Regulation of chemokine mRNA expression in a rat model of vanadium-induced pulmonary inflammation

Environmental and occupational exposure to vanadium dusts results in toxic effects mainly confined to the respiratory system. Using a rat model of acute lung inflammation induced by intratracheal instillation of sodium metavanadate (NaVO3) at the dose of 200 microg V/kg, we investigated the relationship between the cytologic characterization of pulmonary inflammation and the expression of chemokine mRNA. Significant polymorphonuclear leukocyte (PMN) influx (P < 0.01) into the lung was noted 4 h after NaVO3 instillation, whereas alveolar macrophages (AMs) in bronchoalveolar lavage (BAL) cells appeared to decrease significantly. In contrast, neither PMNs nor AMs changed substantially 1 h after NaVO3 instillation. By Northern analysis, macrophage inflammatory protein (MIP)-2 mRNA in BAL cells increased markedly 1 h after NaVO3 instillation and reduced a little bit at 4 h, whereas MIP-1alpha mRNA in BAL cells was expressed relatively high 1 h after NaVO3 instillation, although a basal expression was detected in control group, and returned rapidly nearly to control level at 4 h. Since MIP-2 is a potent PMN chemoattractant and MIP-1alpha is a potent macrophage/monocyte chemoattractant has been well known. The facts that PMN influx was preceded by increased MIP-2 mRNA expression, suggesting that MIP-2 is involved in the development of NaVO3-induced pulmonary inflammation, whereas increased MIP-1alpha mRNA expression was followed by decreased AMs in BAL cells, suggesting AMs might be activated by MIP-1alpha, adherent to the lining surface of the airways and then resistant to be washed out. To delineate the mechanisms of transcriptional activation, we recently cloned the 5'-flanking region of the MIP-2 gene. The promotor region contains consensus binding sites for transcription factor nuclear factor kappaB (NF-kappaB) and activator protein-1 (AP-1). Using electrophoretic mobility shift assay, increased nuclear NF-kappaB, not AP-1, binding activity was detected 1 h after NaVO3 instillation, which correlated with the induction of MIP-2 mRNA. p65 (Rel A) and p50 protein appears to be involved in MIP-2 NF-kappaB binding. Taken together, our studies suggest that MIP-2 is an important mediator of NaVO3-induced pulmonary inflammation in the rat model. In addition, elevated MIP-2 mRNA levels are accompanied by increased NF-kappaB binding activity in BAL cells, suggesting possible MIP-2 transcriptional regulation through NF-kappaB.

Expression and regulation of macrophage inflammatory protein-2 gene by vanadium in mouse macrophages

Environmental and occupational exposure to vanadium (V) dusts results in inflammation mainly confined to the respiratory tract. Macrophages apparently play an important role in mediating the inflammation via the production of many chemokines. In the current study, we investigated whether vanadium can regulate the gene expression of a CXC chemokine macrophage inflammatory protein-2 (MIP-2), and to determine the molecular mechanisms controlling MIP-2 gene expression. A mouse macrophage cell line RAW 264.7 was treated with sodium metavanadate (NaVO3) at the dose of 0.5, 5, or 10 microg/mi V. Northern blot analysis showed that induction of MIP-2 mRNA expression was in a dose-dependent manner. To define the time course of the inflammatory response, RAW 264.7 cells were exposed to 5 microg/ml V, MIP-2 mRNA in macrophages increased markedly as early as 1 h after treatment, maximally induced at 4 h and reduced to 2-fold above control levels by 6 and 8 h. The protein levels of MIP-2 in conditioned media, measured by enzyme-linked immunosorbent assay (ELISA), was well correlated with the levels of MIP-2 mRNA following all of the treatments in the study. In addition, the increase in MIP-2 mRNA expression by vanadium was attenuated by co-treatment with the antioxidant N-acetylcysteine (NAC), at the doses of 10 and 20 mM, suggesting that the induction of MIP-2 mRNA is mediated via the generation of reactive oxygen species (ROS). To further investigate transcriptional regulation of the MIP-2 gene expression by vanadium, we performed RNA decay assay by measuring the half-life of MIP-2 mRNA. Co-treatment of macrophages with the transcriptional inhibitor actinomycin D at 5 microg/ml following exposure to 5 microg/ml V for 4 h revealed complete stabilization of vanadium-induced MIP-2 mRNA and no sign of mRNA degradation, at least, for 6 h, in comparison to the half-life of MIP-2 mRNA was approximately 2.5 h by bacterial lipopolysaccharide (LPS) treatment, supporting post-transcriptional stabilization as the predominant role of MIP-2 gene expression. In conclusion, these observations demonstrate that in vitro vanadium can induce MIP-2 mRNA expression, mediating, at least in part, via the production of ROS. In addition, the increase in MIP-2 mRNA level involves, most likely, post-transcriptional control via increased mRNA stability.

