Abstract
Chronic hypersensitivity pneumonitis (CHP) can be difficult to differentiate from other interstitial lung diseases (ILD). To determine the diagnostic usefulness of a provocation test (PT), 17 patients with CHP induced by avian antigens, 17 with other ILD, and five healthy control subjects were challenged with pigeon serum. After PT, an increase in body temperature (BT) and a decrease in FVC, PaO2 and SaO2 % were observed in all patients with CHP and in three with ILD. No reaction was noticed in healthy subjects. ROC curves showed that for FVC the best cut point was a drop of 16% displaying sensitivity (S): 76%, specificity (SP): 81%, positive predictive value (PPV): 81%, and negative predictive value (NPV): 83%. For a drop of 3 mm Hg in PaO2 or 3% SaO2 , S was 88% for both, SP was 82 and 86%, PPV was 81 and 82%, and NPV was 82 and 86%, respectively. An increase of BT > 0.5o C showed S, 100%; SP, 82%; PPV, 100%; NPV, 86%. A univariate regression analysis confirmed that changes in BT and FVC are predicting values of CHP: RR, 82.5 (CI, 10.43 to 651.76) and 1.21 (CI, 1.06 to 1.36). There were no challenge test complications. These findings suggest that PT is a useful tool for diagnosis of CHP.
The clinical behavior of chronic hypersensitivity pneumonitis (CHP) depends on a number of genetic and environmental variables. Usually, acute hypersensitivity pneumonitis results from intermittent and intense exposure to the causative antigen in the domestic, occupational, or atmospheric environment. The subacute and chronic forms result from a less intense but continuous exposure to inhaled antigens, usually in the domestic environment. The characteristic example is the exposure to domestic caged birds where the chronic inhalation of avian antigens from the bird droppings or the bloom results in hypersensitivity pneumonitis (1). This pattern of exposure leads to a clinical course characterized by progression to diffuse lung fibrosis and chronic respiratory insufficiency (2). In Mexico, chronic pigeon breeder's disease is the most common form of hypersensitivity pneumonitis, often leading to diffuse lung fibrosis, disability, and premature death (2-4). In general, approximately 25% of the patients die in the first 5 yr after diagnosis (3). Differential diagnosis with other forms of interstitial lung diseases (ILD), particularly with idiopathic pulmonary fibrosis (IPF), may be a problem mainly because of the extensive practice of keeping birds at homes (3). Consequently, tissue examination is required to identify the features consistent with CHP (4). Still, a subgroup often remains where it is difficult to rule out CHP (3). Actually, from 125 patients with ILD studied prospectively in our Institute, 30 revealed long exposure to birds, but a biopsy with a pattern compatible with usual interstitial pneumonia and without histologic indicators of HP (3). Therefore, a number of patients with ILD remain without a clear diagnosis even after an open lung biopsy, and very likely they may have a chronic form of hypersensitivity pneumonitis.
In this context, a clinical diagnostic test capable of making unnecessary the histologic evaluation and capable of differentiating patients with CHP from patients with IPF or other ILD, even in the end stages of the disease, might be very useful. The importance of distinguishing CHP from IPF is related to the different prognosis and therapeutical approach. Patients with IPF have a shorter survival than do patients with CHP, although when the extent of fibrosis is similar the survival in both disorders is quite similar (3). Because the persistence of exposure to the offending antigen has been considered a major factor for progressing to fibrosis in CHP, once the diagnosis has been made, recommendations to avoid further exposure should follow. Unfortunately, the different clinical skills so far used to get a certain diagnosis of CHP seem to be as yet incomplete, and often lung biopsy is necessary. Although there are some reports about the utility of challenge tests for diagnosis of CHP, most of them lack reliability in the methods. Therefore, although provocation tests are widely used in asthma, their role for diagnosis in CHP has not yet been fully established. Inhalation tests, in which patients with suspected CHP are given an inhalation challenge with the putative antigen under controlled laboratory conditions, have been used occasionally to assist in confirming the diagnosis. However, their role in the routine clinical practice is unclear. In the present study, we aimed to determine the utility of a provocation test (PT) with avian antigens in terms of its sensitivity and specificity for diagnosis in patients with chronic pigeon breeder's disease.
Between June 1993 and March 1995, 117 patients with interstitial lung diseases were hospitalized at the National Institute of Respiratory Diseases in Mexico City. Seventy-five patients with different ILD and five healthy subjects were interviewed to determine the possibility to be included in this study. Thirty-five were excluded because they were not candidates for lung biopsy because of their high risk for surgery. Two subjects refused to enroll in the study because of the provocation test. Thus, 42 subjects were challenged; however, in three of them it was not possible to establish a definitive diagnosis, and they were excluded. Although these patients had exposure to birds and the specific antibodies against avian antigens were positive, the pathologist was not able to clearly distinguish changes compatible with either CHP or with other specific interstitial lung diseases.
Three different groups of patients were studied. One group was formed by 17 patients in whom pigeon breeder's disease (PBD) had been diagnosed, another group was formed by 17 patients with ILD different from CHP (ILDO), and the last group was formed by five exposed but asymptomatic subjects (AS). The diagnoses of PBD and ILDO were made independently of the results of the challenge test.
