pleural fluid ferritin concentrations-in disease · congestive heart failure-- - 102 (50-840) 4...

J Clin Pathol 1985;38:818-824

Pleural fluid ferritin concentrations-in human diseaseMATTI KLOCKARS,§ TEDDY WEBER,t PIRJO TANNER,tPAHR-EINAR HELLSTROM,t TOM PETTERSSON*

From the *Fourth Department of Medicine, Helsinki University Central Hospital, Helsinki, Finland, thetMinerva Institute for Medical Research, and Medix Laboratories Ltd, Helsinki, tMjolbolsta Hospital,Finland, and the §Institute of Occupational Health, Helsinki, Finland

SUMMARY The concentration of ferritin was measured in the pleural fluid of 108 patients withpleural effusions. In all groups of patients the ferritin concentration was higher in pleural fluidthan in serum. The greatest differences, with up to 100 times more ferritin in the pleural fluid,were found for patients with rheumatoid pleurisy, malignant effusions, and empyema. In patientswith non-malignant inflammatory pleural effusions the concentration of ferritin in pleural fluidcorrelated significantly with other pleural fluid indices of inflammation: there was a positivecorrelation with lactate dehydrogenase activity and a negative correlation with concentrations ofglucose and complement components C3 and C4. Ferritin was detected immunocytochemicallyonly in the macrophages found among the pleural fluid cells.Our study shows that large amounts of ferritin accumulate locally in the pleural cavity in

certain types of pleural inflammation. The accumulation is probably partly the result of increasedlocal reticuloendothelial system activity. Determination of the concentration of ferritin in pleuralfluid may provide corroborative information for differential diagnosis and may further our under-standing of the pathogenetic events that lead to the perpetuation of inflammatory activity inpleural effusions.

Ferritin, an iron storage protein with a molecularweight of about 450 000, in apoferritin form, isfound mainly in the liver and in cells of thereticuloendothelial (mononuclear phagocyte) sys-tem. In healthy adults the concentration of ferritin inserum is positively correlated with the iron stores inthe body.'

Determination of serum ferritin is useful in thediagnosis of iron deficiency anaemia and ironoverload-for example, haemochromatosis.Increased serum concentrations of ferritin havebeen found in patients with severe liver disease-forexample, patients with alcoholic liver disease,2 andviral or drug induced hepatic necrosis.3 High con-centrations of serum ferritin have also been reportedfor patients with malignancies, particularly thosewith cancer of the breast, lung, pancreas, or liver,but also for some patients with leukaemia andHodgkin's disease.'4 In the inflammatory responseferritin possibly behaves as an acute phase pro-tein57'; in rheumatoid arthritis, concentrations offerritin in synovial fluid have been found to beincreased when compared with serum values.8

Accepted for publication 5 February 1985

This study was undertaken to determine the diag-nostic value of measuring the concentration of ferri-tin in the serum and pleural fluid of patients withpleural effusions caused by different diseases.Biochemical and immunological measurements ofinflammatory activity in the pleural cavity were cor-related with the results of ferritin assays in order toelucidate the pathogenetic events of pleural effu-sions.

Patients and methods

One hundred and eight patients with pleural effu-sion were studied. Paired samples of serum andpleural fluid were collected from 70 patients, and forthe remainder only pleural fluid was studied. Thecause of the pleural effusion was determined foreach patient from clinical, laboratory, and radiologi-cal findings. The patients were divided into groupson the basis of their final diag oeis (Table 1).

Total and differential leucocyte counts, erythro-cyte counts, lactate dehydrogenase activity, andconcentrations of total protein, glucose, and com-plement components C3 and C4 were determined asdescribed previously.9

818

on 12 March 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.38.7.818 on 1 July 1985. Dow

nloaded from

http://jcp.bmj.com/

Pleural fluid ferritin concentrations in human disease

Table 1 Groups ofpatients studied

No ofpatents Sex ratio (M:F) Age (yr) Mean (range) Diagnosis

5 5:0 61 (35-76) Parapneumonic effusion5 5:0 71 (56-84) Bacterial empyema

12 10:2 57 (26-80) Tuberculous pleurisy*9 8:1 54 (39-70) Rheumatoid arthritist3 2:1 35 (18-51) Systemic lupus erythematosust

51 29:22 55 (33-84) Pleural effusions caused by cancer: 26 patients withpulmonary carcinoma and 4 with malignantmesothelioma; 12 patients with extrapulmonarycancer; 9 patients with cancer and evidence of livermetastases

4 0:4 77 (56-86) Transudative pleural effusions caused by congestiveheart failure

19 13:6 61 (32-83) Non-specific pleural effusionsDiagnosis was made after the above mentioned types of

pleurisy had been ruled out

*In seven patients the diagnosis was based on a positive culture of Mycobacterium tuberculosis or a pleural biopsy specimen compatiblewith active tuberculosis. In five patients the diagnosis was based on clinical findings and a favourable response to specific antituberculoustreatment.tDiagnosed according to criteria of the American Rheumatism Association.

