DIAGNOSTIC VALUE OF ASCITES ADENOSINE DEAMINASE IN TUBERCULOUS PERITONITIS

751

DIAGNOSTIC VALUE OF ASCITES

ADENOSINE DEAMINASE IN TUBERCULOUS PERITONITIS ISAAC KALVARIA2 PETER BERMAN4 RALPH E. KIRSCH1

MICHAEL D. VOIGT1 CHARLES TREY3 CARL LOMBARD5

UCT/MRC Liver Research Centre, Department of Medicine and Gastroenterology Clinic, University of Cape Town and Groote Schuur Hospital, South Africa;1 Department of Medicine, Division

of Gastroenterology, Hepatology, and Nutrition, University of Florida, Gainesville, USA;2 Division of Gastroenterology, Deaconess Hospital, Boston, Massachusetts, USA;3 Department of Chemical Pathology, University of Cape Town and Groote Schuur

Hospital;4 and MRC Institute for Biostatistics, Tygerberg, South Africa5

The value of ascitic fluid adenosine deaminase activity in distinguishing tuberculosis from other causes of ascites was examined in a retrospective study of 41 patients with bacteriologically confirmed tuberculous peritonitis and 41 control patients, matched for age and sex, with ascites of other causes (12 alcoholic cirrhosis, 5 cryptogenic cirrhosis, 12 malignant disorders, 3 pancreatitis, and 9 miscellaneous causes). The mean ascites adenosine deaminase activity was 99·8 (SD 49·1) in tuberculous patients and 14·8 (8·4) U/l in control patients (p <0·0001). A cutoff of 32·3 U/l had a sensitivity of 95% and specificity of 98% in distinguishing between the two groups. In a subsequent prospective study of 64 patients with ascites, 11 were found to have tuberculosis. Of the others, 23 had cirrhosis (18 alcoholic, 5 cryptogenic), 17 malignant disorders, 3 pancreatitis, 5 cor pulmonale, 3 congestive cardiac failure, 1 systemic mastocytosis, and 1 renal failure and hypothyroidism. The mean ascites adenosine deaminase activity was 112·6 (45·0) U/l in the patients with tuberculous ascites and 16·3 (36·7) U/l (p<0·0001) in those with ascites of other causes. In this study, adenosine deaminase had a sensitivity of 100% and specificity of 96% in discriminating tuberculosis from other causes of ascites. These findings suggest that the ascitic fluid adenosine deaminase activity may be used to identify patients in whom the diagnosis of abdominal tuberculosis must be pursued.

Summary

peritonitis.16 We therefore compared adenosine deaminase activity in 41 patients with proven tuberculous ascites and 41 patients with ascites of other causes and then determined its sensitivity and specificity in the diagnosis of tuberculous ascites in a prospective study of 64 patients with ascites. Patients and Methods In the pilot study the clinical and bacteriological records of all patients with ascites admitted to Groote Schuur Hospital between January, 1982, and January, 1986, were reviewed. All 41 patients with bacteriologically proven tuberculous ascites in whom ascitic fluid adenosine deaminase activity had been determined were included. As controls, 41 patients, matched for age and sex, with

ascites of other causes and whose ascites adenosine deaminase activity had been determined were selected by M. D. V., who was unaware of their adenosine deaminase results. There were 21 female and 20 male patients in each group. The median age of the patients with tuberculous ascites (43, range 16-75 years), was similar to that (48, range 12-80 years) of the controls. 14 of the patients with tuberculosis and 11of the control patients admitted alcohol abuse. Of the 41 patients with tuberculosis, 36 had positive ascitic fluid bacteriology (9 on direct microscopy and 27 on culture), 4 had peritoneal granulomas in which acid-fast bacilli were detected, and 1 patient had caseating peritoneal granulomas plus acid-fast bacilli in his sputum. The control group consisted of 12 patients with alcoholic cirrhosis, 5 with cryptogenic cirrhosis, 12 with ascites due to malignant disorders, 3 with pancreatic ascites, and 9 with miscellaneous causes of ascites-2 cardiac failure, 1 graft-versushost disease, 1 sarcoidosis, 1 Budd-Chiari syndrome, 1 systemic lupus erythematosus, 2 cor pulmonale, and in 1 patient the cause of the ascites was not found. The prospective study included 64 patients with ascites admitted consecutively to the medical wards of Groote Schuur Hospital between February, 1985, and August, 1987. 37 were male and 27 female. The median age was 60 (range 13-85) years. Since the ascitic fluid adenosine deaminase activity was not known and was not considered to be part of the investigative routine in patients with ascites, all participating physicians agreed that the results of this test should be unavailable during the study. The diagnosis of tuberculous ascites was made in 11 patients: 5 had positive ascitic fluid culture; 5 positive peritoneal histology; and 1 strongly suggestive clinical features and ascites which resolved on antituberculous therapy alone. 5 of the patients with tuberculosis