This study was designed according to guidelines of a diagnostic test (5). The steps followed to determine a positive provocation test are shown in Figure 1. Patients suspected of having an ILD because of clinical and radiologic findings were submitted to our study. Changes in the following clinical, imnmunologic, radiologic, and functional parameters were used to analyze the response to the avian antigens: FVC, TLC, residual volume (RV), diffusing capacity (Dl CO), oxygen saturation (SaO2 ), arterial blood gas determinations (PaO2 , PaCO2 ), white blood cell count (WBC), immunoglobulins, C reactive protein, rheumatoid factor, and clinical features (body temperature, presence of malaise, dyspnea, and bronchospasm). An explanation of the rationale of the study was given to the patients, and the patients were then challenged with avian antigen by an investigator blinded to the final diagnosis. After the provocative test, the same variables were evaluated. When all patients were challenged the variables were entered in a statistical analysis (paired t test) to determine the delta pre-post challenge. Only variables that had a significant change from baseline (p < 0.05) were further analyzed by using receiver operator characteristics (ROC) curves in order to find the cut points indicating the best sensitivity and specificity.
Fig. 1. Sequence of steps followed to determine the response to the inhalatory challenge. After the provocation test, all the variables (see Methods) were analyzed by paired t test to determine the delta pre-post challenge. Only variables with a significant change from baseline (p < 0.05) were further analyzed by using receiver operator characteristic (ROC) curves in order to find the cut points indicating the best sensitivity and specificity.
[More] [Minimize]Informed consent was obtained from each subject, and the protocol was approved by the ethical committee of the Institute.
A record of clinical and spirometric measurements was carried out during the following 24 h of the provocation test. At the end of this period pulmonary function tests were repeated. In order to complete the final diagnosis, patients followed the routine protocol of study for patients with ILD by a group of pulmonologists who were not associated with the present investigation. Although they used the baseline pulmonary function tests to complete the diagnosis, they were not aware of the results of the challenge test and were asked not to ascertain them. For diagnosis to be considered as definitive, the patients should have been submitted to open lung biopsy.
Diagnosis of PBD. Diagnosis of CHP induced by avian antigen was obtained according to international criteria, including: (1) exposure to pigeons preceding disease, (2) shortness of breath with partial improvement upon avoidance of the avian antigen exposure, (3) bilateral crackles in both pulmonary bases, (4) bilateral reticulonodular shadowing with some degree of ground-glass attenuation on chest radiographs, (5) restrictive lung pattern, (6) hypoxemia at rest worsening during exercise, (7) positive serum antibodies against avian antigen (AvAbs), and (8) lung histology compatible with HP (2, 6, 7). Briefly, the tissue samples showed diffuse interstitial inflammation of mononuclear predominance, mainly lymphocytes, and frequent multinucleated giant cells in terminal and respiratory bronchioles, as well as in the interalveolar walls. Foamy macrophages were seen in the alveolar spaces, and small and loosely arranged granulomas were observed in the interstitium. Biopsy cultures were negative for bacteria, mycobacteria, and fungi, and no changes suggestive of any other ILD were found.
Diagnosis of other ILD (ILDO). Diagnosis of idiopathic pulmonary fibrosis (n = 13) was made following the same clinical, radiologic, and functional criteria as for CHP. In all selected patients, AvAbs had to be absent. The patients had neither a known cause for ILD nor environmental exposure known or suspected to provoke ILD. Furthermore, the morphologic study of the lung sample obtained by open lung biopsy showed the characteristic features of IPF. The microscopic assessment included patchy alveolar septal fibrosis and interstitial inflammation consisting mostly of mononuclear cells, but also with neutrophils and eosinophils, a variable macrophage accumulation in the alveolar spaces, and cuboidalization of the alveolar epithelium. In addition, biopsies lacked granulomas, vasculitis, microorganisms, and inorganic material by polarized light microscopy (8). Diagnoses of other interstitial lung diseases (n = 4) were made according to conventional criteria and primarily based upon the histopathologic study (9-13). Patients included displayed eosinophilic granuloma, pulmonary microlithiasis, lymphangioleiomyomatosis, and bronchiolitis obliterans organizing pneumonia (BOOP).
All patients with HP or ILDO were symptomatic when the inhalatory test was performed. Thirteen of the 17 patients with ILDO had the antecedent of pigeon exposure, but without a cause-effect relationship, and without specific circulating antibodies.
Antigen-exposed subjects without disease. They were relatives of patients with PBD included in this study. They had exposure to birds and specific serum precipitins against avian antigen but no clinical, radiologic or functional respiratory abnormalities.
Presence of fever, cough, and general discomfort were evaluated before the PT and hourly during the 24 h after the test. Shortness of breath was assessed before and 8 h after PT by using a visual analog scale (VAS), consisting of horizontal line 100 mm in length with descriptors as anchors for the extreme magnitudes (14). Temperature was measured with a body thermometer. General discomfort was assessed by the presence or absence of headache, shivers, arthralgias, or simply malaise.
