The Diagnosis of Hypersensitivity Pneumonitis and the Role of Lung Biopsy

Abstract

Hypersensitivity pneumonitis (HP) refers to a collective noun of diffuse lung diseases encompassing some degree of bronchiolar and interstitial granulomatous inflammation that results from persistent inhalation exposure and consequent immune sensitization to a large potential diversity of (predominantly) organic antigens in predisposed individuals. In suspected cases of HP, forceps transbronchial lung biopsy (TBLB) has been traditionally performed on a case-by-case basis along with bronchoalveolar lavage. This option has been subject to some debate and its use is more restrained in the presence of a chronic fibrotic form of HP—where surgical lung biopsy is classically recommended in the face of the need for a more reliable differentiation from fibrotic idiopathic interstitial pneumonias. We intended to assess the diagnostic contribution of conventional TBLB in the combined multidisciplinary diagnosis of an HP patient cohort. A retrospective evaluation of all the diagnostic elements and level of confidence from all HP cases followed in an interstitial lung disease ILD outpatient clinic of a district hospital center (Centro Hospitalar do Baixo Vouga), from June 2015 to August 2019, and simultaneously evaluated in a multidisciplinary team discussion of the same hospital, comprising an interstitial lung disease dedicated lung physician, a chest radiologist, 2 rheumatologists, and a pathologist. We identified 78 patients (mean age: 70.5 y, interquartile range: 58.5 to 78.0) with a slight female predominance. Most of the patients (61.5%) had chronic/fibrotic HP. The most frequently identified inducing antigens were avian antigens in 59.0% of cases, followed by molds in 20.5%. Of the 72 patients who underwent bronchofibroscopy, 36.1% (n=26) conventional TBLB performed, predominantly in the segments of the right lower lobe with an average number of 3.9 biopsies (SD±1.4) accomplished per patient. In 50.0% of the cases submitted to TBLB, the biopsies showed representative material with histologic features (definite or supportive) which had some degree of contribution for the diagnostic discussion. Among the patients where TBLBs were not performed or whose results were found to be devoid of significant findings, 73.1% were still diagnosed as HP without the need for surgical video-assisted thoracoscopic lung biopsy/transbronchial lung cryobiopsy (VATS/TBLCB) on the grounds of other diagnostic elements; 15.4% of patients were diagnosed with HP after a VATS/TBLCB procedure. Lastly, around 11.5% of patients were considered to have an unacceptable risk for VATS/TBLCB but, on the basis of clinical, radiologic, and immunologic elements received a multidisciplinary provisional diagnosis still with a reasonable level of confidence. Regarding complications with TBLB, there were 2 cases of moderate bleeding (7.6%) and 1 pneumothorax (3.8%) that did not require drainage. Notwithstanding its limitations, TBLB can still have a role in the diagnostic workup of HP, namely in acute/inflammatory HP, adding definite or supportive histologic information for multidisciplinary discussion in up to 50% of cases. TBLB can augment diagnostic yield at the expense of only a minimal increase of risk, as it is a universally available technique that can be performed along with bronchoalveolar lavage. This has important implications, particularly in centers devoid of TBLCB, as a surgical biopsy can be avoided in approximately half of patients who are eventually diagnosed with HP.

KEY POINTS

Notwithstanding its limitations, transbronchial lung biopsy can still have a role in the diagnostic workup of hypersensitivity pneumonitis (HP), namely in acute/inflammatory HP, adding definite or supportive histologic information for multidisciplinary discussion in up to 50% of cases. TBLB can augment diagnostic yield at the expense of only a minimal increase of risk, as it is a universally available technique that can be performed along with bronchoalveolar lavage (BAL). This has important implications, particularly in centers devoid of transbronchial lung cryobiopsy (TBLCB), as a surgical biopsy can be avoided in approximately half of patients who are eventually diagnosed with HP.

