± 16.8 70.9 ± 16.1 4.399∗∗ <0.001  DLCO predicted (%) 62.3 ± 21.7 68.7 ± 22.1 57.4 ± 20.2 3.609∗∗ <0.001 Small airway dysfunction 133 (72.7) 56 (70.9) 77 (74.0) 0.225∗∗ 0.635

Data are presented as n (%), mean ± standard deviation, or medium (Q1, Q3).

∗Z values.

†χ2 values.

§Calculated using the Fisher's exact test.

∗∗t values.

‡The number of cases is 135 with 59 of cases in NFHP group and 76 in FHP group.

||The number of cases is 187 with 82 of cases in NFHP group and 105 in FHP group.

¶The number of cases is 190 with 80 of cases in NFHP group and 110 in FHP group. AFP: Alpha fetoprotein; BAL: Bronchoalveolar lavage; CA: Carbohydrate antigen; CEA: Carcinoembryonic antigen; CHD: Coronary heart disease; COPD: Chronic obstructive pulmonary disease; CYFRA21-1: Cytokeratin 21-1; DLCO: Diffusing capacity of the lungs for carbon monoxide; FEV1: Forced expiratory volume in 1 s; FHP: Fibrotic hypersensitivity pneumonitis; FVC: Forced vital capacity; HP: Hypersensitivity pneumonitis; HPAF: Hypersensitivity pneumonitis with autoimmune features; NFHP: Non-fibrotic hypersensitivity pneumonitis; NSE: Neuron-specific enolase; Pro-GRP: Gastrin releasing peptide precursor; SCC: Squamous cell carcinoma antigen; TLC: Total lung capacity. –: Not available.

As shown in Table 1, serum levels of several tumor markers were higher in the FHP group than in the NFHP group, including carcinoembryonic antigen (CEA) (4.3 [2.5–6.1] ng/mL vs. 2.7 [1.6–4.1] ng/mL, Z = –4.586, P < 0.001), carbohydrate antigen (CA) 125 (CA125) (23.3 [13.9–38.9] IU/mL vs. 14.9 [9.6–23.7] IU/mL, Z = –3.823, P <0.001), CA153 (35.4 [17.1–56.0] IU/mL vs. 15.3 [9.2–30.8] IU/mL, Z = –4.571, P <0.001), gastrin-releasing peptide precursor (pro-GRP) (42.1 [32.1–54.2] pt/mL vs. 35.1 [27.4–42.4] pt/mL, Z = –3.135, P = 0.002), squamous cell carcinoma (SCC) antigen (0.8 [0.5–1.2] ug/L vs. 0.7 [0.5–0.9] ug/L, Z = –2.199, P = 0.028), and cytokeratin 21-1 (CYFRA21-1) (4.2 [2.9–5.5] ng/mL vs. 3.7 [2.4–4.4] ng/mL, Z = –2.182, P = 0.029). Compared with the NFHP group, the percentage of eosinophils in BAL were higher in the FHP group (3.0% [1.0–6.5%] vs. 1.5% [0.5–5.0%], Z = –2.264, P = 0.024). There were 70.1% (131/187) cases characterized with BAL lymphocytosis. But it is noteworthy that lymphocytosis was less pronounced in the FHP group than in the NFHP group (61.0% [64/105] vs. 81.7% [67/82], χ2 = 9.454, P = 0.002).

Regarding ventilatory impairment, 6.3% (12/190), 35.3% (67/190), and 2.1% (4/190) patients exhibited obstructive, restrictive, and mixed ventilation dysfunctions, respectively. Additionally, 79.6% (144/190) patients had diffusion impairment. The extents of restrictive and diffusion impairments were mostly mild and moderate. Small airway dysfunction was prevalent among patients with HP but did not differ significantly between the NFHP and FHP groups. Patients with FHP experienced more severe impairments of lung function than those with NFHP.

On chest CT, GGOs (91.6% [185/202]) were the most common radiographic feature, followed by reticulation (63.9% [129/202]), tractive bronchiectasis (48.5% [98/202]), mosaic perfusion (40.6% [82/202]), centrilobular nodules (35.1% [71/202]), honeycombing (22.8% [46/202]), and emphysema (22.3% [45/202]). Centrilobular nodules were more frequent in the NFHP group (26.1% [30/87] vs. 47.1% [41/115], χ2 = 9.619, P = 0.002), while GGOs (96.5% [111/115] vs. 85.1% [74/87], χ2 = 8.446, P=0.004) and emphysema (31.3% [36/115] vs. 10.3% [9/87], χ2 = 124.258, P < 0.001) were more likely to appear in the FHP patients. Non-caseous granuloma and lymphocytic infiltration were typical histopathological characteristics of HP [Table 2].

Table 2 - Comparison of features of radiology and histopathology between NFHP and FHP patients. Characteristics Total HP (n = 202) NFHP (n = 87) FHP (n = 115) Statistics P values Features of radiology  Centrilobular nodules 71 (35.1) 41 (47.1) 30 (26.1) 9.619∗ 0.002  GGOs 185 (91.6) 74 (85.1) 111 (96.5) 8.446∗ 0.004  Mosaic perfusion 82 (40.6) 42 (48.3) 40 (34.8) 3.739∗ 0.053  Reticulation 129 (63.9) 18 (20.7) 111 (96.5) 123.413∗ <0.001  Honeycombing 46 (22.8) 0 (0) 46 (40.0) <0.001§  Emphysema 45 (22.3) 9 (10.3) 36 (31.3) 12.567∗ <0.001  Tractive bronchiectasis 98 (48.5) 3 (3.4) 95 (82.6) 124.258∗ <0.001 Lesion scores on radiology  GGOs scores 1.3 (0.8–2.0) 1.5 (0.6–2.5) 1.2 (0.8–1.8) –2.203† 0.028  Reticulation scores 0.8 (0–1.4) 0 (0–0.3) 1.3 (0.9–1.9) –10.919† <0.001