Additional environmental causes of interstitial lung disease may be recognized if temporal-spatial clusters of respiratory disease are evaluated. In November 1994 and February 1996, two young male textile workers from a plant in Rhode Island were referred by their pulmonologists to our academic occupational medicine clinic for evaluation of interstitial lung disease. Because the two cases constituted a sentinel occupational health event [5, 6], we alerted the National Institute for Occupational Safety and Health (NIOSH), the union representing the plant's workers, and the company's management. On learning of an earlier outbreak of five cases of nongranulomatous interstitial lung disease at the company's plant in Canada [7] and two additional biopsy-confirmed cases there in 1995, we began an investigation at the Rhode Island facility. This report details the results of that investigation, including the clinicopathologic and epidemiologic features of a previously unrecognized occupational lung disease in the nylon flocking industry [8, 9].

Background

In the flocking industry, short fibers (flock) are cut from cables of parallel synthetic monofilaments (tow) and applied (flocked) to an adhesive-coated substrate. At the index company, nylon tow, previously impregnated with a titanium dioxide delusterant, is usually dyed before it is bath-finished, cut, dried, screened, and bagged as flock. On the flock coating line, a water-based acrylic adhesive is applied to a moving bolt of cotton-polyester fabric before the bolt passes into the flocking room. There, on the basis of its electrostatic charge and vibration by beater bars, flock is embedded into the adhesive. After heat curing, the flocked fabric may be subjected to finishing, embossing, and printing. In the United States, an estimated 2500 workers employed at 12 flock manufacturing companies and 100 flocking companies make products for the upholstery, automobile, carpet, apparel, and novelty industries.

Substances to which workers were exposed at both the Rhode Island and the Canadian plants included bioaerosols; nylon fiber; a flock finish of tannic acid, ammonium ether of potato starch, and an alcohol-ammonium sulfate mixture; an acrylic adhesive containing carbon black; nonfibrous zeolite; heat transfer oil; and thermal degradation products. In 1989, the company's Canadian plant reportedly began processing the thinnest (9 microns) nylon fiber in use; in 1991, this operation was relocated to Rhode Island. Use of the thinner fiber reportedly caused the plant to be blanketed in fine white powder.

Except for a few case reports [10, 11], neither the literature nor industrial experience suggests that exposures in the flocking industry cause interstitial lung disease. The nylon manufacturer's Material Safety Data Sheet notes that: "As shipped, Nylon Polyamide Fiber products do not pose a hazard. Under normal conditions of use, Nylon Polyamide Fiber does not generate respirable fibers or dust." Nylon manufacturers have never conducted inhalation toxicity studies because of the assumption that nylon fiber is too thick to generate respirable fibers (diameter 3 microns) capable of alveolar deposition [12]. During the current investigation, however, NIOSH personnel identified airborne, respirable-size nylon fragments at the Rhode Island plant [8, 9].

Methods

Top Methods Results Discussion References

Case Finding and Case Definition

To identify all outbreak-associated cases of interstitial lung disease at the Rhode Island plant, all currently employed production workers with persistent respiratory symptoms and former employees identified by word of mouth were invited for evaluation. Employees reporting symptoms were evaluated by pulmonary function testing and chest radiography. Spirometry was conducted as recommended by the American Thoracic Society [13], lung volumes were measured by using plethysmography, and diffusing capacity was measured by the single breath method; reference values were those of Crapo and colleagues [14-16], and the severity scale was that of the American Medical Association [17]. All employees with symptoms suggesting pulmonary dysfunction were studied by high-resolution computed tomography (CT) (HiSpeed Advantage Helical Scanner, GE Medical Systems, Milwaukee, Wisconsin). Those with unexplained restrictive lung function, impairment of diffusing capacity, or high-resolution CT findings consistent with interstitial lung disease were referred for bronchoalveolar lavage, biopsy (transbronchial or wedge), or both. Using a four-point ordinal scale, a pulmonary pathologist scored all specimens, including those of the five Canadian patients, with respect to 18 histopathologic features.