INTRODUCTION

HP refers to a collective noun of diffuse lung diseases encompassing bronchiolar and interstitial granulomatous inflammation that results from persistent inhalation exposure and consequent immune sensitization to a large potential diversity of (predominantly) organic antigens in predisposed individuals.^1–3

The diagnosis of HP remains a challenge due to its remarkable clinical polymorphism and absence of clear widely accepted diagnostic criteria.^4,5 Nevertheless, in the current paradigm of multidisciplinary, integrative, and comprehensive diagnostic approach, histologic evaluation is still necessary in a relevant proportion of cases of HP.⁶

ROLE OF BIOPSY IN HP DIAGNOSIS

Conventional Transbronchial Lung Biopsy

Conventional transbronchial lung biopsy (TBLB) is a type of lung biopsy that is carried out with a flexible bronchoscope in a patient with conscious sedation and into a segmental bronchus without direct visualization of the lesion. Some centers realize this technique with radiologic guidance, but diagnostic yield is no different between techniques (Table 1—based on Raghu et al⁷ and Chami et al⁸). The flexible bronchoscope is introduced as far as possible through the chosen segmental/subsegmental bronchus in order to perform a TBLB; then, the biopsy forceps are introduced and advanced as possible; at that moment the forceps is retracted 1 to 2 cm from the pleura and the patient should do a breath in as the forceps is opened. After this point, the patient is instructed to exhale and the forceps is advanced and closed. The forceps is then pulled out under direct visualization. This process is repeated 5 to 6 times to increase the diagnostic yield.^9,10

TABLE 1 - Characteristics of TBLB, SLB, and TBLCB^7,8

	TBLB	SLB	TBLCB
Histopathologic yield
In interstitial lung diseases	37% (95% CI: 32%-42%)	96%-98% (95% CI)	81% (95% CI: 75%-88%)
In diffuse lung diseases	68% (95% CI: 50-86%)		82% (95% CI: 73%-90%)
Bleeding
Moderate to severe	0% (95% CI: 0%-1%)	1% (95% CI: 0%-1%)	11% (95% CI: 7%-15%)
Severe			0% (95% CI: 0%-1%)
Pneumothorax	7% (95% CI: 2%-13%)	4% (95% CI: 2%-5%)	11% (95% CI: 9%-14%)
Respiratory exacerbation	No cases	2% (95% CI: 1%-3%)	Rare (95% CI: 0%-1%
Periprocedural mortality	No cases	2% (95% CI: 1%-3%)	Rare (95% CI: 0%-1%)

CI indicates confidence interval; SLB, surgical lung biopsy; TBLB, conventional transbronchial lung biopsy; TBLCB, transbronchial lung cryobiopsy.

In suspected cases of HP, despite its limited value in the evaluation of chronic fibrotic form of HP—where surgical lung biopsy (SLB) is classically recommended for a more reliable differentiation from fibrotic idiopathic interstitial pneumonias, forceps TBLB, although debatable, has been traditionally performed on a case-by-case basis along with BAL.⁴ The addition of TBLB to BAL has significantly improved the diagnostic yield in granulomatous diseases with bronchiolocentric distribution, regardless of the BAL lymphocyte cutoff used.^11,12

However, several factors may limit its diagnostic accuracy in the study of diffuse lung disease such as sample size, possible crush artifacts, distribution of changes, and difficulty in sampling tissue beyond the peribronchial sheath.^13–15

It is proposed that in the absence of relevant histopathological features retrieved from conventional forceps TBLBs, TBLCB or SLB should be obtained.⁶

Surgical Lung Biopsy

A SLB is a key procedure in the study of patients with interstitial lung diseases (ILDs). It is generally considered or patients where a confident diagnosis cannot be reached after MTD discussion with other diagnostic workup. Those patients must not be present an unacceptable risk for postoperative complications and should be presumed to benefit from histological pattern-based diagnosis.

Although SLB has some associated risks, partially associated with patient´s impaired lung function, the morbidity and mortality rate can be biased by case selection, type of intervention [open surgery vs. video-assisted thoracoscopic lung biopsy (VATS) surgery with a lower risk in the VATS surgery] and center’s expertise.