Pigeon exposure questionnaire. Risk factors for pulmonary diseases, including exposure to pigeons, were recorded by using a standardized questionnaire. For this report, only the most recent exposure preceding symptoms was considered. The magnitude or intensity of exposure was arbitrarily graded according to the number and closeness of the birds as follows: no exposure at all; mild exposure when patients were exposed to five pigeons or less, all of which were kept outside the home; moderate exposure when they had six to 10 birds in the yard or had less than five birds in the home but not in the bedroom; severe exposure when patients had more than 10 birds in the yard, more than five into the home, or at least one in the bedroom. In addition, the questionnaire sought out symptoms related to the exposure to birds. All the patients denied have had systemic symptoms such as malaise, fever, and rhinitis related to the pigeons. Every 2 h after the challenge test, either positive or negative, the same questions were asked.
Laboratory tests. White blood cell count (WBC) and its differential, immunoglobulins (A, E, M, and G), and specific serum precipitins against avian antigen (AvAbs) were measured before and 24 h after the PT. Immunoglobulins were measured by electrophoresis. The AvAbs were assessed by ELISA according to a technique developed in our hospital, and the data were expressed as optical densities (OD) (15). With this technique the cut point to consider a test as positive is 0.21 OD or higher.
Specific serum IgE to pigeon droppings and budgerigar feathers were also measured (Allergosorbent test system for pigeon droppings-specific IgE antibodies and budgerigar feathers-specific IgE antibodies: Allercoat Rapid East, Compact Disc Pac; Sanofi Diagnostics Pasteur, Chaska, MN). Values were reported in Allercoat East units (AEU/ml) as recommended by the manufacturer.
Spirometry. FEV1 and FVC were obtained before the challenge test and after at hourly intervals for a total of 24 h with a portable spirometer (Pony Cosmed; Italia). We followed the American Thoracic Society criteria for the performance test (16).
Lung volumes and diffusing capacity. TLC, residual volume (RV), and Dl CO were obtained before and 24 h after the PT. Volumes were obtained by using a body plethysmograph (E. Jaeger, Würzburg, Germany). All subjects were previously made familiarized with the equipment before the tests. For Dl CO manuevers, at least two tests within 5% or less of variability were considered to be acceptable. Patients with FVC less than 1 L and/or unable to follow the instructions or to hold their breath for 10 s, were not submitted to this test. Before the challenge test, FEV1, FVC, TLC, RV, and Dl CO were made by duplicate determinations within a 24-h interval. The coefficient of variation and intraclass correlation coefficient for FEV1 and FVC were 3% and 0.98, respectively; for TLC, RV, and Dl CO, these were in the range of 5 to 12% and 0.90 to 0.92, respectively. For all the PFT we used reference values reported by Quanjer (17).
Arterial blood gas determinations. These were obtained through a radial puncture while patients breathed room air for at least 30 minutes and measured in a gasometer (IL 1310; Instrumentation Laboratories, Lexington, MA) before and 24 h after the PT.
Oxygen saturation. Arterial Po 2 was measured with a pulse oxymeter (Novametrix Medical Systems Inc.) in two different conditions both before and after the PT. First, while breathing room air and then while breathing supplementary oxygen until an oxygen saturation > 90% was reached.
Antigen. The same pooled pigeon sera used for ELISA was utilized as antigen for the PT. Sera were obtained through cardiac puncture from five to 10 pigeons using an aseptic technique. The samples were further sterilized by passing through a 0.22-μm filter. To eliminate complement, serum was additionally heated to 56° C for 30 min. In order to homogenize the nebulized antigen all the pooled serum samples were routinely made up with saline to a concentration of 30 mg of protein/ml of pigeon serum. This sample was afterwards diluted 100-fold with normal saline, and aliquots of 5 ml each were prepared for the PT. There was an initial dose-ranging experiment. Volunteers with PBD were challenged with 0.9 to 0.15 mg/ml of pigeon serum. The selected dose was 0.3 mg/ml (dilution, 1:100) because it produced around a 15% drop in FVC and moderate fever without relevant side effects.
Inhalation test. The inhalation of the antigen test was carried out in an open space (hospital yard). All patients and control subjects were exposed to the antigen by inhaling through a jet nebulizer (Hudson Inc., Temecula, CA) with an oxygen flow of 10 L/min until finishing the whole volume of the nebulizer. The average time to nebulize the 5 ml was 20 min.
The mean difference between the baseline and the postchallenge (deltas) values of the pulmonary and laboratory tests from the three different groups of subjects were analyzed by using one-way analysis of variance (ANOVA) or Student's t test when comparing only two variables. In order to select the best cut point of the different parameters studied, an analysis of their sensitivity versus 1-specificity (ROC curves) was used. The test with higher sensitivity and the lower 1-specificity was considered to be the most useful. A logistic regression analysis, including the variables tested, was further used to identify parameters that might be predictive of CHP.