Flock worker's lung was defined by persistent respiratory symptoms, previous work in the flocking industry, and histologic evidence of interstitial lung disease that had no better explanation. In the absence of a tissue specimen, the triad of an abnormal distribution of cell types on bronchoalveolar lavage, restrictive lung function, and high-resolution CT findings of diffuse ground-glass opacity or micronodularity served as a surrogate for the histologic criterion.

Blood samples for measurement of complete blood count, leukocyte differential, rheumatoid factor, and antinuclear antibody were obtained from employees who had bronchoalveolar lavage or biopsy. The Johns Hopkins University Dermatology, Allergy, and Clinical Immunology Reference Laboratory (Baltimore, Maryland) analyzed serum specimens for precipitating antibody to pigeon serum, Aspergillus fumigatus (antigens 1, 2, and 6), Aspergillus flavus, Aspergillus niger, Aureobasidium pullulans, and thermophilic actinomycetes; for specific IgE antibodies to a mixture of 8 to 10 common aeroallergens (Phadiatop, Pharmacia & Upjohn, Inc., Kalamazoo, Michigan); and for potato-specific IgE.

Cohort Investigation

The study cohort comprised all current and former production workers, employed at the Rhode Island facility on or after 15 June 1990, who worked for at least 18 months before 15 September 1996 (n = 165). The cohort was based on the company's annual vacation lists from 15 June 1990 through 15 June 1996, which tabulate all hourly workers employed continuously during the previous 6 months. Cohort membership was restricted to those who had worked for at least 18 months because employees working for shorter periods might not appear on vacation list. The company provided employment histories for salaried production workers who satisfied the criteria for cohort membership. Job histories were recorded for cohort members with accessible personnel files. The files of former employees were scrutinized for evidence of lung disease.

After they had worked for 18 months, cohort members contributed person-years at risk for time worked subsequently from 15 June 1990 to 15 September 1996. Crude incidence rates of flock worker's lung and all interstitial lung disease were calculated. General population estimates for age- and sex-specific incidence rates of pulmonary fibrosis-idiopathic pulmonary fibrosis and for sex-specific incidence rates of all interstitial lung disease were obtained from an interstitial lung disease registry maintained for Bernalillo County, New Mexico [1]. This registry provides the only such published incidence data for a general population in the United States. The racial and ethnic distributions of Bernalillo County and the study cohort were similar [1, 9]. By using the registry's estimates [1] and weights based on the age and sex distribution of the study cohort, standardized incidence ratios for pulmonary fibrosis-idiopathic pulmonary fibrosis and for all interstitial lung disease were calculated. To calculate the former, we counted cases of flock worker's lung as cases of pulmonary fibrosis-idiopathic pulmonary fibrosis because they probably would have been assigned to this broader diagnostic category if the occupational association had been overlooked. We used exact Poisson calculations to derive 95% CIs for the estimated standardized incidence ratios [18, 19]. Computations were done by using SAS software (SAS Institute, Inc., Cary, North Carolina).

Results

Top Methods Results Discussion References

Case Finding

We clinically evaluated 39 of the 148 current production workers and 3 former employees. We also reviewed questionnaire responses, spirometry results, and chest radiographs of an additional 78 production workers who authorized us to review their NIOSH-generated screening records [9] (Figure 1). We found 7 men and 1 woman (mean age ±SD, 43 ± 12 years) with flock worker's lung; their demographic and clinical features are shown in Table 1 and Table 4, Table 2 and Table 5 and Table 3 and Table 6. Two reported stable dry cough and dyspnea; 6 reported gradually worsening dry cough and dyspnea; and 3 reported intermittent atypical chest pain. Median latency was 6 years from date of hire to symptom onset and 15 months from symptom onset to diagnosis. Only 2 persons reported work-related symptoms. No specific job category was associated with illness. Crackles were auscultated in all but one patient (patient 7). One patient had normal lung function, 2 had mild airway obstruction, and 5 had both moderate-to-severe diffusion impairment and mild-to-moderate restriction.