Table 2 Median concentration ofserum ferritin for each group ofpatients

Diagnosis Men Women

Serum ferritin concentration No ofpatents Serum ferritin concentration No ofpatients(pg/l) Median (range) (pgll) Median (range)

Inflammatory pleural effusion* 210 (50-1580) 31 138 (51-430) 7Empyema 269 (55-1750) 5 - -Congestive heart failure -- - 102 (50-840) 4Cancer 230 (78-560) 12 165 (18-290) 8(3 patients with liver metastases

were excluded)

*Tuberculosis, pneumonia, rheumatoid arthritis, systemic lupus erythematosus, and non-specific pleural effusions.

Pleural fluid was analysed cytologically andstained and cultured for the presence of bacteria,including M tuberculosis. For measurement of ferri-tin concentration blood and pleural fluid were col-lected on the same day. After centrifugation theserum and pleural fluid were stored at - 20°C untilassayed. Ferritin was measured by animmunoradiometric method (Ramco LaboratoriesInc, Houston, Texas)'"0 with human spleen ferritinas the standard. Recovery of spleen ferritin added topleural fluid samples was 93-118%, mean 108%(n = 6). Dilution curves of pleural fluid samplesparalleled those of the standards, which indicatesimmunological identity of pleural fluid and spleenferritin in the assay. The reference values for theconcentration of ferritin in serum were 20-120 ,ug/lfor women and 40-240 ,ug/l for men. Because ofdifferences between men's and women's concentra-tions of ferritin, the results are presented separately.Immunocytochemical visualisation of ferritin in

pleural fluid cells was performed using animmunoperoxidase technique.'2 Cytocentrifugedpleural fluid cells were fixed in formol-calcium(Baker's solution). The cells were then washed, andendogenous peroxidase in the cells was inactivatedby treatment with methanol and hydrogen peroxide.

The cells were washed again in phosphate bufferedsaline and then treated for 30 min in successionwith rabbit antihuman ferritin (dilution 1/200,Dakopatts, Copenhagen, Denmark), sheep antirab-bit gammaglobulin (1/20), and rabbit antihorserad-ish peroxidase (1/150) Between each application ofantiserum the cells were washed twice for 5 min withphosphate buffered saline. After incubation withhorseradish peroxidase the enzyme was identified bythe method of Graham and Karnovsky.'3 Controlstaining was performed either with an antiserumabsorbed with ferritin (Sigma Chemical Company,USA) or by omitting the primary antiserum.

Statistical significance was tested by Wilcoxon'srank test and linear regression analysis.

Results

Table 2 gives the median concentration of ferritin inserum for each group of patients. No significant dif-ferences were found between the values.

Figs. 1 and 2 show the concentrations of ferritin inthe pleural fluid of patients with non-malignant dis-eases and of patients with cancer, respectively.Median values are given in Table 3. Significant dif-ferences between the median values of ferritin in the

819



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Klockars, Weber, Tanner, Hellstrom, Pettersson

Pneum Empyemac Tbc Non-specific RA SLE CHF(5) (5) (10) (2) (13) (6) (8) (1) (2) (1) (4)

)00* 0

*0 0 0 0

Range M 540-1300F

860-23 000 390- 2050 39-34 500 1420-72900 188-1750- 440-1300 180-1290 - - 68-1700

Fig. 1 Concentrations offerritin in pleural fluid. Values for men (@) and women (0) are shownseparately. Pneum = pneumonia; Tbc = tuberculosis; RA = rheumatoid arthritis; SLE = systemic lupuserythematosus; CHF = congestive heart failure.

