[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.158.98.119. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
A, Papanicolaou-stained smear of bronchoalveolar lavage fluid shows the globules stained green, orange, or orange in the center coated with a green rim (original magnification ×400). B, The globules appear to be dark blue in the Riu-stained smear (original magnification ×400). C, In the combined alcian blue and PAS (periodic acid–Schiff)-stained smear, the globules show strong PAS staining with a lack of alcian blue staining (original magnification ×400).

A, Papanicolaou-stained smear of bronchoalveolar lavage fluid shows the globules stained green, orange, or orange in the center coated with a green rim (original magnification ×400). B, The globules appear to be dark blue in the Riu-stained smear (original magnification ×400). C, In the combined alcian blue and PAS (periodic acid–Schiff)-stained smear, the globules show strong PAS staining with a lack of alcian blue staining (original magnification ×400).

Table 1. 
Clinical Data of 7 Patients With Pulmonary Alveolar Proteinosis*
Clinical Data of 7 Patients With Pulmonary Alveolar Proteinosis*
Table 2. 
Clinical Diagnoses of 128 Patients With Various Pulmonary Disorders
Clinical Diagnoses of 128 Patients With Various Pulmonary Disorders
Table 3. 
The Number of Globules in Papanicolaou-Stained Smears of Bronchoalveolar Lavage Fluid Specimens*
The Number of Globules in Papanicolaou-Stained Smears of Bronchoalveolar Lavage Fluid Specimens*
1.
Rosen  SHCastleman  BLiebow  AA Pulmonary alveolar proteinosis. N Engl J Med. 1958;2581123- 1142Article
2.
Bedrossian  CWLuna  MAConklin  RHMiller  WC Alveolar proteinosis as a consequence of immunosuppression: a hypothesis based on clinical and pathologic observations. Hum Pathol. 1980;11(suppl 5)527- 535
3.
Prakash  UBBarham  SSCarpenter  HADines  DEMarsh  HM Pulmonary alveolar phospholipoproteinosis: experience with 34 cases and a review. Mayo Clin Proc. 1987;62499- 518Article
4.
Goldstein  LSKavuru  MSCurtis-McCarthy  PChristie  HAFarver  CStoller  JK Pulmonary alveolar proteinosis: clinical features and outcomes. Chest. 1998;1141357- 1362Article
5.
Martin  RJRogers  RMMyers  NM Pulmonary alveolar proteinosis: shunt fraction and lactic acid dehydrogenase concentration as aids to diagnosis. Am Rev Respir Dis. 1978;1171059- 1062
6.
Mermolja  MRott  TDebeljak  A Cytology of bronchoalveolar lavage in some rare pulmonary disorders: pulmonary alveolar proteinosis and amiodarone pulmonary toxicity. Cytopathology. 1994;59- 16Article
7.
Cordonnier  CFleury-Feith  JEscudier  EAtassi  KBernaudin  JF Secondary alveolar proteinosis is a reversible cause of respiratory failure in leukemic patients. Am J Respir Crit Care Med. 1994;149788- 794Article
8.
Burkhalter  ASilverman  JFHopkins  MBGeisinger  KR Bronchoalveolar lavage cytology in pulmonary alveolar proteinosis. Am J Clin Pathol. 1996;106504- 510
9.
Mikami  TYamamoto  YYokoyama  MOkayasu  I Pulmonary alveolar proteinosis: diagnosis using routinely processed smears of bronchoalveolar lavage fluid. J Clin Pathol. 1997;50981- 984Article
10.
Sosolik  RCGammon  RRJulius  CJAyers  LW Pulmonary alveolar proteinosis: a report of two cases with diagnostic features in bronchoalveolar lavage specimens. Acta Cytol. 1998;42377- 383Article
11.
Takemura  TFukuda  YHarrison  MFerrans  VJ Ultrastructural, histochemical, and freeze-fracture evaluation of multilamellated structures in human pulmonary alveolar proteinosis. Am J Anat. 1987;179258- 268Article
12.
Lee  CHLan  RSTsai  YHChiang  YCWang  WJ Riu's stain in the diagnosis pulmonary cryptococcosis: introduction of a new diagnostic method. Chest. 1988;93467- 470Article
13.
Wang  BMStern  EJSchmidt  RAPierson  DJ Diagnosing pulmonary alveolar proteinosis: a review and an update. Chest. 1997;111460- 466Article
14.
Ramirez  J Pulmonary alveolar proteinosis: treatment by massive bronchoalveolar lavage. Arch Intern Med. 1967;119147- 156Article
15.
Selecky  PAWasserman  KBenfield  JRLippmann  M The clinical and physiological effect of whole-lung lavage in pulmonary alveolar proteinosis: a ten-year experience. Ann Thorac Surg. 1977;24451- 461Article
16.
Du Bois  RMMcAllister  WABranthwaite  MA Alveolar proteinosis: diagnosis and treatment over a 10-year period. Thorax. 1983;38360- 363Article
17.
Greaves  TSStrigle  SM The recognition of Pneumocystis carinii in routine Papanicolaou-stained smears. Acta Cytol. 1985;29714- 720
18.
Ghali  VSGarcia  RLSkolom  J Fluorescence of Pneumonitis carinii in Papanicolaou smears. Hum Pathol. 1984;15907- 909Article
Original Investigation
February 26, 2001