View larger version (18K): [in this window] [in a new window]   Figure 1. Distribution of study participants by extent of evaluation. Two additional former employees who were not members of the cohort were also clinically evaluated.

Radiography

On radiography, three patients had normal results, four had diffuse reticulonodular infiltrates, and one had patchy areas of consolidation. On high-resolution CT, one patient had diffuse micronodularity and patchy areas of mild ground-glass opacity (Figure 2 A), three had patchy areas of mild ground-glass opacity (Figure 3 A), two had patchy areas of both consolidation and ground-glass opacity (Figure 4 A), and two had bilateral peripheral honeycombing (Figure 5 A).

View larger version (84K): [in this window] [in a new window]   Figure 2. Patient 4. A. High-resolution computed tomographic scan showing micronodularity and patchy areas of mild ground-glass opacity. B. Open-lung biopsy specimen showing nodules of lymphocytes (arrows) in inflamed and consolidated parenchyma (hematoxylin-eosin stain; original magnification, x96).

View larger version (60K): [in this window] [in a new window]   Figure 3. Patient 5. A. High-resolution computed tomographic scan showing patchy areas of subtle ground-glass opacity. B. Transbronchial biopsy specimen showing lymphocytic infiltration of a bronchiole and a germinal center (asterisk) (hematoxylin-eosin stain; original magnification, x83).

View larger version (84K): [in this window] [in a new window]   Figure 4. Patient 8. A. High-resolution computed tomographic scan showing patchy areas of consolidation and ground-glass opacity. B. Open-lung biopsy specimen showing organizing fibrous tissue within alveoli (thick arrows) and widespread lymphocytic infiltrates with perivascular lymphoid nodules (thin arrows) (hematoxylin-eosin stain; original magnification, x96).

View larger version (127K): [in this window] [in a new window]   Figure 5. Patient 2. A. High-resolution computed tomographic scan showing diffuse reticular densities and honeycombing consistent with interstitial fibrosis. B. Open-lung biopsy specimen showing a longitudinally sectioned bronchiole with severe lymphocytic bronchiolitis and parenchymal infiltrates (lower right) (hematoxylin-eosin stain; original magnification, x38). C. Higher magnification of a region similar to the lower right portion of panel B, showing lymphocytic infiltration of alveolar walls (arrow) and intra-alveolar macrophages and multinucleated giant cells (hematoxylin-eosin stain; original magnification, x380).

Blood Studies

Testing of peripheral blood revealed eosinophilia (15%) in patient 6 and antinuclear antibody (1:320) in patient 8. Precipitating antibody was present in patient 2 (Aspergillus niger, Aspergillus fumigatus 1) and patient 4 (Aspergillus niger). The Phadiotop aeroallergen multiscreen test was positive in patients 4, 6, and 8. Potato-specific IgE was detected in patients 4 and 6.

Pathology

Six patients had bronchoalveolar lavage; two of these patients had more than 30% lymphocytes, three had more than 25% eosinophils, and one had normal cell populations. The results of transbronchial (n = 2) and wedge (n = 5) biopsies are shown in Table 3 and Table 6. Six of the seven patients who underwent biopsy had findings (Figure 2 B, Figure 3 B, Figure 5 B and Figure 5 C) of nonspecific interstitial pneumonia, as the term has been used either in HIV-infected patients [20, 21] or for the idiopathic interstitial pneumonias [3]. The seventh patient who underwent biopsy (patient 8) had findings indicating bronchiolitis obliterans organizing pneumonia (Figure 4 B). All seven patients had nodular, peribronchovascular interstitial lymphoid infiltrates, generally of moderate intensity, and five had germinal centers. Six had lymphocytic bronchiolitis, five had lymphocytic infiltration of alveolar septa, and five had interstitial fibrosis that was usually mild. Six showed a generally mild increase in the number of intra-alveolar macrophages, four had interstitial eosinophilia that was mild in all but one case (patient 6), and none had diffuse alveolar damage. All findings were seen in a predominantly peribronchovascular distribution.