SLB has the advantage of providing a larger biopsy size and is particularly valuable in distinguishing chronic fibrotic HP from other causes of usual interstitial pneumonia, including idiopathic pulmonary fibrosis, and nonspecific interstitial pneumonia. The reason for this distinction is the histologic findings in some patients with suspected HP, which overlap with usual interstitial pneumonia and nonspecific interstitial pneumonia.¹⁶

Transbronchial Lung Cryobiopsy

TBLCB is a procedure that uses a long flexible cryoprobe with metal tips than can be advanced through the channel of the bronchoscope. The tip quickly freezes an area of peripheral lung by way of the Joule-Thomson effect: the compressed gas is released at a high flow rate into the tip of the probe, where it expands and cools in seconds to a temperature of −79°C (using carbon dioxide) or –89°C (using nitrous oxide).

The bronchoscope and cryoprobe with attached frozen biopsy are then removed en-bloc from the airway, because cryobiopsy is significantly larger than the working channel of the bronchoscope. The specimen is then submerged in saline to release the biopsy. The biopsy specimens are larger than those obtained with TBLB and usually do not have the crush artifacts.¹⁷

TBLCB has proved to have a high diagnostic yield in the diagnosis of HP,¹⁸ but it is still hindered by a relatively scarce accessibility (the studies reporting similar diagnostic yield are only valid for some experienced centers), a significant pneumothorax rate, and a relative contraindication in patients with severe lung function impairment (predicted diffusion capacity <35% and predicted forced vital capacity <50%).¹⁹

In the study discussed below, we tried to find out whether there is still some role for the conventional forceps TBLB in the diagnostic workup of HP.

Our Experience With Conventional Transbronchial Biopsy in HP

This study assesses the diagnostic contribution of TBLB in the multidisciplinary diagnosis of an HP patient cohort.

We made a retrospective assessment of the diagnostic elements and the level of confidence from all cases with an HP diagnosis followed in an ILD outpatient clinic in a district hospital center (Centro Hospitalar do Baixo Vouga), from June 2015 to August 2019, and simultaneously evaluated in a ILD multidisciplinary team discussion, comprising an ILD dedicated pulmonologist, a chest radiologist, a pathologist and 2 rheumatologists.

Data were collected from the patient’s records, comprehensive clinical data, chest high-resolution computed tomography (HRCT), lung function tests, BAL, and other immunologic workup and biopsy reports (TBLB, TBLCB and VATS). Special attention was given to TBLB, namely number attempted per patient, anatomic location targeted, and complications. All the multidisciplinary team meeting records regarding the decision for advanced sampling procedures (VATS or TBLCB) and the diagnostic confidence level achieved were reviewed.²⁰

TBLBs were performed with Olympus forceps, by different physicians, during a diagnostic flexible bronchoscopy under conscious sedation with iv midazolam and, generally, targeting a Richmond Agitation-Sedation Scale (RASS)-3 sedation level. Preprocedural evaluation protocol demanded suspension of anticoagulant and antiplatelet drugs. Iatrogenic bleeding after the procedure was classified based on the bronchoscopy reports: as mild, when estimated as <20 mL and self-limited; moderate, when estimated to be 20 to 100 mL and/or requesting an active hemostatic endobronchial procedure; and severe, when estimated with a blood volume of >100 mL and/or requesting advanced airway control. Iatrogenic pneumothorax and the need for needle exsufflation or chest drain were also acknowledged.

TBLB biopsy fragments were 10% formalin fixed, paraffin embedded, and hematoxylin and eosin stained for analysis. The samples were blinded for review regarding lung representativity and relevant histologic changes depicted. According to the literature, a differentiation was made concerning the representation of lung parenchyma in the samples and the presence/absence of supportive features of HP (more or less relevant).^21,22

Statistical Analysis

Qualitative variables are expressed as absolute value and percentage, whereas quantitative variables are expressed as means and SD for normally distributed data and median and interquartile range (IQR) for non-normally distributed data. The descriptive analysis relied on IBM SPSS Statistics 25.

