Most cases of HP, whether acute or subacute, include the following histologic features in variable proportions [21]:

The bronchiolitis may include variable degrees of peribronchiolar fibrosis and hyperplasia of the bronchiolar epithelium, a characteristic but non-specific finding [22].

In some patients with insidious onset disease, emphysema may be a prominent component. Non-specific interstitial pneumonitis (NSIP), usual interstitial pneumonia (UIP) and bronchiolitis obliterans organizing pneumonia might be the sole histological expressions of the disease [20]. In patients in whom the pattern of fibrosis is consistent with UIP or NSIP, the presence of giant cells, granulomas, areas of interstitial granulomatous inflammation, or peri-bronchiolar fibrosis should suggest the diagnosis of HP.

Chest radiography is used to determine the cause of illness and rule out the others. Up to 20 % of individuals with active HP have normal chest X-rays [23].

HRCT reflects histopathologic findings with precision by showing direct (mainly centrilobular opacities and ground-glass opacities) and indirect (mainly hyperluscent area and air-trapping) images.

Ground-glass opacities are usually bilateral and symmetric, but sometimes patchy; they generally predominate in the middle part of the lungs. These opacities usually represent chronic interstitial inflammation but occasionally may be caused by fibrosis or organizing pneumonia. Another characteristic direct feature is the numerous small round centrilobular opacities, usually less than 5 mm in diameter (Fig. 29.3). These centrilobular ground-glass opacities sometimes have relatively well-defined borders and are referred to as nodules. These abnormalities represent cellular bronchiolitis, peribronchiolar interstitial inflammation, or, less frequently, focal organizing pneumonia. The third major finding is hypoattenuation and hypovascularity of scattered secondary lobules giving sometimes a mosaic aspect (Fig. 29.4). Hypoattenuating regions that persist on expiratory HRCT are indicative of air-trapping (Fig. 29.5), which is caused by bronchiolar obstruction, often favored by the presence of mucus. The combination of patchy ground-glass opacities, normal regions and air-trapping on HRCT is sometimes referred to as the headcheese sign [21]. Bronchiolar wall thickening may also occur. Lung cysts have been found occasionally and are probably caused by obstruction of bronchioles (Fig. 29.5). Focal consolidation, which often represents organizing pneumonia, is rarely present. Mediastinal lymph-node enlargement does not rule out the diagnosis of HP [17]. Pulmonary arteries are occasionally enlarged, presumably reflecting pulmonary arterial hypertension [24]. In type-1 patients, especially farmer’s lung patients, centrilobular emphysema may develop [17]. When fibrosis is present, HRCT demonstrates reticulation, mainly in the middle portion of the lungs or evenly throughout the lungs but with sparing of the extreme bases [21, 25]. Honeycombing may occur, resembling that in patients with idiopathic pulmonary fibrosis (IPF) [26]. When fibrotic HP is compared with IPF or NSIP, the CT-features favoring a diagnosis of HP are lobular air-trapping, centrilobular ground-glass opacities and absence of lower-lobe predominance [21, 26].

BAL plays a major role in the investigation of patients suspected of having HP in that a normal number of lymphocytes rules out all but residual disease [27]. BAL is considered normal when the percentage of lymphocytes is less 30 % in non-smokers and 20 % in smokers. Nevertheless, a significant increase in neutrophils can be observed when BAL is carried out shortly after exposure [28]. In this situation, the BAL sample can consist mainly of macrophages and neutrophils, with less than 30 % of lymphocytes. Soon afterwards, lymphocytes predominate. Lymphocytosis is also seen in BAL fluid from patients with a history of HP but with no recent activity, making it a very instructive tool for retrospective diagnosis. BAL lymphocytosis is, however, poorly specific as it may be seen in many other lung diseases [29]. A classic finding is a ratio of CD4 to CD8 lymphocyte subsets that is less than 1 (the normal ratio is 1.8). CD4/CD8 ratio is probably affected by the clinical form of HP, exposure to tobacco smoke, the type and dose of inhaled antigen and the time elapsed since antigen exposure. A recent multicenter study in France retrospectively assessed the diagnostic value of BAL in 139 HP cases, mainly farmer’s lung: only 34 % of the subjects showed a CD4/CD8 ratio less than 1 [30]. Hence, the usefulness of the CD4/CD8 ratio to diagnose HP is highly questionable.