Tissue specimens of the five Canadian case-patients who had biopsy, including the three previously reported to have desquamative interstitial pneumonia [7], showed lymphocytic interstitial infiltrates in all cases and germinal centers in three, typically in a peribronchovascular distribution. Although intra-alveolar collections of macrophages were conspicuous, their heterogeneity was incompatible with a diagnosis of desquamative interstitial pneumonia [2, 3]. Areas of diffuse alveolar damage were dominant in two cases. Neither the Canadian nor the Rhode Island specimens revealed granulomas or birefringent particles.

Natural History

The patients from Rhode Island all had symptomatic, radiographic, and functional improvement within weeks to months of leaving work, but none have recovered completely. Immunosuppressive therapy was prescribed only for patients 1 (1985 to 1990), 2 (1992 to 1997), and 3 (1994 to 1996). Patient 2 required oxygen for 3 years after leaving work but not for the past year, during which time his diffusing capacity has remained stable. After patient 6 left work, his potato-specific IgE titer decreased; the eosinophilia of his blood, bronchoalveolar lavage, and bronchial mucosa disappeared; and his only methacholine challenge showed normal airway responsiveness. Patient 8 (who had bronchiolitis obliterans organizing pneumonia) improved rapidly without immunosuppressive therapy after he left work.

Additional Possible Cases

Of the 42 clinically evaluated workers, at least 7 who did not receive a diagnosis of flock worker's lung had abnormalities, mainly findings on high-resolution CT, compatible with such a diagnosis. Only 2, however, underwent bronchoalveolar lavage, biopsy, or both, and none satisfied the histologic criterion. One had restrictive lung function, normal radiographs, lymphocytic lavage, and no tissue specimen. A second had reversible airway obstruction; unilateral, patchy ground-glass opacity; normal findings on lavage; eosinophilic mucosal infiltrates; and a meager transbronchial biopsy specimen showing respiratory bronchiolitis. A third had prominent mosaic ground-glass opacity, moderate diffusion impairment, and mild airflow obstruction. Four others had patchy areas of mild ground-glass opacity; two of the four had ill-defined centrilobular nodules, and two had mild-to-moderate pulmonary function abnormalities. We recommended further evaluation for 41 of the 78 additional employees whose NIOSH screening records we reviewed; recommendations were based on symptom history alone for 35 employees and on additional abnormalities on spirometry or chest radiography for 6. No follow-up data are available.

Cohort Investigation

The 162 men and 3 women who made up the study cohort contributed 660 and 9 person-years at risk, respectively, over the age strata 18 to 34 years (207 person-years at risk from men, 1 person-year at risk from women), 35 to 44 years (289 person-years at risk from men, 8 from women), 45 to 54 years (99 person-years at risk from men, 0 from women), 55 to 64 years (58 person-years at risk from men, 0 from women), and 65 to 74 years (7 person-years at risk from men, 0 from women). Seven of the 8 case-patients (patients 2 through 8) were members of the study cohort, yielding a crude incidence of flock worker's lung of 1046 cases per 100 000 person-years. The crude incidence of all interstitial lung disease in the study cohort (n = 10, including one case of pulmonary histiocytosis X and two cases of talcosis due to nonplant exposures) was 1495 cases per 100 000 person-years.

In the Bernalillo County reference population of 480 577 persons, 202 incident cases of all interstitial lung disease were seen during the 2-year period from 1988 to 1990; sex-specific incidence rates were 31.5 cases per 100 000 person-years for men and 26.1 cases per 100 000 person-years for women [1]. Pulmonary fibrosis-idiopathic pulmonary fibrosis accounted for 91 of these cases, yielding age- and sex-specific incidence rates of 4.0 for men and 0 for women 35 to 44 years of age, 2.2 for men and 4.0 for women 45 to 54 years of age, 14.2 for men and 17.5 for women 55 to 64 years of age, 71.0 for men and 33.1 for women 65 to 74 years of age, and 152.9 for men and 89.6 for women 75 years of age and older per 100 000 person-years [1].