The baseline characteristics of patients and control subjects are shown in Table 1. CHP was diagnosed in 17 and 17 presented with ILDO. In 13 of these, IPF was diagnosed, one with eosinophilic granuloma, one with microlithiasis, one with lymphangioleiomyomatosis, and one with BOOP.
| Group | PBD (n = 17) | ILDO (n = 17) | AS (n = 5) | p Value | ||||
|---|---|---|---|---|---|---|---|---|
| Age | 42 ± 12 | 54 ± 10 | 32 ± 12 | |||||
| Sex, M/F | 2/15 | 10/7 | 2/3 | |||||
| Bird time exposure, mo (median) | 180 | 8 | 32 | < 0.001 | ||||
| AvAbs, OD | 1.12 ± 0.70 | 0.061 ± 0.05 | 1.078 ± 0.97 | < 0.001* | ||||
| Pulmonary function | ||||||||
| FVC, % pred | 66 ± 13 | 69 ± 12 | 97 ± 11 | < 0.001* | ||||
| FEV1, % pred | 62 ± 12 | 65 ± 10 | 97 ± 16 | < 0.001* | ||||
| FEV1/FVC | 84 ± 5 | 80 ± 8 | 83 ± 7 | NS* | ||||
| Dl CO, % pred | 43 ± 12 | 59 ± 9 | 113 ± 13 | < 0.001* | ||||
| TLC, % pred | 64 ± 12 | 64 ± 16 | 110 ± 14 | < 0.001* | ||||
| PaO2 , mm Hg | 54 ± 8 | 51 ± 4 | — | NS† | ||||
| PaCO2 , mm Hg | 33 ± 5 | 35 ± 4 | — | NS† | ||||
| SaO2 , % | 89 ± 5 | 87 ± 5 | 96 ± 2 | < 0.003* |
To determine the diagnosis of CHP, patients must fulfill the following major criteria: (1) exposure to pigeons preceding disease, (2) positive specific serum antibodies, (3) clinical, radiologic, and functional abnormalities compatible with ILD, (4) partial improvement of dyspnea upon avoidance of the avian antigen exposure, and (5) lung histology compatible with HP.
Exposure to pigeons was domestic in all cases. Pigeons were kept mostly as pets in small numbers (1 to 10) usually running freely in the house. Only one of the patients was a professional pigeon breeder exposed to approximately 1,000 birds in specially designed lofts. This patient also had two budgerigars at home. Thirteen patients with ILDO were also exposed to pigeons. However, the duration of the exposure was longer in patients with CHP (median, 180 mo) than in patients with ILDO (median, 8 mo). In addition, only patients with CHP and AS exhibited positive precipitant AvAbs.
Total IgE was in the normal range, but four of the 17 patients with CHP also exhibited positive pigeon-droppings-specific IgE antibodies, ranging from 0.37 to 0.68 AEU/ml.
Twenty patients exhibited a significant delta pre-post challenge, 17 of them corresponding to those in the HP group.
Clinical symptoms, and laboratory tests. Data are shown in Table 2 as mean delta from baseline. All patients displaying a positive test showed at least one systemic symptom such as headache, shivers, arthralgias, or simply malaise. These symptoms were observed from the first 4 h. Two patients, apparently those with the most severe reaction, vomited. Fever was the major symptom and was observed in all the patients who had a positive response, usually preceding other clinical symptoms. Its time of onset and maximal peak are shown in Figure 2 for the three groups of patients. In the group with PBD, the peak mean increase from baseline was 2.35° C (p < 0.0001), whereas for the control groups, ILDO and SA, it was 0.5° C (p > 0.05) and 0.05° C (p > 0.05), respectively. Five patients had symptoms of rhinitis, and 17 displayed sneezing and eye-itching. Fourteen of the 17 patients with PBD recalled having had these same symptoms recurrently in the past. Shortness of breath did not increase significantly in any group.
| PBD | ILDO | AS | p Values | |||||
|---|---|---|---|---|---|---|---|---|
| Leukocytes | −2,198 | −371 | −740 | NS | ||||
| Neutrophils, % | −5 | 2 | 5.4 | NS | ||||
| Lymphocytes, % | 5 | 1 | −4 | NS | ||||
| IgAU.I. | −160 | 13 | −16 | NS | ||||
| IgMU.I. | 31 | 5 | −14 | NS | ||||
| IgEU.I. | 56 | −95 | 1 | NS | ||||
| AvAbs, O.D. | −0.041 | −0.02 | −0.004 | NS | ||||
| VAS, cm | 0.4 | 0.09 | 0 | NS | ||||
| Body temp, ° C* | 2.35 | −0.53 | −0.05 | < 0.001 |
Fig. 2. Time-course of body temperature after provocation test. A significant difference was observed between CHP, ILDO, and AS groups (p < 0.0001).
[More] [Minimize]No changes were observed in WBC, RF, RCP, immunoglobulins, and AvAbs, after the provocation test (Table 2). No patients showed immediate reactions such as bronchospasm or wheezes.