Diagnosis of Pulmonary Alveolar ProteinosisUsefulness of Papanicolaou-Stained Smears of Bronchoalveolar Lavage Fluid

Author Affiliations

From the Chest Department, Veterans General Hospital–Taipei (Drs Chou, Lin, and Chang and Mss Tung and Liou), and the School of Medicine, National Yang-Ming University (Dr Chang), Taipei, Taiwan.

Arch Intern Med. 2001;161(4):562-566. doi:10.1001/archinte.161.4.562
Abstract

Background  The globules (stained green, orange, or orange in the center coated with a green rim) seen in Papanicolaou-stained smears of bronchoalveolar lavage fluid are suggested to be characteristic of pulmonary alveolar proteinosis (PAP).

Objective  To evaluate the usefulness of Papanicolaou-stained smears of bronchoalveolar lavage fluid in aiding a diagnosis of PAP.

Methods  Papanicolaou-stained smears of bronchoalveolar lavage fluid obtained from 7 patients (5 idiopathic, 2 secondary) with PAP were evaluated. To serve as controls, the smears of 11 normal subjects and 128 patients with other pulmonary disorders were also examined. The findings on the presence and number of globules were recorded. To differentiate PAP from other pulmonary disorders, the highest globule value obtained from the control group was chosen as the cutoff point.

Results  The characteristic globules were not found in normal subjects and only found in 6 of 128 patients with other pulmonary disorders. Their clinical diagnoses were Sjögren syndrome in 2 cases; polymyositis, idiopathic pulmonary fibrosis, asbestosis, and hypersensitivity pneumonitis in 1 case each. The numbers of globules in these 6 patients were 1, 3, 17, 7, 3, and 2. In contrast, more than 100 globules were found in all patients with PAP. The number of globules was highly sensitive and specific in aiding a diagnosis of PAP when the cutoff value was set at 18.

Conclusion  The globules seen in Papanicolaou-stained smears of bronchoalveolar lavage fluid may be valuable in aiding a diagnosis of PAP, especially when the number of globules is more than 18.

PULMONARY alveolar proteinosis (PAP) is a rare disease first described by Rosen et al1 in 1958. It is characterized by deposition within the air spaces of granular extracellular material composed of protein and lipids that is periodic acid–Schiff (PAS)-positive and diastase resistant.2,3 The onset of PAP is usually insidious and the initial symptoms are nonspecific, often causing a significant delay in the diagnosis. Patients with PAP may progress rapidly into severe respiratory distress. Therefore, a prompt diagnosis with the initiation of appropriate therapy with bronchoalveolar lavage (BAL) has clinical merit.

A diagnosis of PAP is based mainly on histopathologic findings obtained with transbronchial lung biopsy (TBLB) or open lung biopsy specimens. However, TBLB specimens are often so small that they do not provide specific evidence. In a 1998 article,4 the diagnostic yield of a TBLB specimen was only 29% in 68 patients with PAP. Bronchoalveolar lavage has proved valuable in providing a diagnosis of PAP as reported in some articles510; however, ultrastructural examination of BAL fluid (BALF) specimens and/or specific stains such as PAS and combined alcian blue and PAS stains are usually required.6,811 Clinically, it is impractical to process all routine BALF specimens for electron microscopy or special stains.