The sex-adjusted incidence ratio for all interstitial lung disease in the Rhode Island cohort was 48 (95% CI, 23 to 88). The standardized incidence ratio for pulmonary fibrosis-idiopathic pulmonary fibrosis (used as a surrogate for flock worker's lung) was 258 (CI, 104 to 530). The standardized incidence ratio would have been 478 (CI, 255 to 818) if the six other suspected cases of flock worker's lung had been confirmed histologically and would have been 147 (CI, 40 to 377) if only the 4 case-patients (patients 2, 3, 4, and 8) with moderate-to-severe disease were counted.

Discussion

Top Methods Results Discussion References

The eight-case cluster of interstitial lung disease described here reflects an incidence rate approximately 250 times greater than background rates. Together with the cluster at the company's Canadian facility [7], the cases represent what seems to be the largest unexplained continuing outbreak of nongranulomatous interstitial lung disease under investigation.

Numerous clinicopathologic findings in our patients are noteworthy. First, three patients had normal chest radiographs, normal or minimally abnormal pulmonary function, and subtle abnormalities on high-resolution CT that were initially over-looked. Second, three patients had bronchoalveolar lavage eosinophilia of a degree (25% to 40%) typically seen only in chronic eosinophilic pneumonia [22-28]. Third, although the observed histopathologic findings are most readily encompassed by the term nonspecific interstitial pneumonia [3, 20, 21], considerable variation was seen. All patients, however, had nodular peribronchovascular lymphocytic infiltrates, and most had diffuse lymphocytic interstitial infiltrates, germinal centers, and lymphocytic bronchiolitis. Fourth, half of the affected workers, including two receiving corticosteroids, had documented disease progression while they continued to work. Finally, all patients, even those with honeycombing, improved substantially after leaving the plant.

With respect to alternative diagnoses, the absence of granulomata from all 12 biopsy specimens from workers in Rhode Island and Canada makes hypersensitivity pneumonitis untenable because 66% to 100% of biopsy specimens from patients with this condition have granulomata [29-32]. The marked bronchoalveolar lavage eosinophilia seen in three patients is also inconsistent with a diagnosis of hypersensitivity pneumonitis [22-26]. Usual interstitial pneumonia was originally diagnosed in patient 2, but the patient's intense nodular and diffuse lymphocytic bronchiolitis, improvement and stabilization of pulmonary function after retirement, and now 6-year survival argue against this diagnosis [2]. Although respiratory bronchiolitis-associated interstitial lung disease was suggested in patient 7 and the bronchiolitis obliterans organizing pneumonia in patient 8 may have been idiopathic, lymphoid nodules have not been reported in these conditions [33-37]. Moreover, after leaving work, patient 8 recovered rapidly without receiving corticosteroid therapy. Finally, although the bronchiolocentric lymphoid nodules characterizing our patients occur in cases of follicular bronchitis-bronchiolitis [38, 39], the immunologic disorders typically associated with this condition were absent.

The limitations of our study should be noted. Observation bias, the peculiarities of the cohort's construction, or the choice of pulmonary fibrosis-idiopathic pulmonary fibrosis as a basis for comparison may have inflated the standardized risk estimates. The workforce was scrutinized more intensely than the comparison population, but five patients had moderate-to-severe pulmonary disease and all eight had seen physicians for chest symptoms. In fact, the study probably underestimated the true risk because many symptomatic cohort members had limited, if any, evaluation and because diagnostic criteria were conservative. Finally, although our use of incidence rates of pulmonary fibrosis-idiopathic pulmonary fibrosis to derive standardized incidence ratios for flock worker's lung may be questioned, the experiences of patients 1 and 2 suggest that if affected workers had remained at the plant, they may have developed progressive fibrotic lung disease.