Introduction TUBERCULOUS peritonitis is an important cause of ascites

in developing countries.1-3 Since both pulmonary and

extrapulmonary tuberculosis

are

becoming

more common

in Europe and North America, tuberculosis should be considered as a cause of ascites in these areas also. 4-8 Clinical, biochemical, and bacteriological tests used up to now for the diagnosis of tuberculous ascites are insensitive, non-specific, or take a long time.9,10 Laparoscopy and laparotomy are hazardous in these ill patientsll,12 and are reserved for those in whom the diagnosis of abdominal tuberculosis is strongly suspected. Since tuberculous peritonitis occurs in many patients with other disorders that can cause ascites,2,13 this diagnostic suspicion is often lacking. The activity of adenosine deaminase, which catalyses the conversion of adenosine to inosine and is released by lymphocytes and macrophages during the cellular immune response, is useful in diagnosis of tuberculous meningitisl4 and pleural effusions,15 but it has not been widely studied in tuberculous

Fig 1-Retrospective study activity.

of ascites adenosine deaminase

(ADA)

Middle horizontal line of box plot median; box middle 50% of values; vertical solid line = 1 x (fourth spread); open circles outlying values. =

=

=

752

Fig 2-Ascites lactate dehydrogenase activity, protein concentration, lyniphocyte count, and glucose concentration in prospective study. admitted excessive alcohol consumption. The non-tuberculous group consisted of 23 patients with cirrhosis (18 alcoholic, 5 cryptogenic), 17 with ascites due to malignant disorders, 3 with pancreatitis, 5 with ascites due to cor pulmonale, 3 with congestive cardiac failure, 1 with systemic mastocytosis, and 1 with renal failure and hypothyroidism. In the prospective study each patient underwent diagnostic paracentesis on admission before any therapy was given. Ascitic fluid protein and glucose concentrations, white cell count, differential count, lactate dehydrogenase activity, and adenosine deaminase activity were determined in every case, and fluid was sent for cytological and microbiological examination, including culture. Adenosine deaminase activity was determined spectro-

photometrically. 17 Non-parametric

tests (Wilcoxon signed rank and two sample carried out on both data sets to determine differences between patients with tuberculous ascites and the control subjects. Univariate logistic regression was used as a discriminant analysis, to determine the optimum cutoff, in the pilot study. This cutoff was

test)

were

used to classify the patients of the prospective study and to validate the sensitivity and specificity. Sensitivities and specificities, other than for adenosine deaminase, were calculated from the prospective data. Because of the small sample size, these should be considered

exploratory. Results Pilot

Study

Adenosine deaminase activity was significantly higher in tuberculous peritonitis patients than in controls (998 [SD 491] vs 148 [8’4] U/13- p<0-0001). A cutoff of 32.3 U/l distinguished the groups with a sensitivity of 95% and a specificity of 98% (fig 1).

Prospective Study There were significant differences between the tuberculosis patients and the controls in ascites lactate

753 DIAGNOSTIC TESTS FOR TUBERCULOUS PERITONITIS

*100-sensitivity (%).

dehydrogenase activity (p < 0001), protein concentration and glucose (p<00001), lymphocyte count (p<0001), concentration (p < 0’03) (fig 2). The sensitivity, specificity, false-positive, and false-negative rates for each of these measurements are shown in the table. Although all the tuberculous patients had low ascites glucose concentrations, so did 23% of the controls, including patients with uncomplicated cirrhosis and ascites due to malignant disorders.