Cytologic findings of BALF specimens from patients with PAP have been reported,510 but the results of Papanicolaou-stained smears are insufficient for a definite diagnosis. A 1997 study9 suggested that a routinely processed cytologic smear of BALF with a Papanicolaou stain could be valuable in providing a diagnosis of PAP. The globules (stained green, orange, or orange in the center coated with a green rim) seen in Papanicolaou-stained smears as seen under a light microscope are the multilamellated structures characteristic of PAP as seen under an electron microscope and can be used as diagnostic evidence of PAP. To evaluate the usefulness of Papanicolaou-stained cytologic smears of BALF in aiding a diagnosis of PAP, 7 patients with PAP, 11 healthy subjects, and 128 patients with other pulmonary disorders were studied.

PATIENTS AND METHODS

From February 1997 to April 2000, 7 (5 idiopathic, 2 secondary) consecutive patients with PAP were included in this study. The diagnosis of PAP was established by pathologic examination of TBLB specimens with or without electron microscopy in 7 patients. In all 7 patients, BAL was performed to show the presence of PAS-positive intra-alveolar material. The clinical information about the patients with PAP is given in Table 1.

Bronchoalveolar lavage was performed prior to bronchial brushing or TBLB through a fiberoptic bronchoscope (model BF 20 or P20; Olympus, Tokyo, Japan) wedged in a segmental bronchus of the right middle lobe or lingula or other appropriate location, using 3 aliquots of a 50-mL sterile isotonic sodium chloride solution. Aspirates were pooled into a siliconized container and kept on ice during transport. Part of the retrieved BALF was subjected to Papanicolaou and Riu staining. Some slides were stored for subsequent staining by PAS and combined alcian blue and PAS.

To serve as controls, Papanicolaou-stained cytologic smears of BALF specimens from 11 healthy subjects and 128 patients with other pulmonary disorders were studied. To avoid bias, BAL was done by one of us (S.-C.C.) following the aforementioned protocol and the retrieved BALF was processed using the same protocol. Those with unknown entities of their pulmonary disorders were excluded. The clinical diagnoses of 128 patients with other pulmonary disorders are summarized in Table 2.

After a demonstration of characteristic globules seen in a Papanicolaou-stained smear of BALF specimen from the reported case, all Papanicolaou-stained smears were screened first by cytologic technicians (S.-M.T. and R.-D.L.), then were read separately by 2 physicians (C.-W.C. and F.-C.L.). Both cytologists and physicians were blind to the final diagnoses when they evaluated the smears. The findings on the presence and number of globules were recorded separately and then discussed, and the consensus findings were documented for the study. When the 2 readers could not reach consensus, the case was presented to a third expert reader (S.-C.C.), and the adjudicated reading became the final one.

REPORT OF A CASE

A 40-year-old housewife was admitted to the hospital in February 1997 because of exertional dyspnea for 1 year and shortness of breath for 2 months. She was found to have hypertension after pregnancy in 1990 and received antihypertensive treatment as needed. For 3 years she also had type 2 diabetes mellitus treated by diet only. One year prior to this admission, she began to experience exertional dyspnea and was admitted to another hospital 4 months later. After a series of examinations, idiopathic pulmonary fibrosis was diagnosed and the patient underwent steroid treatment. Two months before the admission, productive cough with whitish sputum, chest tightness, and shortness of breath developed in addition to progression of exertional dyspnea. On admission, a frontal chest radiograph revealed diffuse pulmonary infiltrates in both lungs, with lower lung fields predominantly affected. Thoracic high-resolution computed tomography revealed a diffuse ground-glass appearance in both lungs and a paved-stone arrangement. The findings of radiographic studies were in keeping with those of PAP. An elevated serum lactate dehydrogenase level (389 U/L) was noted. Pulmonary function testing showed mildly restrictive ventilatory defect with marked reduction of diffusing capacity for carbon monoxide (DLCO, 43 % of predicted value). Results of arterial blood gas studies, measured while the patient was breathing room air, were as follows: PaO2, 47.9 mm Hg; PaCO2, 34.6 mm Hg; and pH, 7.412. The patient underwent bronchoscopy with BAL and TBLB. Bronchoalveolar lavage was performed via a segmental bronchus of the right middle lobe, and TBLB was done via the right lower lobe after BAL. In gross appearance the retrieved lavage fluid was opaque and milky. A Papanicolaou-stained cytologic smear of BALF showed numerous, dense, amorphous globules stained green, orange, or orange in the center coated with a green rim dispersed within a background of finely granular amorphous material devoid of inflammatory cells (Figure 1, A). These globules appeared to be dark blue in a Riu-stained12 (a modified Wright-stain) smear (Figure 1, B). Strong positive PAS staining was noted in a PAS-stained smear. The combined alcian blue and PAS stain was performed later and showed globules with strong PAS staining of the proteinaceous material with a lack of significant alcian blue staining (Figure 1, C). Pathologic examination of TBLB specimens evidenced a diagnosis of PAP. Recognition of characteristic cytologic features of BALF specimens seen in the Papanicolaou-stained smear from this patient with PAP prompted us to evaluate the usefulness of Papanicolaou-stained smears of BALF in aiding a diagnosis of PAP.