Although the cause of flock worker's lung remains unknown, there are only several etiologic candidates. Nylon fiber and its three-component finish accounted for most of the airborne particulates in the Rhode Island plant [8, 9]. Potato starch and tannic acid, two constituents of the finish, have adverse pulmonary effects. In the potato processing industry [40], employees experience work-related wheezing and chest tightness but not nongranulomatous interstitial lung disease, and air concentrations of total dust, endotoxin, and potato antigens are substantially higher than those encountered by the most intensely exposed workers at the Rhode Island plant [8, 9]. Consequently, potato starch may contribute to flock worker's lung but is unlikely to be the complete cause of the condition. The same is probably true of tannic acid, a ubiquitous constituent of woody plants that has been shown to modulate alveolar macrophage function [41]. The third finish constituent, a mixture of aliphatic alcohols and ammonium sulfates, is believed to pose no risk.

The absence of carbon dust lung macules exonerates carbon black [42]. Nonfibrous zeolite, an aluminosilicate, is not known to cause lung disease [43]; furthermore, silicatosis would be extremely unlikely in the absence of birefringent particles on biopsy. The etiologic agent may be a novel compound generated through the thermal degradation of nylon, finish, or adhesive, but this is speculation. Nylon 6,6 melts at 254 to 263 °C; at higher temperatures, ammonia, various amines, hydrogen cyanide, and nitrogen dioxide may be generated. Acute exposure to high concentrations of nitrogen dioxide may cause bronchiolitis obliterans organizing pneumonia, but this gas was not detected at the Rhode Island facility [8, 9].

Investigators of the earlier outbreak at the company's Canadian plant postulated an etiologic role for mycotoxin [7]. They found Fusarium species contaminating the plant's adhesive but detected no mycotoxin. Mycotoxin is unlikely to be the cause of flock worker's lung for three reasons: 1) extensive sampling at the Rhode Island plant revealed only modest air concentrations of fungal spores and, in bulk samples, no Stachybotrys atra (the only mycotoxin-containing spore shown to cause inhalation injury in experimental animals) [44, 45]; 2) additional workers developed interstitial lung disease at the Canadian plant after measures to minimize mold growth were implemented; and 3) no convincing evidence indicates that inhalation exposure to mycotoxins causes chronic interstitial lung disease.

In 1975, Pimentel and colleagues [10] reported diffuse interstitial lung disease in two patients with occupational exposure to nylon. Biopsies revealed chronic cellular infiltrates and fibrosis of both the bronchial walls and surrounding interstitium and birefringent fibers with physicochemical properties suggestive of nylon. In an accompanying experimental study [10], 28 guinea pigs were sprayed with 2 g of pulverized nylon dust three times daily for 325 days. Lung sections from half of the animals revealed "... oedema, reticular fibres, and especially, granulomatous cellular proliferations made up of histiocytes and fibroblasts ... localized in the interalveolar septa ... ." The study has been criticized [12], but its results were unchallenged by a better-designed study. Preliminary results of recent experimental studies conducted by NIOSH investigators, however, provide evidence that respirable-size nylon fragments have substantial pulmonary toxicity [9].

In 1993, a fatal epidemic of bronchiolitis obliterans organizing pneumonia occurred among textile dye sprayers in Spain and Algeria [46, 47]. This "Ardystil" syndrome was attributed to the introduction of a new spray-on dye formulation containing a polyamide-amine polymer made from diethylenetriamine and hexanedioic acid. Surprisingly, intratracheal instillation studies of Syrian golden hamsters revealed both the new and old dye formulations to be highly toxic and fibrogenic [48]. The epidemic and experimental studies provided the first convincing evidence that recurrent inhalation exposure to synthetic polymers, previously believed to be inert, can cause severe pulmonary injury [48]. Moreover, one of the two polymers is similar to nylon 6,6, which is itself a polyamide, formed by reacting hexamethylenediamine and hexanedioic acid. Of note, bronchiolitis obliterans organizing pneumonia, the signature pathology of the Ardystil syndrome [46, 47], was found in one, possibly two, of the Rhode Island cases (patients 3 and 8). It remains to be seen, however, whether respirable nylon fragments are generated in the flocking industry at concentrations high enough to cause pulmonary injury, and, if so, whether they cause flock worker's lung, as preliminary experimental findings have suggested [9].