The presence of circulating antibodies against the putative offending antigen(s) is useful for diagnosis. Indeed, the results of the HP study demonstrated that positive serum antibodies are a significant predictor of HP (odds ratio: 5.3; 95 % IC: 2.7–10.4) after taking antigen exposure into account [4]. The selection of antigens to be tested often needs to be determined locally according to the prevalent antigens. In Eastern France, by using a panel of antigens shown to be responsible for farmer’s lung rather than a classic standardized panel, serological tests showed high figures for both negative predictive values and specificity (from 89 to 95 % according to prevalence of the disease). In that study however, sensitivity was low with 11 out of the 31 tested HP cases negative. Additional investigations by microbiological samplings in the environment of these negative subjects allowed to identify antigens not present in the initial panel. Eight of the 11 false-negative HP cases became positive after testing these unusual antigens. These results highlight the importance of (1) the use of antigens proven to be representative of exposure, (2) the knowledge of unusual antigenic exposure and (3) sampling the patient’s environment.

Several methods of determining precipitins or total IgG antibodies have been described. Enzyme-linked immunosorbent assay (ELISA) is largely used because this technique is standardized. The consistency of 4 serological techniques (electrosyneresis, Ouchterlony double-diffusion, ELISA and Western-blot) was recently evaluated in France. Electrosyneresis on cellulose acetate was the most relevant diagnostic tool to discriminate farmer’s lung from healthy exposed farmers (sensitivity 87 %, specificity 100 %) [31].

All of these considerations concern tests against organic antigens: microorganisms, moulds or animal proteins. In the case of HP due to chemical substances, isocyanates for example, serological tests have not been demonstrated to be useful or even interpretable.

Pulmonary function tests (PFTs) have no discriminative properties in differentiating HP from other diffuse infiltrative lung diseases [4]. At the time of diagnosis, PFTs can be normal, or classically show a restrictive pattern (especially in type 2 HP) or sometimes reveal an obstructive pattern. In chronic disease, the pattern can be restrictive, but at least in farmer’s lung, and probably in other HP due to microorganisms, the most frequent profile is an obstructive defect resulting from chronic airway involvement or emphysema [32]. Decreased carbon monoxide diffusion capacity (DLCO) is generally considered to be consistently present in HP [33, 34]. But if we strictly define abnormal values as a DLCO <80 % predicted, some HP cases have normal results [3]. Importantly, the improvement of DLCO is very low, for up to 2 years in the large follow up study by Kokkarinen et al. [34]. This slow recovery of DLCO is in agreement with the slow recovery of both histological changes and lymphocytic infiltration in HP. Hypoxemia is common, especially with exercise. Hypoxemia (at rest or at exercise) and lung volume defect recover quickly, within a few months after an acute episode [34, 35]. These abnormalities are nevertheless non-specific since similar changes are found in most interstitial lung diseases (ILDs).

The objective of this prospective multicenter cohort study was to develop a clinical prediction rule for the diagnosis of active HP. Such a rule aims at helping clinicians to reach a more accurate estimate of probability of HP and to decide whether further investigation is needed to rule it either in or out [36]. Consecutive adult patients presenting with a pulmonary syndrome for which active HP was considered whatever the possible diagnosis were included in the study. The investigators had to classify each patient as HP or non-HP. The final diagnosis was based on BAL findings, HRCT and, when necessary, other procedures including surgical lung biopsy.

Multivariate analysis identified 6 significant independent predictors of HP (Table 29.2). The clinical prediction model produces an equation that expresses the probability of HP as a function of the statistically significant variables. From this equation, a table of probability for combinations of predictors was constructed (Table 29.3). For instance, farmers presenting with recurrent episodes of respiratory symptoms, inspiratory crackles and positive testing for the corresponding precipitating antibodies, the probability of HP would be 81 % (Table 29.3). A patient with progressive dyspnoea and inspiratory crackles as the only criteria in favor of HP would have a probability of less than 1 % [36]. Further investigations such as HRCT and/or BAL would be indicated in the first case; diagnosis would be excluded in the second.

These propositions cannot be considered as actual diagnostic criteria in that they have not been validated. They represent a tool for standardized diagnostic management to be used by specialists of interstitial lung diseases (Table 29.4).

Their use requires:

Tracking the culprit antigen(s) is the most important thing to do when a patient presents with a respiratory syndrome for which HP is considered among the possible diagnoses. When the antigenic source is identified, the probability that a patient is suffering from HP is multiplied by 18 as compared with a patient for whom no exposure is found [4]. The presence of serum precipitins (or IgG) directed against the antigen(s) multiplies the probability of HP by 5 again [4]. Furthermore, when HP is diagnosed, identifying the precise etiology facilitates the only efficient treatment – avoiding the antigenic source. It is therefore crucial to be able to refer to a comprehensive and easily understood list of rare or newly identified causes.

However, identification of a specific antigen should not be a sine qua non for diagnosis, because new knowledge about antigens is generated from identifying new sources of exposure [2]. Some authors speak of HP from “unidentified causes”. In our opinion, it would be more appropriate to say HP from “causes not yet identified”.

In addition to the unusual etiologies or etiological circumstances presented in Table 29.5, we present below a brief overview of recently discovered or poorly known causes or circumstances of HP.