RESULTS

We identified 78 patients (mean age: 70.5 y, IQR: 58.5 to 78.0) with an established HP diagnosis in the predefined timeframe after multidisciplinary discussion. There was a discrete female gender predominance (n=46, 59.0%) and a higher mean age in male patients (75.0 y, IQR: 63.3 to 75.0) compared with female patients (mean age of 65.5 y, IQR: 55.8 to 76.5).

Of the 78 HP patients, 61.5% presented chronic/fibrotic HP and the remaining cases showed inflammatory features, without clear established fibrosis on HRCT. The most frequently identified inducing antigens were avian antigens (parakeet, canary, cockatiel, pigeon, chicken, blackbirds, and other exotic birds) representing 59.0% of the cases, followed by molds (20.5%) and a set of other antigens such as isocyanates, bacteria, antigen combination, and cases without identifiable causal antigen (Table 2).

TABLE 2 - HP Cases by Causal Antigen Source

	n (%)
Acute/inflammatory HP	30 (38.5)
Avian-induced	19 (24.4)
Mold-induced	6 (7.7)
Avian and/or mold-induced	3 (3.8)
Isocyanate-induced	1 (1.3)
No identifiable antigen	1 (1.3)
Chronic/fibrotic HP	48 (61.5)
Avian-induced	27 (34.6)
Mold-induced	10 (12.8)
Avian and or mold-induced	6 (7.7)
Isocyanate-induced	3 (3.8)
Other antigens	1 (1.3)
No identifiable antigen	1 (1.3)

HP indicates hypersensitivity pneumonitis.

Of the 72 patients who underwent flexible bronchoscopy, 36.1% (n=26) had TBLB performed, mostly in right lower lobe segments with a mean of 3.9 biopsy fragments (SD±1.4) sampled per patient. Six patients refused to undergo bronchoscopy or had previously performed the examination in another hospital setting.

In those 26 patients, 26.9% had inadequate samples due to insufficient representativeness of the lung parenchyma, 23.1% showed normal parenchyma and the remaining 50.0% had representative material with histologic features (definite or supportive) that contributed for the diagnostic discussion (Fig. 1). The most prevalent significant findings were: lymphoplasmacytic inflammatory infiltrate (53.8%), bronchiolocentric fibrotic changes (38.5%), intra-alveolar xanthomatous macrophages (30.8%), alveolar septal thickening (30.8%), cellular bronchiolitis (15.4%) and, with 7.7% each, loose granulomas, small foci of organizing pneumonia/Masson bodies, multinucleated giant cells and fibroblast foci. The diagnostic yield for definite and/or supportive histologic features was 38.4% for the subgroup of chronic/fibrotic HP and 61.5% among acute/inflammatory HP patients submitted to TBLB. Globally, 19.2% of cases submitted to TBLB presented highly relevant supporting features (loose granulomas, small foci of organizing pneumonia, multinucleated giant cells or cellular bronchiolitis, accompanied or not by other less relevant features) and another 30.8% were found to have one or more supporting features (lymphoplasmacytic infiltration, bronchiolocentric fibrosis, intra-alveolar foamy macrophages and fibroblastic foci).

FIGURE 1: Representativeness and histopathologic features of the TBLB performed. TBLB indicates conventional forceps transbronchial lung biopsy.

In 50% of the cases submitted to TBLB, a multidisciplinary confident diagnosis was reached without the need for a sequential advanced biopsy procedure. Among the patients where TBLBs were not performed or whose results were devoid of significant findings (Fig. 2): 73.1% still were diagnosed as HP without the need for VATS/TBLCB on the grounds of other diagnostic elements (significant antigen exposure, HRCT pattern, BAL lymphocytic profile, elevated titers for antigen-specific IgGs, results of home-visiting or domestic aerobiological study); 15.4% of patients were diagnosed with HP after a VATS/TBLCB procedure; and lastly, ~11.5% of patients were considered to have an unacceptable risk for VATS/TBLCB although they still received a multidisciplinary provisional diagnosis with an acceptable level of confidence on the grounds of clinical, radiologic, and immunologic elements, according to previously suggested diagnostic framework.²⁰

FIGURE 2: Diagnostic diagram of HP cases where TBLBs were devoid of diagnostic contribution or simply not performed. BAL indicates bronchoalveolar lavage; HP, hypersensitivity pneumonitis; HRCT, high-resolution computed tomography; TBLB, conventional forceps transbronchial lung biopsy; TBLCB, transbronchial lung cryobiopsy; VATS, video-assisted thoracoscopic lung biopsy.