Although the specific cause of flock worker's lung remains in question, the recognition of this occupational lung disease in workers in two textile plants in two countries is a finding of substantial public health importance relevant to physicians caring for patients employed in the flocking industry.

Appendix: Index Cases

First Referral

In February 1993, a 34-year-old, previously asymptomatic textile worker (patient 3) suddenly developed headache, pleuritic chest pain, and dyspnea 3 hours into his work shift. At an emergency department, physical examination was normal except for tachypnea. A chest radiograph was of poor quality. A room air arterial blood gas test revealed a PaO₂ of 62 mm Hg, a PaCO₂ of 34 mm Hg, and a pH of 7.46. The patient's symptoms disappeared spontaneously in a few days, and he resumed working without limitation. In April 1993, he began to experience work-related dyspnea that resolved shortly after the end of each work shift. By January 1994, the patient had dyspnea on effort even outside of work and had a chronic, productive cough. At the time, results of physical examination, chest radiography, and pulmonary function tests were normal except for a low diffusing capacity (45% of predicted value). Over the ensuing 5 months, the patient developed increasing dyspnea on effort and mild wheezing, only to become asymptomatic toward the end of his 2-week summer vacation. Within a week of resuming work, he was again increasingly short of breath.

When the patient was reevaluated in September 1994, chest auscultation revealed diffuse crackles. Pulmonary function tests showed both moderate restriction and a further deterioration in diffusing capacity (29% of predicted value). Chest radiographs showed subtle diffuse reticulonodular infiltrates and mild ground-glass opacity. High-resolution CT showed diffuse, striking ground-glass opacity and patchy areas of consolidation, predominantly in the lower lobes. Complete blood count showed a mildly elevated leukocyte count and normal differential. The erythrocyte sedimentation rate was 48 mm/h. Serologic studies for antinuclear antibody, antineutrophil cytoplasmic antibody, and rheumatoid factor were negative.

A consulting pulmonologist diagnosed occupational hypersensitivity pneumonitis; removed the patient from work; and prescribed prednisone, 50 mg/d. One week later, bronchoscopy and bilateral bronchoalveolar lavage revealed airway mucosa with normal appearance and 24% to 47% neutrophils, 23% to 33% eosinophils, 13% lymphocytes, 9% to 19% macrophages, and 8% to 11% monocytoid cells. The patient was removed from work for 28 months and treated with corticosteroids. Two months after leaving work, he was feeling better and his spirometry results and lung function were almost normal, but his diffusing capacity remained low (38% of predicted value).

In November 1994, the patient was referred to the Brown University Program in Occupational Medicine, based at the Memorial Hospital of Rhode Island. His occupational history revealed that since May 1992, he had been employed as a flock-coating line attendant at a Rhode Island facility that manufactured primarily nylon flock and flocked fabric. Although the patient's primary duty was to apply an acrylic adhesive to moving bolts of cloth, he worked next to the flock room, where finely powdered nylon fiber (flock) was deposited onto cloth. A serum sample did not reveal precipitating antibody to a panel of antigens.

A diagnosis of hypersensitivity pneumonitis was considered unlikely for three reasons: 1) the company's Material Safety Data Sheets revealed no chemicals known to cause hypersensitivity pneumonitis, 2) the patient's serum revealed no precipitating antibody to the bioaerosols most often implicated in hypersensitivity pneumonitis, and 3) the degree of eosinophilia seen in the patient's bronchoalveolar lavage fluid had not previously been reported in hypersensitivity pneumonitis. Nevertheless, the Memorial Hospital team visited the plant. They found visible flock in the workroom air and much standing water but no visible mold growth. Volumetric air sampling revealed concentrations of viable fungal spores (primarily cladosporium and penicillium) at the patient's work station (298 colony-forming units/mm³) that were lower than those outside the plant. A diagnostic open-lung biopsy was recommended but was declined.