Pulmonary function tests. The mean differences (deltas) for the variables studied before and after the PT are shown on Table 3. By using ANOVA the mean delta for FVC, TLC, PaO2 , and SaO2 were statistically greater in the group with PBD than in the other groups. However, when the analysis was made within groups, these changes (361, 303, and 8 mm Hg, and 7%, respectively) were significant only within patients with PBD, whereas in the other groups the falls were not statistically significant. The individual changes for patients with HP are shown in Table 4. Additionally, changes in FVC in the three groups are illustrated in Figure 3
| PBD | ILDO | AS | p Values | |||||
|---|---|---|---|---|---|---|---|---|
| FVC, ml | −361 | −134 | −160 | < 0.001 | ||||
| TLC, ml | −303 | −120 | −112 | < 0.001 | ||||
| Dl CO, ml/min/mm Hg | −1.01 | −0.63 | −0.20 | NS | ||||
| PaO2 , mm Hg | −8 | −2 | — | < 0.001 | ||||
| PaCO2 , mm Hg | −2 | −1 | — | NS | ||||
| SaO2 , % | −7 | −1 | 0 | < 0.001 |
| Patient No. | FEV1 | FVC | FEV1/FVC | PaO2 | SaO2 | Temperature | ||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pre | Post | Pre | Post | Pre | Post | Pre | Post | Pre | Post | Pre | Post | |||||||||||||
| 1 | 1,500 | 1,065 | 1,845 | 1,284 | 81 | 83 | 43 | 38 | 87 | 80 | 36.0 | 38.8 | ||||||||||||
| 2 | 1,890 | 1,750 | 2,360 | 1,884 | 80 | 93 | 49 | 43 | 88 | 80 | 36.0 | 39.0 | ||||||||||||
| 3 | 1,330 | 1,110 | 1,505 | 1,280 | 88 | 86 | 50 | 44 | 89 | 86 | 36.0 | 39.6 | ||||||||||||
| 4 | 1,330 | 1,120 | 1,635 | 1,360 | 81 | 82 | 62 | 39 | 92 | 85 | 36.0 | 39.3 | ||||||||||||
| 5 | 1,940 | 1,720 | 2,160 | 1,960 | 90 | 88 | 64 | 60 | 94 | 94 | 36.0 | 37.0 | ||||||||||||
| 6 | 2,100 | 1,820 | 3,000 | 2,786 | 71 | 65 | 49 | 46 | 89 | 89 | 36.5 | 37.5 | ||||||||||||
| 7 | 1,330 | 760 | 1,670 | 960 | 80 | 79 | 47 | 40 | 85 | 80 | 36.0 | 38.8 | ||||||||||||
| 8 | 1,500 | 1,120 | 1,885 | 1,380 | 80 | 81 | 49 | 47 | 83 | 77 | 36.5 | 38.0 | ||||||||||||
| 9 | 1,770 | 1,600 | 2,185 | 1,825 | 81 | 87 | 50 | 46 | 88 | 85 | 36.5 | 39.0 | ||||||||||||
| 10 | 1,760 | 1,320 | 1,905 | 1,320 | 92 | 100 | 63 | 50 | 95 | 91 | 36.0 | 37.6 | ||||||||||||
| 11 | 1,440 | 1,380 | 1,855 | 1,660 | 78 | 83 | 65 | 59 | 94 | 90 | 36.4 | 38.8 | ||||||||||||
| 12 | 1,080 | 1,040 | 1,255 | 1,202 | 86 | 86 | 42 | 40 | 86 | 81 | 36.0 | 38.0 | ||||||||||||
| 13 | 1,660 | 1,516 | 1,940 | 1,516 | 85 | 100 | 57 | 43 | 92 | 82 | 36.0 | 39.0 | ||||||||||||
| 14 | 995 | 950 | 1,185 | 1,042 | 84 | 91 | 40 | 36 | 75 | 67 | 36.0 | 38.3 | ||||||||||||
| 15 | 1,270 | 1,050 | 1,500 | 1,082 | 96 | 97 | 60 | 43 | 93 | 83 | 36.0 | 38.5 | ||||||||||||
| 16 | 1,200 | 1,146 | 1,320 | 1,146 | 90 | 100 | 65 | 54 | 95 | 85 | 36.0 | 39.0 | ||||||||||||
| 17 | 1,590 | 1,100 | 1,750 | 1,134 | 91 | 97 | 57 | 44 | 91 | 82 | 36.0 | 38.0 | ||||||||||||
Fig. 3. Effect of provocation tests on FVC in asymptomatic subjects (AS), pigeon breeder's disease (PBD), and interstitial lung diseases (ILD) other than HP.
[More] [Minimize]According to these results, FVC, PaO2 , SaO2 , and body temperature were used to construct ROC curves. TLC was excluded from ROC analysis because of its variability (12%) and low sensitivity.
ROC curves for temperature, FVC, PaO2 , and SaO2 , respectively, are shown in Figure 4. It can be seen that an increase of 0.5° C in temperature (point a) was the best cut point displaying a sensitivity (S) of 100%, a specificity (SP) of 82%, a positive predictive value of (PPV) of 100%, and a negative predictive value (NPV) of 86%. For FVC the best cut point was for a drop of 16% from baseline (point c) displaying a S of 76%, a SP of 81%, a PPV of 81%, and a NPV of 83%. For SaO2 and PaO2 the best cut points were for a drop of 3% or 3 mm Hg, respectively, showing a S of 88% for both, a SP of 82 and 86%, respectively, a PPV of 81 and 82%, and a NPV of 82 and 86%, respectively.
Fig. 4. (Left panel ) ROC curves for body temperature and FVC. a = an increase in temperature of 0.5° C; b = an increase > 1° C; c = an increase > 1.5° C; letter d = an increase > 2° C; and e = an increase > 2.5° C. Δ = FVC where; a = a decrease ⩾ 5%; b = a decrease ⩾ 10%; c = a decrease ⩾ 16%; d = a decrease ⩾ 18%; and e = a decrease ⩾ 20%. (Right panel ) ROC curves for SaO2 and PaO2 . a = a decrease ⩾ 3%; b = a decrease ⩾ 4%; and c = a decrease ⩾ 5%. Δ = PaO2 where a = a decrease ⩾ 3 mm Hg; b = a decrease ⩾ 4 mm Hg; c = a decrease ⩾ 5 mm Hg; d = a decrease ⩾ 6 mm Hg; e = a decrease ⩾ 7 mm Hg; and f = a decrease ⩾ 8 mm Hg.