RESULTS

The cytologic features in 11 BAL samples from 7 patients with PAP were similar. The gross appearance of the retrieved BALF was opaque and/or milky in all specimens and sediment was abundant. The cells were poorly preserved and abundant granular extracellular material was present. Total cell and differential cell counts could not be measured in 9 specimens obtained from 5 patients with idiopathic PAP.

Viewed under a light microscope, Papanicolaou-stained smears of BALF obtained from patients with PAP contained abundant amorphous granular and clumped extracellular material with a scarcity of cells. There were numerous globules stained green, orange, or orange in the center coated with a green rim interspersed in the amorphous material (Figure 1, A). The globules seen as solitary elements in Papanicolaou-stained smears were more than 100 globules per slide in all 7 patients with PAP. These globules appeared to be dark blue in Riu-stained smears (Figure 1, B). The globules and the amorphous material were PAS positive. Combined alcian blue and PAS-stained smears showed the globules with positive PAS staining of the proteinaceous material with a lack of significant alcian blue staining (Figure 1, C) in all patients with PAP.

In contrast, the globules could not be found in Papanicolaou-stained smears of BALF obtained from 11 healthy subjects. Furthermore, the globules were only found in 6 of 128 patients with other pulmonary disorders. The clinical diagnoses of the 6 patients were as follows: Sjögren syndrome in 2 cases and polymyositis, idiopathic pulmonary fibrosis, asbestosis, and hypersensitivity pneumonitis in 1 case each. The numbers of globules in these 6 patients were 1, 3, 17, 7, 3, and 2, respectively. Compared with those who had PAP, the smears obtained from patients with other pulmonary disorders revealed fewer globules, a small amount of extracellular material, and a greater number of cells. Globules with positive PAS staining were not found in combined alcian blue and PAS-stained smears of BALF obtained from all 6 patients with other pulmonary disorders.

The highest globule value seen in Papanicolaou-stained smears of BALF obtained from the control group was 17. To distinguish PAP from other pulmonary disorders, the value of 18 globules was chosen as a cutoff point. Using this cutoff point, the globules seen in Papanicolaou-stained smears seemed to be highly sensitive and specific in aiding a diagnosis of PAP (Table 3).

COMMENT

Pulmonary alveolar proteinosis is a disease characterized by phospholipoprotein material filling in the alveoli and terminal bronchial trees. Its diagnosis is mainly based on histopathologic findings obtained with TBLB or open lung biopsy specimens. However, the diagnostic yield of TBLB is usually unsatisfactory.3,4 Open lung biopsy may fail to ensure a diagnosis owing to sampling error and may not be without morbidity or mortality. Compared with TBLB and/or open lung biopsy, BAL can sample a much wider area and is safer for the patient.

With the advent of improved bronchoscopic techniques, BAL has become the standard for both the diagnosis9,13 and therapy3,1416 for PAP. The retrieved BALF of PAP is grossly opaque, milky white, or muddy. In Papanicolaou-stained smears of BALF specimens, the globules stained green, orange, or orange in the center with a green rim have been described to be characteristic of PAP.69 However, the characteristic cytologic findings in Papanicolaou-stained smears are considered insufficient for a diagnosis of PAP and ultrastructural examination of BALF is usually needed.6,811 Clinically, electron microscopy and special stains like PAS and combined alcian blue and PAS were not routinely performed for all BALF specimens because a diagnosis of PAP is not usually expected. This may explain in part why in most of the reported cases the diagnosis of PAP was established by open lung biopsy or often missed prior to autopsy.