Regarding complications with TBLB, there were 2 cases of moderate bleeding (7.6%) and 1 pneumothorax (3.8%) that did not require drainage.

DISCUSSION

TBLBs were performed in only 36.1% of our HP cohort probably for 2 main reasons. First, in the patients with chronic/fibrotic (majority) HP, there is a low probability of adding diagnostic value with this sampling method. These patients were preferably referred for VATS (frequently) or TBLCB (occasionally) when a biopsy was judged necessary. Secondly, cases where TBLB execution was dismissed due to an already very high initial clinico-radiological suspicion for HP, where an expectable compatible BAL lymphocytosis would be sufficient to support an integrative diagnostic assumption. We found that in half of the cases with a previous HP suspicion where TBLBs were performed, representative lung samples with relevant histologic abnormalities were obtained. Although rarely providing a pattern-based diagnosis, several histological features (eg, foci of organizing pneumonia, loose granulomas, bronchiolitis, multinucleated giant cells, marked lymphoplasmacytic infiltration, bronchiolocentric fibrosis, intra-alveolar xanthomatous macrophages, etc.) proved to be pertinent in the setting of a multidisciplinary team discussion. The good safety profile of TBLB is, again, perfectly highlighted by the present study.

In cases of ILD where histological analysis is required for a reliable diagnostic embodiment, VATS SLB remains the “gold standard.”^23,24 However, as known, this sampling method is not achievable in an important fraction of patients due to unacceptable risk (advanced age, poor lung function, comorbidities) or just due to patient refusal. Despite its unquestionable relevance and growing affirmation, TBLCB stands as the closer alternative to SLB but it does not replace it. Besides, it also requires a readjustment of resources (starting with the availability of an anesthesiologist in the room), has a relevant risk of pneumothorax and remains a technique not available in-house in the majority of hospitals.^2,20,25–28

In our study, as in a previous work, 50% of cases submitted to TBLB had histologic abnormalities important for the diagnosis (we found definite features in 19.2% of cases and supportive findings in 30.8%). In these patients, a SLB could be avoided. Previous literature has shown characteristic findings of HP in 11% to 44% of TBLB.^13,29,30 Actually, 1 study demonstrated an increase in diagnostic yield from 23% to 24% to 52% to 58% with the addition of TBLB to BAL and found a 46.6% yield specifically from TBLB among the subgroup of HP patients with <20% lymphocytes on BAL.¹¹ Therefore, it can be argued that performing TBLB as a first histologic approach can still be valuable specially in the clinical setting of inflammatory HP on HRCT and in centers devoid of TBLCB. It must be emphasized that a “negative” or “inadequate” TBLB should not be used to rule out HP, unless it establishes a clear alternative histopathological diagnosis.

Regarding complications associated with TBLB, the pneumothorax rate was (3.8%) in line with the previous literature (1% to 7%),^7,8,31 whereas the bleeding rate was slightly higher (moderate bleeding in 7.6%) than previous reports (0 to 5.7%),^7,8,32,33 which can be justified by our sample size and absence of severe bleeding.

There are several limitations to this study. First, this was a retrospective study from a single center. Second, there is a selection bias as several chronic/fibrotic HP cases were not submitted to TBLB, as stated, due to the preconceived notion of low additional value. Third, the value of BAL itself for the diagnostic assumption was not accessed. Finally, the absence of a control group with other non-HP diagnoses (including unclassifiable disease) who also performed TBLB hindered a sensitivity and specificity analysis, allowing only a diagnostic yield estimate. Nonetheless, all cases were discussed by the same ILD multidisciplinary team after a thorough diagnostic assessment.