[More] [Minimize]In order to further find out which is the best predictive test for the diagnosis of CHP, univariate and multiple-logistic regression analyses were made taking into account the changes of the variables showing significant postchallenge modifications. The univariate model confirmed that the changes in body temperature and FVC are predicting factors of CHP (RR, 82.5; CI, 10.43 to 651.76 and RR, 1.21; CI, 1.069 to 1.369, respectively). In the multivariate model, however, only the body temperature continued to be a predictor of CHP (body temperature: RR, 31.99; CI, 3.41 to 300.04; FVC: RR, 1.118; CI, 0.99 to 1.25).
This study has shown that provocation tests are useful in the diagnosis of hypersensitivity pneumonitis. Using strict criteria we found that fever and declines in lung function are common after inhalation challenge in patients with histologically proven CHP. Importantly, clinical findings were documented even in patients without prior history of such episodes despite ongoing environmental exposure. Thus, after the challenge test 14 of the 17 patients with chronic PBD, who previously were not aware of “acute” symptoms, did report having had similar episodes but dismissing them as unimportant and related to some short “flu-syndrome-like” rather than to any specific exposure to their birds. Therefore, the challenge test when positive became a useful way to convince the patients to make every effort to remove the exposure (birds) from the environment.
The most common problem facing the studies on diagnostic tests is the inability to establish a gold standard. Thus, for example, in the report by Hendrick and colleagues (18) only six patients out of the entire group of 29 subjects had a definitive diagnosis; however, none of them had a diagnosis confirmed by histologic sampling of the lung. In the remaining 23 patients, the diagnosis was not conclusive. The major weakness of that study was that the results of the challenge test were used as part of the diagnosis, and thus there was an inclusion bias.
The present study was made following a strict methodology of a diagnostic test (19). We were extremely careful to make up the precise diagnosis (gold standard). The group of pulmonologists who made the diagnosis had to agree with each other to consider the diagnosis as definitive. When disagreements existed a second or third evaluation was made to reconsider the case. The work was rigorously double-blinded in that the researcher (AR) who performed the challenge test and recorded PFT and clinical symptoms after the PT never knew the definitive diagnosis from the pulmonologists. On the other hand, these clinicians did not know the results of the challenge test because at the time of their evaluations, the diagnosis of a positive test was not yet established. The issue of including the histologic assessment of CHP as an additional criterion was further supported for the blindness of the study to the definitive diagnosis.
Fever was the single clinical variable that showed the most significant difference with the control groups. It is unlikely to explain fever as a side effect of the pooled serum because it was obtained under sterile conditions and, previously, to be used was heated and filtered. Moreover, a nonspecific febrile reaction should occur in the three groups. The increase in body temperature displayed a sensitivity of 100% and a specificity of 81%. This finding was further highlighted by the model of logistic regression that showed a strong predictive value. Three patients who had a false positive test explained the relatively low specificity observed for this variable as well as for other clinical tests. In a retrospective analysis, at least two of these patients might have been considered to have CHP. They had the antecedent of pigeon exposure, and the values for antibodies against avian antigen measured by the ELISA were very close to the threshold of a positive test. These patients had also exhibited a drop in FVC of 20 and 27%, an increase in body temperature of 2.5 and 2.7° C, and a decrease in PaO2 of 8 and 9 mm Hg, respectively, indicating an important positive response to the avian antigen. If these two patients were included as CHP rather than ILDO, the specificity and the negative predictive value for fever would increment to 94 and 96% instead to 81 and 86%, respectively.
Changes in the pulmonary function test were not better indicators of CHP than were changes in body temperature. Nevertheless, the sensitivity for a drop in FVC of 16% or more was just 76%, and a model of univariated logistic regression did reveal a significant predictive value. The lack of sensitivity for TLC and Dl CO may be explained by the interval of 24 h between the challenge and the tests since the most severe lung response appears to occur at 6 to 8 h after the challenge.
Although the drop in PaO2 and SaO2 showed better sensitivity and specificity than the drop in FVC, the logistic regression analysis was not useful to predict CHP.
Chest radiographs, total number of leukocytes and differentials, and immunoglobulins did not prove to be good indicators of a positive test. A possible explanation of this lack of change may be that the measurements of these variables were performed 24 h after the challenge. The abnormalities tended to resolve spontaneously over a period of 18 to 24 h. In this study, the initial changes were observed after 4 h, whereas their maximal response was observed at 8 to 10 h.
Interestingly, we found that 4 of 17 patients with HP presented positive, although at low levels, pigeon-droppings-specific IgE antibodies. Nevertheless, they did not show an early response, and they did not exhibit any difference in symptomatology or functional test abnormalities in comparison with the IgE-negative patients. As mentioned, in all patients with PBD, the changes in PFT were observed several hours after challenge, coinciding with systemic symptoms. No patients displayed immediate reactions such as bronchospasm or wheezes. Therefore, the response observed after challenge cannot be explained by an IgE-mediated asthmatic response.