It is of clinical significance if a routinely processed cytologic-staining method like a Papanicolaou stain proves to be useful in providing a diagnosis of PAP. The findings of a 1997 article9 confirmed that the globules stained green, orange, or orange in the center coated with a green rim seen in Papanicolaou-stained smears of BALF specimens were the multilamellated structures, characteristic of PAP when viewed under an electron microscope and concluded that these globules can be used as diagnostic evidence of PAP.

At variance with the results of the study by Mikami et al,9 the findings in our study indicated that the globules were not exclusively found in patients with PAP. The globules described in Papanicolaou-stained smears of BALF could be found in 6 of 128 patients with other pulmonary disorders. Their clinical diagnoses were Sjögren syndrome in 2 cases and polymyositis, idiopathic pulmonary fibrosis, asbestosis, and hypersensitivity pneumonitis in 1 case each. The number of globules in these 6 patients was 1, 3, 17, 7, 3, and 2, respectively. Of clinical interest, the globules with positive PAS staining were not found in the combined alcian blue and PAS-stained smears obtained from all 6 patients with other pulmonary disorders. This may be explained in part by the scarcity of globules. Another explanation is that the globules lack protein.

Compared with those who had other pulmonary disorders, the globules were numerous and there were more than 100 globules per slide in all 7 patients with idiopathic or secondary PAP. In addition, the presence of a large amount of amorphous and/or granular extracellular material and a scarcity of cells seemed to be valuable in distinguishing PAP from other pulmonary disorders. Furthermore, globules with positive PAS staining seen in combined alcian blue and PAS could be found in all 7 patients with PAP. These findings can be of clinical significance in differentiating PAP from other pulmonary disorders and in supporting a diagnosis of PAP.

It is important that a differential diagnosis be made between PAP and Pneumocystis carinii pneumonitis. In P carinii pneumonitis, the Papanicolaou-stained smears of BALF specimens contain characteristic findings described as a foamy mass or foamy exudate.17,18 The globules characteristic of PAP can be easily differentiated from this foamy mass or exudate described in P carinii pneumonitis by Papanicolaou staining. When the differentiation is difficult, combined alcian blue and PAS and Grocott methenamine silver staining methods can be used to make a differential diagnosis.

Immunosuppression seems to play a pivotal role in the defective alveolar clearance that subsequently leads to PAP.2 With the increased use of immunosuppressive and/or cytotoxic agents to treat a variety of immune-related diseases, the improved long-term survival of the organ transplants, and the improved outcome of the patients with different malignant neoplasms undergoing chemotherapy, secondary PAP can be expected to occur more frequently. It is important to have a high index of clinical suspicion to make a diagnosis. The characteristic gross appearance of a retrieved BALF specimen can be of help but may be easily missed owing to limited experience. Special stains for detecting PAS-positive material in a BALF specimen may not be done as needed. These may cause a significant delay in diagnosis or underdiagnosis of PAP.7 The presence of numerous globules in routinely processed Papanicolaou-stained smears of BALF specimens can be valuable in aiding a diagnosis of PAP as described in this article. A diagnosis of PAP can be further confirmed by PAS stain, combined alcian blue and PAS stain, or electron microscopy of BALF specimens. To avoid an unnecessary delay or miss in making a diagnosis of PAP, it is suggested to reserve some slides of BALF specimens for the patient undergoing BAL.

CONCLUSIONS

Our results suggested that the globules (stained green, orange, or orange in the center coated with a green rim) seen in Papanicolaou-stained smears of a BALF specimen can be of significant value in making a diagnosis of PAP, especially when the number of globules is more than 18. Other cytologic findings on the presence of a large amount of granular and/or amorphous extracellular material and the scarcity of cells may further support a diagnosis of PAP. When diagnosis is difficult, special stains like PAS and combined alcian blue and PAS stains or electron microscopy of BALF specimens may confirm the diagnosis of PAP.