Serum hepatocyte growth factor levels in patients with inflammatory lung diseases

HGF is a pulmotrophic factor in the regeneration of an injured lung. However, the physiological role of HGF in vivo remains largely unknown. We studied HGF in patients with inflammatory lung diseases to investigate the clinical significance of HGF and compared with C-reactive protein (CRP) in inflammatory lung diseases. Forty-seven patients with inflammatory lung diseases (16 tuberculosis, 18 pneumonia, and 13 chronic obstructive pulmonary disease (COPD)) were studied. Fifty normal, healthy individuals were analyzed as normal control subjects. Serum HGF levels were measured by enzyme-linked immunosorbent assays (ELISA). Serum CRP levels were also performed. The mean +/- SE numbers of serum HGF levels in the patients with inflammatory lung diseases (4.33 +/- 0.41 ng/ml) were significantly elevated when compared with those in normal control subjects (0.36 +/- 0.02 ng/ml) (p < 0.0001). Serum HGF levels in patients with COPD was significantly lower than those were with tuberculosis or pneumonia (p < 0.05). There was a significant correlation between serum HGF levels and CRP in inflammatory pulmonary diseases (r = 0.48, p = 0.00087). The significantly decreased serum HGF levels in patients with improved inflammatory lung diseases were also observed subsequently. Our results suggest that secreted HGF may play an important role in bronchial epithelium reconstruction during respiratory inflammation.

Functional characterization of recombinant rat macrophage inflammatory protein-1 alpha and mRNA expression in pulmonary inflammation

Chemokines are important inflammatory mediators that function by activating and recruiting leukocytes to an inflamed tissue. We have recently cDNA cloned the rat chemokine macrophage inflammatory protein-1 alpha (MIP-1 alpha) (1). In the present study, we characterize the biological function of recombinant MIP-1 alpha protein and describe expression of its mRNA both in vitro and in a rat model of lung inflammation. In vitro rat rMIP-1 alpha protein was chemotactic for both polymorphonuclear leukocytes (PMNs) and macrophages with maximal activity at 50 nM for both cell types. In in vivo studies, we found that intratracheal instillation of 1 and 5 micrograms of rMIP-1 alpha resulted in a significant (P < 0.05) influx of cells, primarily monocytes/macrophages, into the airspace of the lungs after 6 h. Mean numbers of lavagable PMNs were not elevated significantly (P < 0.05) for either dose of MIP-1 alpha. As a model of inflammation, rats were intratracheally instilled with 0.1 mg/kg bacterial lipopolysaccharide (LPS). Bronchoalveolar lavage (BAL) was performed 3 h later. Instillation of LPS resulted in an acute neutrophilia, but no significant change in lavagable macrophages. BAL cells from control animals (saline instilled) displayed no basal mRNA expression of either MIP-1 alpha or MIP-2 (positive control). In contrast, both MIP-1 alpha and MIP-2 mRNA levels increased markedly in BAL cells from rats instilled with LPS. The rat alveolar macrophage cell line (NR8383) also showed increased MIP-1 alpha mRNA levels in response to LPS (10 micrograms/ml) with a maximal increase after 6-8 h. The induction of MIP-1 alpha mRNA expression by LPS in NR8383 cells was attenuated by cotreatment with the antioxidants N-acetylcysteine and dimethylsulfoxide, suggesting that the induction of MIP-1 alpha mRNA by LPS is mediated via the generation of reactive oxygen species. We conclude that MIP-1 alpha is a potent chemoattractant for macrophages in vivo, and its mRNA expression in macrophages and BAL cells in response to inflammatory stimuli suggests a fundamental role in acute pulmonary inflammation.