In January 1995, bronchoalveolar lavage revealed 35% to 56% macrophages, 31% to 33% eosinophils, 5% to 24% neutrophils, and 7% to 8% lymphocytes. Over the next year, the patient's pulmonary function continued to improve but thereafter remained stable with a persistent moderate impairment in diffusing capacity (57% of predicted value). Over the same period, the patient's high-resolution CT scan almost completely returned normal. In late 1996, the patient's corticosteroid therapy was discontinued without complication. In February 1997, the patient returned to work at one of the company's off-site warehouses. Shortly after being transferred to another warehouse, he developed symptoms of mucosal irritation and quit permanently.

Second Referral

In December 1995, a 28-year-old man (patient 4) sought the assistance of a pulmonologist for a minimally productive cough of 18 months' duration and dyspnea on effort and intermittent pleuritic chest pain of 6 months' duration. His symptoms showed little temporal variability, and review of systems was otherwise noncontributory. He had no known allergies; was using no medications; and denied both domestic and recreational exposure to chemicals, pets, birds, barns, stored vegetable matter, mineral dusts, and water-damaged building materials. The patient reported having had mild, childhood-onset asthma that had disappeared when he was in his late teens. An occupational history revealed that for the previous 7 years, the patient had worked at the same manufacturing facility where patient 3 had been employed. He spent his first 3 years of employment on the flock-coating line. Thereafter, except for 1 month of work in the flock room at the end of 1993, he worked continuously in the shipping department, a job that required him to travel regularly throughout the plant.

Physical examination revealed no abnormalities other than diffuse crackles on chest auscultation. Chest radiographs showed bilateral diffuse reticulonodular infiltrates. High-resolution CT showed diffuse micronodular densities and patchy, mild ground-glass opacity (Figure 2 A). Pulmonary function tests revealed moderate restriction and moderate impairment in diffusing capacity (Table 2 and Table 5). Methacholine challenge revealed a moderate degree of airway hyperresponsiveness (provocative concentration of methacholine causing a 20% decrease in FEV₁ from baseline, 5 mg/mL). Complete blood count, including leukocyte differential, was normal, and antinuclear antibody and rheumatoid factor were absent from serum.

Bronchoscopy revealed diffuse inflammation of the bronchial mucosa and occlusion of the posterior segment of the left lower lobe by an adherent, white, granular material. On biopsy, the granular material revealed diffuse lymphocytic infiltration so dense that two pathologists raised the possibility of lymphoma. Bilateral bronchoalveolar lavage revealed 52% to 65% lymphocytes, 30% to 32% macrophages, 5% to 11% neutrophils, and 2% to 5% eosinophils. Subsequent open-lung biopsy showed diffuse interstitial lung disease characterized by bronchiolocentric nodular and diffuse interstitial lymphoid infiltrates, germinal centers, and mild interstitial fibrosis without granulomas or birefringent particles (Figure 2 B).

The patient was removed from work and referred to Memorial Hospital of Rhode Island with a presumptive diagnosis of hypersensitivity pneumonitis. While out of work, he gradually recovered without medication. Ten months later, he remained minimally symptomatic, and both his high-resolution CT scan and pulmonary function tests showed only mild abnormalities. His serum was found to contain precipitating antibody to Aspergillus niger but not to any other antigen tested.

Note Added in Proof: In the 3 months before publication of this paper, the diagnosis of flock worker's lung was confirmed by biopsy in two additional young men who had become symptomatic 3 to 12 months earlier while employed at the Rhode Island plant.

Dr. Crausman: Division of Pulmonary and Critical Care Medicine, Memorial Hospital of Rhode Island, Pawtucket, RI 02860. Ms. Durand: Department of Environmental Health, Box 357234, University of Washington, Seattle, WA 98195.

Dr. Nayer: Internal Medicine Residency Program, Memorial Hospital of Rhode Island, Pawtucket, RI 02860.

Dr. Kuhn: Department of Pathology, Memorial Hospital of Rhode Island, Pawtucket, RI 02860.