Table 3 Median concentration offerritin in pleural-fluid for each group ofpatients

Diagnosis Men Women

Pleural feuid ferritin No ofpatients Pleural flid ferritin No ofpatientsconcentration (g/lt) Median concentration (pgll) Median

Pneumonia 915 5 -Empyema 7000 5 -

Tuberculosis 1685 10 870 2Non-specific pleural effusions 780 13 350 6Rheumatoid arthritis 5390 8 1800 1Systemic lupus erythematosus 970 2 1450 1Congestive heart failure - - 1265 4Pulmonary carcinoma 1765 22 1200 8Extrapulmonary carcinoma 3900 2 585 10Cancer and evidence of liver 7900 5 1895 4

metastases

820

25000 -

10000-

I-

4,1000

2 2500-

1000 -

500 -



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


ExtraPulmonary pulmonary Livercarcinoma carcinoma metastases

(8) j (10) (4)0 00

0

0

0

00

Range 355-3675 61-1790 610-10700 230-5700 1750-6050 2715-9550

Fig. 2 Concentration offerritin in the pleural fluid ofmen (X) and women (0) with cancer. The medianconcentration is indicated by the bar and the range is seen below the figure.

0

Oooo

0

0

0

0

0

0

0

.0 0.-

pleural fluid were as follows: men with rheumatoidarthritis had significantly higher concentrations thanmen with non-specific pleural effusions (p < 0 05)or pneumonia (p < 0-01); men with empyema hadsignificantly higher concentrations than men withnon-specific pleural effusions (p < 0.05); and menwith cancer and signs of liver metastases hadsignificantly higher concentrations than men withnon-specific pleural effusions (p < 0-05), tuber-culosis (p < 0-001), pneumonia (p < 0.05), orcancer of the lung without liver metastases(p < 0-05).

Table 4 shows the clinical data, serum and pleuralfluid ferritin concentrations, and other indices ofinflammatory activity in the pleural fluid of ninepatients with rheumatoid arthritis. High pleural fluidferritin concentrations are associated with increasedlactate dehydrogenase values and low concentra-tions of complement components in the pleuralfluid.

Table 5 shows the relation between the ferritinconcentrations in pleural fluid and other laboratoryindices of pleural inflammation in men with non-malignant inflammatory pleural effusions-that is,patients with tuberculosis, pneumonia, rheumatoidarthritis, systemic lupus erythematosus, and non-specific pleural effusions. There was a highlysignificant correlation (p < 0.001) between the con-centration of ferritin and lactate dehydrogenaseactivity and an inverse correlation between the con-centrations of ferritin and complement componentC3.Immunocytochemical staining of ferritin in

pleural fluid cells showed strong positive staining invacuolated macrophages (Fig. 3a), whereasmesothelial cells, granulocytes, and lymphocyteswere not stained for ferritin. Fig. 3b shows theabsence of specific staining in a cell preparation inwhich primary rabbit antiferritin antiserum wasreplaced with an antiserum absorbed with purified

10000 -

5000-

-k2500-c

.-

to 1000-

a-

500

ExtraPulmonary pulmonary Livercarcinoma carcinoma metastQses

(22) (2) (5)0.

00 0@0

0~~

0

00

000

0

a

821

e



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Klockars, Weber, Tanner, Hellstrom, Pettersson

Table 4 Clinical data for nine patients with seropositive rheumatoid arthritis and pleural effusion

Sex Age (yr) Serum ferritin Pleural fluid

Ferritin Leucocyte count Protein Glucose Lactate dehydrogenase C3 (gll) C4 (gll)(pglI) (x I04/l) (gll) (mmolll) (U/1)

M 51 388 27500 0-62 76 0-1 6472 0-05 0 01F 62 51 1800 1-25 34 0 3776 0 04 0-02M 54 81 1495 1-70 39 0 3 944 0-16 0-01M 70 120 9000 0-57 35 0-2 2016 0 05 0-02M 39 859 72900 ND 54 0-1 ND ND NDM 55 195 1420 3 90 57 1-1 891 0-22 0 03M 51 189 1626 4-20 49 1-2 1316 0 34 0-02M 44 50 35900 0 55 67 0 3925 0-05 0-02M 58 1580 1775 ND ND 0-6 ND ND ND

M = men, F = women.ND = not done.

Table 5 Relation between the concentrations ofpleuralfluid ferritin and pleural fluid indices ofdisease activity inmen with non-malignant inflammatory pleural effusions

Correlation (r) Significance (p)

Absolute leucocyte count -0-05 NS(n = 31)

Total protein (n = 29) 0-33 NSGlucose (n = 29) -0-51 <0-01Lactate dehydrogenase 0-71 <0-001

(n = 29)C3 (n = 31) -0-56 <0-001C4 (n = 31) -0-42 <0.05

NS = not significant.

human ferritin.