Heparin-binding epidermal growth factor-like growth factor and transforming growth factor-alpha in human non-small cell lung cancers

Transforming growth factor-alpha (TGF-alpha), a member of the epidermal growth factor (EGF) family that binds to the EGF receptor (EGFR), is thought to function in an autocrine manner in non-small cell lung cancers (NSCLC). Heparin-binding EGF-like growth factor (HB-EGF), a novel member of the EGF family, also binds to EGFR. To compare the expression of HB-EGF, TGF-alpha and EGFR genes in NSCLC and normal lung tissue, we measured the levels of messenger RNA (mRNA) for these genes in human NSCLC and normal lung tissues by Northern hybridization, reverse transcription-polymerase chain reaction (RT-PCR), and in situ hybridization. A total of eight specimens (paired tumor tissue and normal lung tissue) were harvested from four patients who underwent resection of primary resectable NSCLC. HB-EGF was not expressed in either tumor tissue or normal lung tissue, while EGFR and TGF-alpha were expressed in all samples. TGF-alpha was overexpressed in all tumor tissue samples by several hundred-fold, while the expression of EGFR was not significantly different in tumor tissue and normal lung tissue. There was no correlation between the expression of TGF-alpha and EGFR. In situ hybridization showed that TGF-alpha mRNA was localized mainly in the cancer cells of tumor tissues and in the macrophages of alveoli in normal lung tissue. Our results showed that HB-EGF plays no role in the growth of NSCLC, and that there was no significant overexpression of EGFR in tumor tissue. TGF-alpha may play a major role in the growth of NSCLC. This supports a new direction in rational NSCLC treatment.

Comparison of cytokeratin fragment 19 (CYFRA 21-1), tissue polypeptide antigen (TPA) and carcinoembryonic antigen (CEA) as tumour markers in bronchogenic carcinoma

To elevate the diagnostic value of the serum cytokeratin 19 fragment (CYFRA 21-1) and compare it with carcinoembryonic antigen (CEA) and tissue polypeptide antigen (TPA) in bronchogenic carcinoma, the sera of 161 patients (58 with benign pulmonary disease and 103 with bronchogenic carcinoma) was investigated using immunoradiometric assay. Sensitivities for CYFRA 21-1, CEA and TPA (using 3.5 ng ml-1, 5.0 ng ml-1, 110 U l-1, respectively, cut-off values corresponding to a 95% specificity for benign pulmonary disease) in bronchogenic carcinoma were 64, 47 and 61%, respectively. Positive CYFRA 21-1 levels were identified in 75% of patients with squamous cell carcinoma (n = 36), in 67% with adenocarcinoma (n = 45), in 17% with large cell carcinoma (n = 6), and in 50% with small cell lung cancer (SCLC) (n = 16). However, CYFRA 21-1 levels were not significantly different between squamous cell carcinoma and the other histological types. The sensitivity of the combined measurement of CYFRA 21-1 with any other tumour marker was significantly higher than that of CYFRA 21-1 measurement alone. Elevated CYFRA 21-1 levels were observed in 44% of Stages I and II (n = 18) and 72% of Stage III and IV (n = 69) patients with non-small cell lung cancer (P < 0.05). A significant inter-marker correlation was observed between CYFRA 21-1 and TPA (n = 103, r = 0.448, P < 0.0001). Twenty-one patients were monitored by CYFRA 21-1, and significantly different changes in progressive patients (P = 0.0058) and regressive patients (P = 0.016) were obtained. These results indicate that CYFRA 21-1 may be not only a sensitive tumour marker in the diagnosis of bronchogenic carcinoma, but also a useful marker for the monitoring of bronchogenic carcinoma.

Spirometry in life-long non-smoking, healthy Chinese women in Taiwan

To establish the spirometric values for normal, healthy Chinese women in Taiwan, the spirometry of 506 life-long non-smoking, healthy Chinese women was examined, including 140 subjects over the age of 60 years. Significant correlations among age, height and forced vital capacity (FVC), forced expiratory volume in 1 s (FEV1), FEV1/FVC%, peak expiratory flow (PEF), Vmax75, Vmax50, Vmax25 were found. However, there were no significant correlations between age and FEV1/FVC%, nor age and Vmax25 in the elderly group. FEV1 (FEV1*) and FVC (FVC*) were standardized to the overall mean height for elderly women using Cole's formula. The decline in FEV1* and FVC* with age were observed. The predicted value for the average 70-year-old woman with a height of 1.5 m derived from the present study is compared with those from other surveys of the elderly. The values from the present study are somewhat higher than the values from the Hong Kong study. The authors believe the fact that all of the present subjects were life-long non-smokers might explain the differences.