The results of this study clearly show that to determine a positive test for diagnosis of CHP an increment of body temperature is better than pulmonary function tests. These results are consistent with those reported by Hendrick and colleagues (18) and Hargreave and Pepys (20).
In general terms, and according to our findings, the provocation test with avian antigen can differentiate patients with chronic PBD from those with IPF and other fibrotic disorders. In considering that patients with CHP induced by avian antigens had restrictive functional patterns similar to patients with ILDO, a clinical test capable of differentiating one from the other is required. This need is more obvious if the lung tissue assessment might not be sufficient to mark the differences with certainty, as occurs in advanced ILD. In the context of the differential diagnosis between CHP and IPF and other interstitial lung diseases, particularly when patients have antecedents of exposure to birds and positive AvAbs, an increase of body temperature after a challenge test with avian antigen is highly indicative of CHP. In this regard, the sensitivity of the test indicates that when a patient exhibits an increase of body temperature equal to or more than 0.5° C the probability of having PBD is 100%. Likewise, in terms of specificity, the usefulness of the test indicates that when a patient with ILD other than HP induced by avian antigen has a negative response, the probability of not to have PBD is 82%. By contrast, a previous study in our institute suggests that after incorporating clinical, immunologic, functional, and even histopathologic data, the ability of making a differential diagnosis between PBD and IPF in countries with high prevalence of antigen exposure (avian antigens in this case) is lower than the provocative challenge (3). The putative advantage of the inhalatory challenge implies that the use of a single test might hopefully avoid further use of invasive techniques such as lung biopsy.
After this study was finished, we prospectively evaluated the provocation test in an additional nine patients suspected of having HP who fulfilled the following major criteria: exposure to avian antigens with positive antigen-specific IgG antibodies, symptoms compatible with HP, highly compatible high resolution computed tomography, BAL fluid lymphocytosis, and histopathology compatible with HP. Provocation tests were positive in all of them, with an increase in temperature of > 0.5° C. A decrease of more than 16% in FVC was observed in eight of these nine patients. By contrast, when two patients with HP provoked by corn dust were challenged with avian antigen, there was no response (ΔFVC of 50 and 42 ml; ΔPaO2 of 1 and 0 mm Hg; and a Δ of body temperature of 0.1 and 0.3° C, respectively; p > 0.05).
Five asymptomatic subjects, who were relatives of the patients sharing the same exposure and epidemiologic history of patients with CHP were also challenged. They had similar values of AbAvs, but without clinical constraints or alterations in pulmonary function tests. It was suggested long ago that these exposed but asymptomatic subjects could have a subclinical and unnoticed disease (2, 21). Interestingly, we found that the response to the provocation test in these asymptomatic subjects was always negative, suggesting that their immune system did not develop a pathologic response, or that the time and intensity of exposure were not sufficient to develop one. In either case, this observation supports the strength of the provocation test to discern subjects who indeed have the disease from those who do not.
In order to know the safety of the challenge test in terms of the long-term outcome, changes in the pulmonary function test of patients submitted to the provocation test were compared with changes observed in a group of 10 patients who were not exposed to the challenge test. No differences were observed after 3 mo of observation.
In summary, our results showed that the provocation test with avian antigens can properly identify patients with CHP and so might be used for diagnostic purposes of this disease. Because the chronic form of hypersensitivity pneumonitis is characterized by an insidious onset of the disease, without typical acute symptoms, the advantage of doing the provocation test in the laboratory is that, with the appropriate dose of antigen, the patient presents with an acute, clearly distinguishable, reaction, which regularly does not occur in the natural environment under the mentioned conditions. Additionally, the different variables are better controlled.
However, a “natural and environmental” challenge test for hypersensitivity pneumonitis might also be considered. If we leave them out of their home exposure for 24 to 72 h, and we ask them to return with a close monitoring of FVC (perhaps peak flow), body temperature and general symptoms recorded in our questionnaire for at least 10 h, clinical and epidemiologic diagnosis of HP associated with bird exposure may be made. The advantages of this clinical approach diminishes risks, costs, time, and sophisticated studies to get a precise diagnosis and might be useful for other different types of antigen- induced diseases. Further research on this epidemiologic issue is needed not only for hypersensitivity pneumonitis induced by avian antigen but also for other antigen exposure.
The benefit of the challenge test is the avoidance of surgical intervention, mainly in patients in which the risk of a biopsy is greater than the advantage of the information obtained. An additional benefit is that it helps convince patients to remove birds from their homes. Likewise, these results might be extended to a clinical-epidemiologic approach for the diagnosis of this and other forms of hypersensitivity pneumonitis. Actually, inhalation challenge has been recently used in a nationwide epidemiologic study in Japan in several types of HP that have different prevalence. Although not clearly standardized, the test was a useful piece for diagnosis (22).
In conclusion, our results strongly suggest that provocation inhalatory tests may play an important role in the diagnosis of chronic HP. A suggested diagnostic algorithm is displayed in Figure 5.
Fig. 5. Diagnostic evaluation of a patient suspected of having subacute/chronic hypersensitivity pneumonitis. A suggested approach.