Discussion

Determination of biochemical, immunological, andcellular variables in pleural fluid has helped in boththe diagnosis and understanding of the pathogeneticevents of exudative pleural effusions.'4 In the pres-ent study the concentration of ferritin was higher inpleural fluid than in serum for all the groups ofpatients, including patients with transudative effu-sions caused by congestive heart failure. The con-centrations of ferritin in the pleural fluid were fourto one hundred times higher than values in theblood, the greatest difference being for patients withempyema, cancer, or rheumatoid arthritis. Althoughthere are reports of increased serum ferritin con-centrations in patients with some types of cancer, 4

we did not find significantly higher values in patientswith cancer of the lung when compared with patientswho had inflammatory lung diseases. Measurementof ferritin concentration in pleural fluid, however,gave some hint of the diagnosis of malignant pleuraleffusion. Patients with cancer but without livermetastases had high concentrations of ferritin(above 1000 ,ug/l) in pleural fluid; such high valueswere found for two thirds of the men and one thirdof the women. Furthermore, patients with cancerand with liver metastases had even higher ferritin

concentrations in pleural fluid.The turnover and metabolism of ferritin in effu-

sions is poorly understood. A delayed diffusion oflocally produced, high molecular weight ferritinfrom pleural fluid into the circulation could accountfor the high ferritin concentrations. In acute infec-tions and inflammatory conditions the redistributionof iron from plasma and red blood cell compart-ments may result in an increased synthesis of ferritinby the stimulated reticuloendothelial cells. In addi-tion, ferritin may function as an acute phase reactantthat mirrors disease activity independently of themetabolism of iron.5 '5 The mononuclear phagocytesof pleural fluid may have reticuloendothelial systemactivity within the pleural cavity aside from the poss-ible reticuloendothelial like function of the pleuralmembrane. Of all blood cells, monocytes have thehighest concentration of ferritin'6; macrophageswere the only cells in pleural fluid in which wedetected ferritin immunocytochemically.Human leucocytes contain different ratios of

isoferritins.'' Because our assay measures only thespleen type ferritin, total pleural fluid ferritin con-centrations have probably been underestimated.

In patients with cancer and liver metastases boththe release of liver cell ferritin and the possiblyreduced liver parenchymal cell clearance of ferritinmay be responsible for the increased ferritin con-centrations in body fluids.' Besides local tissue nec-rosis, the production of isoferritins by malignantcells may increase the ferritin concentration inpleural fluid.' Iron released into the pleural cavityeven during minor bleeding may also stimulate thelocal synthesis of ferritin.'8 Phagocytic events closeto the tumour may stimulate the local release offerritin from mononuclear phagocytes, which containand synthesise it. It is even possible that ferritinreleased from liver cells due to liver cell damageultimately reaches the pleural cavity. Fluid passesfrom the peritoneal to the pleural space by means ofthe abundant lymphatics in the diaphragm. Observa-tions of the flow routes of particles and proteins

822



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


9 '~~~~~~~~~~~~~~~~~~~~~~~~~.Afik

0

'L.IP

a,

3_l .. ~~a_*. ___

Fig. 3 Specific immunoperoxidase staining offerritin in pleural fluid macrophages (a) and absence ofstaining in control smears (b).

I

V

have suggested that the diaphragmatic lymphaticsystem plays the most important part in the absorp-tion of peritoneal accumulations of fluid.'9

According to the hypothesis presented by Blake etal, free iron and ferritin are implicated in theinflammation of joints in rheumatoid arthritis.20High concentrations of ferritin and ferric iron havebeen found in the synovial fluid of patients withrheumatoid arthritis.5 The phagocytic lining cells ofthe synovial membrane have many of the features ofa reticuloendothelial organ.2' In rheumatoid arth-ritis many of the biochemical changes noted in thesynovial fluid correspond with those seen in thepleural fluid-for example, decreased concentra-tions of glucose and complement components (C3and C4) and increased concentrations of protein andlactate dehydrogenase activity.9 In rheumaticexudates the increased concentration of proteinssuch as lysozyme, fibronectin, and ferritin-all sub-stances produced by activated mononuclearphagocytes-may reflect saturation and defectivefunction of the reticuloendothelial system inrheumatoid arthritis.22 In fact, the suggestion ofBlake et a120 that in men with rheumatoid arthritisthe greater iron stores are likely to cause severe

complications and maintain the local inflammatoryprocess agrees well with the known fact thatrheumatoid pleural effusion is seen mainly in men.