Cytokeratin fragment 19 (CYFRA 21-1) as a tumor marker in non-small cell lung cancer

To evaluate the diagnostic value of the serum cytokeratin 19 fragment (CYFRA 21-1) in bronchogenic carcinoma, we investigated the sera of 138 patients (58 with benign pulmonary disease and 80 with non-small cell lung cancer (NSCLC)) using immunoradiometric assay. The mean (SD) value of serum CYFRA 21-1 in NSCLC (13.26 (16.54)) was significantly higher than in benign lung diseases (1.74 (1.55)) (p < 0.0001). Sensitivity for CYFRA 21-1 (using 3.5 ng/ml, a cut-off value corresponding to a 95% specificity for benign pulmonary disease) in NSCL was 62%. Positive CYFRA 21-1 levels were significantly higher in 75% of patients with squamous cell carcinoma (n = 36) than in 53% with other NSCLC (n = 44) (p < 0.05). CYFRA 21-1 levels were significantly different between squamous cell carcinoma (17.28 (19.94)) and the other NSCLC (9.96 (12.44)) (P < 0.05). Elevated CYFRA 21-1 levels in patients with stage III and IV disease (n = 64, 18.19 (26.51)) were significantly higher than in stage I and II (n = 16, 4.41 (5.76)) (p < 0.02). The positive rate of CYFRA 21-1 in tumor stage I and II was only 37%. Our results indicate that CYFRA 21-1 may be a useful tumor marker in NSCLC, especially in squamous cell carcinoma. However, CYFRA 21-1 cannot be used for the diagnosis of early stage disease of NSCLC. CYFRA 21-1 may also contribute to the monitoring of NSCLC.

Allergic bronchopulmonary aspergillosis: a case report

Aspergillus-associated pulmonary diseases are aspergilloma, invasive aspergillosis, and allergic bronchopulmonary aspergillosis. Allergic bronchopulmonary aspergillosis is caused by a complex of immunologic reactions to the presence of the Aspergillus species colonizing the bronchial trees. The disease is not common in Taiwan. The major diagnostic criteria for allergic bronchopulmonary aspergillosis are 1) bronchial asthma, 2) pulmonary infiltration, 3) peripheral eosinophilia, 4) positive skin test to Aspergillus fumigatus, 5) serum precipitin to Aspergillus fumigatus, 6) elevated serum Ig E, and 7) central bronchiectasis. We report a case who has had a chronic asthmatic-like cough for 5 years. He worked in a silo for two years before he was troubled by the disease. He was admitted to hospitals four times in the past, and received five bronchoscopic examinations and one open lung biopsy without definite diagnosis. Sputum eosinophilia directed our attention to the differentiation of eosinophilic lung diseases. A bronchogram which revealed central brochiectasis helped us to make the diagnosis of allergic bronchopulmonary aspergillosis, despite negative sputum culture for Aspergillus fumigatus and negative serum precipitin to Aspergillus fumigatus.

Nucleolar organizer regions in small cell carcinoma of lung

Agyrophilic staining of nucleolar organizer regions (NOR) has been used to differentiate between cells of small cell carcinoma and lymphocytes in tissue specimens. We used cytologic smears which were previously Papanicolaou-destained to study the one-step agyrophilic staining technique for nucleolar organizer regions (AgNOR) in cell of small cell carcinoma and lymphocyte. The purposes of this study were to assess the feasibility and usefulness of AgNOR staining in diagnostic cytology and to try to set up a procedure that could be used on prestained smears for retrospective study. While the NOR of each lymphocyte appeared to be one round dot after AgNOR staining, the NOR of cell of small cell carcinoma showed dots, slightly variable in size and shape. The mean number of NOR was significantly higher (p < 0.01) in cells of small cell carcinoma (4.7 +/- 0.7) than in lymphocytes (1.4 +/- 0.4). In conclusion, AgNOR staining was demonstrated to be a useful method to differentiate between cells of small cell carcinoma and lymphocyte in Papanicolaou-destained smears.

Nucleolar organizer regions in smears of pleural effusion

Agyrophil staining was applied to nucleolar organizer regions (NOR) to differentiate cells of adenocarcinoma and histiomesotheliosis in pleural effusion. The smears were either nuclearly unstained, but cytoplasmically counterstained by Papanicolaou method or Papanicolaou-destained before agyrophil staining of nucleolar organizer regions (AgNOR). The purposes of this study were to assess the feasibility and usefulness of AgNOR staining in diagnostic cytology and to try to set up a procedure that could be used on prestained smears for retrospective study. All smears showed good background and cellular outline. The distribution of NOR was either intranuclearly or in the nucleoplasm diffusely. In previously nuclearly unstained smears, NOR showed granular to powder-like appearance. The mean number of NOR in adenocarcinoma (38.4 +/- 12.5) was significantly higher than that in histiomesotheliosis (15.6 +/- 2.9). In Papanicolaou-destained smears, the NOR showed confluent dots. The mean number of NOR was much lower as compared to that of previously nuclearly unstained smears. Furthermore, the mean number of NOR in adenocarcinoma (3.6 +/- 1.4) showed no significant difference with that of histiomesotheliosis (2.7 +/- 0.8). In conclusion, AgNOR staining is one of the methods to differentiate benign from malignant cells, but not in Papanicolaou-destained smears.