Back to top
Article Information

Accepted for publication September 14, 2000.

Corresponding author: Shi-Chuan Chang, MD, Chest Department, Veterans General Hospital–Taipei, 201 Section 2, Shih-Pai Road, Shih-Pai, Taipei, Taiwan 112, Republic of China (e-mail: scchang@vghtpe.gov.tw).

References
1.
Rosen  SHCastleman  BLiebow  AA Pulmonary alveolar proteinosis. N Engl J Med. 1958;2581123- 1142Article
2.
Bedrossian  CWLuna  MAConklin  RHMiller  WC Alveolar proteinosis as a consequence of immunosuppression: a hypothesis based on clinical and pathologic observations. Hum Pathol. 1980;11(suppl 5)527- 535
3.
Prakash  UBBarham  SSCarpenter  HADines  DEMarsh  HM Pulmonary alveolar phospholipoproteinosis: experience with 34 cases and a review. Mayo Clin Proc. 1987;62499- 518Article
4.
Goldstein  LSKavuru  MSCurtis-McCarthy  PChristie  HAFarver  CStoller  JK Pulmonary alveolar proteinosis: clinical features and outcomes. Chest. 1998;1141357- 1362Article
5.
Martin  RJRogers  RMMyers  NM Pulmonary alveolar proteinosis: shunt fraction and lactic acid dehydrogenase concentration as aids to diagnosis. Am Rev Respir Dis. 1978;1171059- 1062
6.
Mermolja  MRott  TDebeljak  A Cytology of bronchoalveolar lavage in some rare pulmonary disorders: pulmonary alveolar proteinosis and amiodarone pulmonary toxicity. Cytopathology. 1994;59- 16Article
7.
Cordonnier  CFleury-Feith  JEscudier  EAtassi  KBernaudin  JF Secondary alveolar proteinosis is a reversible cause of respiratory failure in leukemic patients. Am J Respir Crit Care Med. 1994;149788- 794Article
8.
Burkhalter  ASilverman  JFHopkins  MBGeisinger  KR Bronchoalveolar lavage cytology in pulmonary alveolar proteinosis. Am J Clin Pathol. 1996;106504- 510
9.
Mikami  TYamamoto  YYokoyama  MOkayasu  I Pulmonary alveolar proteinosis: diagnosis using routinely processed smears of bronchoalveolar lavage fluid. J Clin Pathol. 1997;50981- 984Article
10.
Sosolik  RCGammon  RRJulius  CJAyers  LW Pulmonary alveolar proteinosis: a report of two cases with diagnostic features in bronchoalveolar lavage specimens. Acta Cytol. 1998;42377- 383Article
11.
Takemura  TFukuda  YHarrison  MFerrans  VJ Ultrastructural, histochemical, and freeze-fracture evaluation of multilamellated structures in human pulmonary alveolar proteinosis. Am J Anat. 1987;179258- 268Article
12.
Lee  CHLan  RSTsai  YHChiang  YCWang  WJ Riu's stain in the diagnosis pulmonary cryptococcosis: introduction of a new diagnostic method. Chest. 1988;93467- 470Article
13.
Wang  BMStern  EJSchmidt  RAPierson  DJ Diagnosing pulmonary alveolar proteinosis: a review and an update. Chest. 1997;111460- 466Article
14.
Ramirez  J Pulmonary alveolar proteinosis: treatment by massive bronchoalveolar lavage. Arch Intern Med. 1967;119147- 156Article
15.
Selecky  PAWasserman  KBenfield  JRLippmann  M The clinical and physiological effect of whole-lung lavage in pulmonary alveolar proteinosis: a ten-year experience. Ann Thorac Surg. 1977;24451- 461Article
16.
Du Bois  RMMcAllister  WABranthwaite  MA Alveolar proteinosis: diagnosis and treatment over a 10-year period. Thorax. 1983;38360- 363Article
17.
Greaves  TSStrigle  SM The recognition of Pneumocystis carinii in routine Papanicolaou-stained smears. Acta Cytol. 1985;29714- 720
18.
Ghali  VSGarcia  RLSkolom  J Fluorescence of Pneumonitis carinii in Papanicolaou smears. Hum Pathol. 1984;15907- 909Article
×