This study was supported by grants from the SigridJusdius Foundation and Finska LiikaresiillskapetHelsinki, Finland.

References

'Worwood M. Ferritin in human tissues and serum. Clin Haematol1982; 11:275-307.

2 Lipschitz DA, Cook JD, Finch CA. A clinical evaluation ofserum ferritin as an index of iron stores. N Engl J Med1975;290: 1213-6.

3Prieto J, Barry M, Sherlock S. Serum ferritin in patients with ironoverload and with acute and chronic liver diseases. Gastro-enterology 1975;68:525-33.

4Roeser HP. Iron metabolism in inflammation and malignant dis-ease. In: Jacobs A, Worwood M, eds. Iron in biochemistry andmedicine. London: Academic Press, 1980:605-40.

Elin RJ, Wolf SM, Finch CA. Effect of induced fever on serumiron and feruitin concentrations in man. Blood 1977;49:147-53.

6Konijn AM, Hershko C. Ferritin synthesis in inflammation. I.Pathogenesis of impaired iron release. Br J Haematol1977;37:7-16.

Birgegard G, Haligren R, Killander A, Stromberg A, Venge P,Wide L. Serum ferritin during infection. A longitudinal study.Scand J Haematol 1978;21:333-40.

823



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

824

8 Blake DR, Bacon PA, Eastham EJ, Brigham K. Synovial fluidferritin in rheumatoid arthritis. Br Med J 1980;281:715-6.

9 Pettersson T, Klockars M, Hellstrom P-E. Chemical andimmunological features of pleural effusions: comparison be-tween rheumatoid arthritis and other diseases. Thorax1982;37:354-61.

Addison GM, Beamish MR, Hales CN, Hodgkins M, Jacobs A,Llewellin P. An immunoradiometric assay for ferritin in theserum of normal subjects and patients with iron deficiency andiron overload. J Clin Pathol 1972;25:326-9.

Miles LEM, Lipschitz DA, Bieber CP, Cook JD. Measurementof serum ferritin by a 2-site immunoradiometric assay. AnalBiochem 1974;61: 209-24.

2 Klockars M, Reitamo S. Tissue distribution of lysozyme in man.J Histochem Cytochem 1975;23:932-40.

Graham RC Jr, Karnovsky MJ. The early stages of absorption ininjected horseradish peroxidase in the proximal tubules ofmouse kidney: ultrastructural cytochemistry by a new techni-que. J Histochem Cytochem 1966; 14:291-302.

4 Pettersson T. Cellular immunity. In: Chretien J, ed. Biologicalresponses of the pleura in health and disease. New York:Marcel Dekker (in press).

'5 Rothwell RS, Davis P. Relationship between serum ferritin,anaemia, and disease activity in acute and chronic rheumatoidarthritis. Rheumatol Int 1981; 1: 65-7.

Klockars, Weber, Tanner, Hellstrom, Pettersson6 Summers M, Worwood M, Jacobs A. Ferritin in normal erythro-

cytes, lymphocytes, polymorphs and monocytes. Br JHaematol 1974;28: 19-26.

'' Jones BM, Worwood M, Jacobs A. Isoferritins in normalleucocytes. Br J Haematol 1983; 55: 73-81.

8 Fineberg RA, Greenberg DM. Ferritin biosynthesis. II Accelera-tion of synthesis by the administration of iron. J Biol Chem1955; 214:97-106.

9 Courtice FC, Simmonds WJ. Physiological significance of lymphdrainage of the serous cavities and lungs. Physiol Rev1954;34:419-48.

20 Blake DR, Hall ND, Bacon PA, Dieppe PA, Halliwell B, Gut-teridge JMC. The importance of iron in rheumatoid disease.Lancet 1981;ii: 1142-4.

21 Muirden KD. Ferritin in synovial cells in patients withrheumatoid arthritis. Ann Rheum Dis 1966;25:387-401.

22 Williams BD, Pussell BA, Lockwood CM, Cotton C. Defectivereticuloendothelial function in rheumatoid arthritis. Lancet1979;i: 1311-4.

Requests for reprints to: Dr T Pettersson, Fourth Depart-ment of Medicine, Helsinki University Central Hospital,Unionsgatan 38, SF-00170 Helsinki, Finland.



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

pleural fluid ferritin concentrations-in disease · congestive heart failure-- - 102 (50-840) 4...

Documents