Clinical study of Mycoplasma pneumoniae pneumonia

From 1982 to 1991, we experienced 76 patients with Mycoplasma pneumoniae pneumonia which were confirmed by serologic tests. There were 32 (42%) male and 44 (58%) female patients. One patient had underlying disease of diabetes mellitus while the other patients were in good health. The age ranged from 9 months old to 72 years old. All the patients complained of fever and coughing; 63% had dry cough and 37% had sputum production. Upper respiratory tract complaints such as rhinorrhea, sore throat, or earache were noted in 57% of the patients. Fifty-five percent of the patients had GI symptoms of anorexia, nausea, vomiting, or diarrhea. Other complaints included myalgia/arthralgia (29%), headache (30%), and general malaise (32%). Dyspnea (17%) and chest pain (20%) were occasional complaints. Seventy-one percent of the patients had WBC counts < 10000/cu mm and 29% > 10000/cu mm. The mean value of C-reactive protein (CRP) was 53.1 micrograms/ml, while 16% of the patients had a CRP value above 100 micrograms/ml. Thirty-one percent of the patients were noted to have a transient elevation of serum transaminase. Four different patterns of infiltration were seen in chest radiographic manifestation: 1) peribronchial and perivascular interstitial infiltrates (18.4%), 2) nonhomogeneous patchy consolidations (22.4%), 3) homogeneous acinar consolidations (27.6%), and 4) mixed interstitial and alveolar infiltrates (27.6%). Interstitial infiltration was more commonly seen in pediatric than adult patients (46% vs 20%). Other features of the radiologic manifestation were as follows: unilateral lesions in 80% of patients, single lobe lesions in 77%, lower lobe predominant in 69%, pleural effusion in 7%, and radiographic deterioration in 10%. Mycoplasmal pneumonia should be considered in the differential diagnosis of community-acquired pneumonias.

Flow cytometric DNA analysis of pleural effusions

A total of 71 cases of pleural effusion in patients with and without cancer were analyzed by conventional cytology and flow cytometry (FCM) in order to detect cells with an abnormal DNA content (aneuploidy). For cytologic examination, the samples were prepared using standard techniques. Sample for FCM analysis were centrifuged and exposed to hypotonic solution containing detergent and propidium iodide. Thirty-eight patients had pleural effusion due to benign disease, whilst 33 patients had primary lung cancer. All 38 patients with benign pleural effusions showed FCM diploidy. There were 17 aneuploidy (52%) and 16 diploidy (48%) in the 33 patients with lung cancer by FCM analysis. Four of these 33 effusions were cytologically negative, however, FCM showed aneuploidy in 2 of these 4 patients. Based on these results, FCM analysis combined with conventional cytopathology yielded 100% specificity, 94% sensitivity and 100% predictive value of positive result. There were no false-positive results but 2 false-negative results. These findings suggest that FCM is a rapid and useful technique in the analysis of pleural effusion and can be a very useful adjunct to conventional cytopathology.

Application of staining of nucleolar organizer regions in cytological smears of the bronchus

An argyrophil technique for the staining nucleolar organizer regions (AgNOR) was applied to cytological preparations obtained from bronchoscopic brushing materials. The number of AgNOR has been thought to be related to cellular activation. To differentiate malignant cells from non-malignant atypical cells, this study was carried out in 20 cases of adenocarcinoma of the lung (mean AgNOR: 18.34), 12 cases of pulmonary inflammatory diseases (mean AgNOR: 6.54) and 10 normal bronchial epithelial specimens for control (mean AgNOR: 2.07). On the basis of AgNOR number, we could differentiate between the three groups. The differences observed were statistically highly significant (p less than 0.0001). Moreover, the nucleolar organizer regions (NOR) in cancer cells were found the more irregularly distributed and more variable in sized than those in atypical and normal bronchial columnar cells. We suggest that the AgNOR technique will find increasing application as a complementary test in diagnostic cytopathology.