[More] [Minimize]The writers would like to thank Dr. Talmadge King for his valuable comments and suggestions.
Supported in part by CONACYT Grant No. F643-M9406.
| 1. | Longbottom J. L.Pigeon breeder's disease: quantitative immunoelectrophoretic studies of pigeon bloom antigen. Clin. Exp. Allergy191989619624 |
| 2. | Selman M., Chapela R., Raghu G.Hypersensitivity pneumonitis: clinical manifestations, diagnosis, pathogenesis and therapeutic strategies. Semin. Respir. Med.141993353364 |
| 3. | Perez-Padilla R., Salas J., Chapela R., Sanchez M., Carrillo G., Pérez R., Sansores R., Gaxiola M., Selman M.Mortality in Mexican patients with chronic pigeon breeder's lung compared with those with usual interstitial pneumonia. Am. Rev. Respir. Dis.14819934953 |
| 4. | Selman M., R. Chapela, J. Salas, R. Sansores, G. Carrillo, M. Sánchez, and R. Barrios. 1990. Hypersensitivity pneumonitis: clinical approach an integral concept about its pathogenesis: a Mexican point of view. In M. Selman and R. Barrios, editors. Interstitial Pulmonary Diseases: Selected Topics. CRC Press, Boca Raton, FL. 171–196. |
| 5. | Department of Clinical Epidemiology and Biostatistics. McMaster University Health Sciences CenterHow to read a clinical journal to learn about a diagnostic test. CMA J.1241981703710 |
| 6. | Kawanami O., Basset F., Barrios R., Lacronique J. G., Ferrans V. J., Crystal R. G.Hypersensitivity pneumonitis in man: light and electron microscopic studies of 18 lung biopsies. Am. J. Pathol.1101983277291 |
| 7. | Coleman A., Colby T. V.Histologic diagnosis of extrinsic allergic alveolitis. Am. J. Surg. Pathol.121988514518 |
| 8. | Crystal R. G., Bitterman P. B., Rennard S. I., Hance A. J., Keogh B. A.Interstitial lung disease of unknown cause: disorders characterized by chronic inflammation of the lower respiratory tract. N. Engl. J. Med.3101984154166 |
| 9. | Schwarz, M., and T. E. King. 1993. Interstitial Lung Disease, 2nd ed. Mosby Year Book Inc, St. Louis, MO. |
| 10. | Travis W. D., Borok Z., Roum J. H., Zhang J., Feuerstein I., Ferrans V. J., Crystal R. G.Pulmonary Langerhans cell granulomatosis (Histiocytosis X): a clinicopathologic study of 48 cases. Am. J. Surg. Pathol.171993971986 |
| 11. | Prakash U. B. S., Barham S. S., Rosenow E. C., Brown M. L., Payne W. S.Pulmonary alveolar microlithiasis: a review including ultrastructural and pulmonary function studies. Mayo Clin. Proc.581983290300 |
| 12. | Kitaichi M., Nishimura K., Itoh H., Izumi T.Pulmonary lymphangioleiomyomatosis: a report of 46 patients including a clinicopathologic study of prognostic factors. Am. J. Respir. Crit. Care Med.1511995527533 |
| 13. | Colby T. V., Myers J. L.Clinical and histologic spectrum of bronchiolitis obliterans, including bronchiolitis obliterans organizing pneumonia. Semin. Respir. Med.131992119133 |
| 14. | Ramı́rez A., Sansores R., Carrillo G., Salas J., Selman M.Validación de una escala análoga visual para medir disnea en pacientes con enfermedad intersticial difusa. Rev. Invest. Clin.461994479486 |
| 15. | Bañales J. L., Vazquez I., Mendoza F., Baltazares M., Nava A., Selman M.On the correct determination of reference values for seric antibodies against pigeon serum antigen using a group of healthy blood donors. Arch. Med. Res.281997289291 |
| 16. | American Thoracic SocietyStandardization of spirometry: 1987 update. Am. Rev. Respir. Dis.136198712851298 |
| 17. | Quanjer, P. 1983. Standardized lung function testing. Bull. Eur. Physiopathol. Respir. 19(Suppl. 5):1–95. |
| 18. | Hendrick D. J., Marchall R., Faux J. A., Krall J. M.Positive “alveolar” responses to antigen inhalation provocation test: their validity and recognition. Thorax351980415427 |
| 19. | Ransohoff D., Feinstein A.Problems of spectrum and bias in evaluating of diagnostic test. N. Engl. J. Med.2991978926929 |
| 20. | Hargreave F. E., Pepys J.Allergic respiratory reactions in bird fanciers provoked by allergen inhalation provocation tests: relation to clinical features and allergic mechanisms. J. Allergy Clin. Immunol.501972157173 |
| 21. | Cormier Y., Belanger J., Laviolette M.Persistent bronchoalveolar lymphocytosis in asymptomatic farmers. Am. Rev. Respir. Dis.1331986843847 |
| 22. | Yoshida K., Suga M., Nishiura Y., Arima K., Yoneda R., Tamura M., Ando M.Occupational hypersensitivity pneumonitis in Japan: data on a nationwide epidemiological study. Occup. Environ. Med.521995570574 |