Comparison of pulmonary tuberculosis in younger and elderly patients

Pulmonary tuberculosis remains a significant clinical and public health problem in the elderly population. To describe age-related differences in disease manifestations, a comparison of the clinical features, predisposing factors, diagnostic approaches and radiographic findings in cases of pulmonary tuberculosis among 52 young and 62 elderly patients was performed. The elderly patients had a higher number of underlying disease than younger patients (p less than 0.05). Prior to admission, symptoms occurring with equal frequency in both younger and elderly patients included coughing, malaise, and weight loss. Elderly patients had significantly higher incidences of negative reactions to the PPD test (p less than 0.05). Radiographic findings revealed that upper lung field infiltrates were still common in both groups, but the elderly had more severe lung field involvement (two or more lobes affected), and more frequent pleural reactions than younger patients (p less than 0.05). Since there were non-specific clinical features, false negative skin test and complex radiographic manifestations, tuberculosis was frequently not suspected in the differential diagnosis, especially among elderly patients with multiple medical problems. We suggest that physicians need to have a high level of suspicion and awareness of varied manifestations for tuberculosis, especially elderly patients.

Quantitative C-reactive protein in pulmonary infections

A total of 207 cases were selected to evaluate the diagnostic value of C-reactive protein (CRP) in pulmonary infections. The mean +/- SD of CRP values in various pulmonary infections were as follows: 18.62 +/- 11.34 micrograms/ml for 32 cases of exudative-fibrotic tuberculosis; 15.98 +/- 16.66 micrograms/ml for 15 cases of tuberculous pneumonia; 25.61 +/- 18.96 micrograms/ml for 29 cases of tuberculous effusion; 16.66 +/- 10.18 micrograms/ml for 11 cases of tuberculous cavity; 81.1 +/- 24.9 micrograms/ml for 10 cases of miliary tuberculosis; 36.4 +/- 22.1 micrograms/ml for 19 cases of mycoplasmal pneumonia; 241 +/- 72 micrograms/ml for 38 cases of bacterial pneumonia; 225 +/- 65 micrograms/ml for 30 cases of bacterial pneumonia with effusion; 169 +/- 50 micrograms/ml for 16 cases of lung abscess. The CRP values of other pulmonary infections were as follows: 20.6, 20.8 micrograms/ml for two cases of Strongyloides stercoralis pneumonia; 7.4, 1.6 micrograms/ml for two cases of aspergilloma; 11.2, 12.4, 7.6 micrograms/ml for three cases of Pneumocystis carinii pneumonia. Serial changes in CRP values in 13 cases of well-treated bacterial pneumonia showed that values of CRP decreased to below half of the initial value on the 3rd to 4th day, and returned to about normal value on the 10th to 13th day. The study suggested that: a) various types of infections had different levels of CRP values, b) level of CRP values was determined both by the pathogen and the severity of inflammation, c) serial CRP values in bacterial infection could be used as a guide in treatment.

Diagnostic value of ferritin measurements in patients with pleural effusions

The diagnostic value of ferritin measurements has been evaluated in 59 patients with pleural effusions (PE). The Mean(+/- standard error [SE]) PE-ferritin level was 1913 +/- 720 ng/ml in 10 patients with nonmalignant nontuberculous (nonTB) exudates (Group I), 594 +/- 110 ng/ml in 14 patients with TB effusions (Group II), 1286 +/- 194 ng/ml in 28 patients with malignant exudates (Group III), and 280 +/- 60 ng/ml in 7 patients with transudates (Group IV). Significant difference was found between Group II and Group III (p less than 0.05). The mean (+/- SE) ratio of PE/serum (PE/S) ferritin in Groups I, II, III, and IV was 3.5 +/- 1.1, 10.5 +/- 5.1, 4.9 +/- 1.0, and 2.0 +/- 0.7, respectively, and it showed no significant difference among Groups I, II, and III. There were good correlations between PE-ferritin and PE-LDH or S-ferritin, but no correlation between PE-ferritin and PE-protein. PE-ferritin revealed a considerable overlap among the exudative groups, and was of no value in the differentiation between malignant and nonmalignant (Group I + II) exudates. In the discrimination between exudates and transudates, the diagnostic accuracy (86%) of PE-ferritin was inferior to those (all 98%) of PE-protein, PE-LDH, PE/S protein ratio, and PE/S LDH ratio. Using PE-ferritin greater than or equal to 500 ng/ml in distinguishing malignant exudates from TB effusions, the sensitivity was 79%, specificity 57% and accuracy 71%.(ABSTRACT TRUNCATED AT 250 WORDS)