diagnosis of tuberculous ascites largely depended on ascitic fluid culture (45 % of patients), with a delay of up to 6 weeks, or on invasive surgical procedures (45 % of patients). Adenosine deaminase activity has been useful in patients with pleural effusions15 or meningitis,14 in whom tuberculosis must be considered present until it is disproved. In a prospective study, Martinez-Vazquez and colleagues found a clear difference in adenosine deaminase activity between 10 patients with tuberculous peritonitis and 56 other patients with ascites (23 had cirrhosis, 17 ascites due to malignant disorders, 8 cirrhosis with bacterial peritonitis, 2 ascites due to hypoproteinaemia, and 6 miscellaneous causes of ascites) .16 Our findings confirm those of MartinezVazquez et al .16 A potential weakness of our study is that so few patients had non-tuberculous infective ascites. However, Martinez-Vazquez et al found normal adenosine deaminase activities in their 8 patients with bacterial peritonitis.16 The high sensitivity and very small overlap in both studies suggest that adenosine deaminase activity is better than other tests used in diagnosing tuberculous peritonitis. We confirmed the strong association of alcoholism with tuberculosis in this study; 43% of the tuberculous patients admitted alcohol abuse. Martinez-Vazquez et al16 found that adenosine deaminase activity was diagnostically accurate in this group, but our only false-negative results occurred in 2 alcoholic tuberculous patients. Both had cirrhosis and had had ascites for more than a year before tuberculosis was diagnosed. Thus, although the accuracy of adenosine deaminase activity is probably good in patients with underlying disease, some false-negative results occur. 3 of 29 patients with malignant ascites had false-positive adenosine deaminase results. 2 of these had positive cytology and in the third the diagnosis was made at endoscopy. 1 patient with surgical peritonitis also had false-positive results. He presented with acute abdominal pain. A clinical diagnosis of pancreatitis was confirmed by finding an ascites amylase activity of 385 000 U /1 and Candida albicans was subsequently cultured from his ascitic fluid. We found a greater range of adenosine deaminase activity

Fig 3--Ascites adenosine deaminase activity in prospective study. Adenosine deaminase activity was significantly higher in tuberculous than in control ascites (1126 [45.0] vs 16 3 [36’7] V/l; p < 00001; fig 3). The sensitivity was 100% and the specificity 96% giving a false-positive rate of 4% and no false-negative results. 2 of the 3 patients with false-positive adenosine deaminase activity had malignant disorders and the third had pancreatic ascites with fungal infection. The diagnosis was made by ascitic fluid cytology in 1, by means of ascites amylase activity (385 000 U/1) in 1, and at endoscopy in 1 (ascites due to peritoneal spread from stomach

We thank Dr Hal Conn for

helpful advice.

Correspondence should be addressed to R. E. K., UCTMRC Liver Research Centre, University of Cape Town Medical School, Observatory 7925, South Africa.

carcinoma).

REFERENCES

Discussion The

in tuberculous peritonitis than Martinez-Vazquez et al,16 possibly because we had a larger study group or a higher prevalence of underlying disease in our patients. The optimum cutoff point in our study is lower than that proposed by Martinez-Vazquez.16 Our study suggests that measurement of ascitic fluid adenosine deaminase activity is sufficiently useful in identifying those patients in whom the diagnosis of tuberculosis peritonitis should be actively pursued to justify the routine use of this measurement, at least in areas where tuberculosis is endemic.

early diagnosis of tuberculous ascites is important; the despite availability of effective treatment, the coexistence of malnutrition and compromised immunity often have an adverse effect on the outcome. By means of standard chemical and haematological analysis of ascitic fluid, the false-negative diagnostic rate varied from 0% to 36% and the false-positive rate from 8% to 27%. In our patients the

1 Collie A 2. Menzies

Extrapulmonary tuberculosis Epidemiol Comments 1987; 14: 1-21. RI, Alsen H, Fitzgerald M, Mohapeloa RG. Tuberculous peritonitis in Lesotho. Tubercle 1986, 67: 47-54. 3. Nwokolo C. Ascites in Africa. Br Med J 1967; i: 33-37 4. Leads from the MMWR Tuberculoshias-United States, 1985 and possible impact of human T-lymphotrophic virus type III lymphadenopathy-associated virus infection. JAMA 1986; 255: 1012-13. 5 Pitchenik AE, Cole C, Russel VW, Fischl MA, Spira TJ, Snider DE. Tuberculosis, atypical mycobacteriosis and the acquired immunodeficiency syndrome among Haitian and non-Haitian patients in South Florida Ann Intern Med 1984, 101: 641-45

754

Preliminary Communication INDUCTION OF GRAFT-VERSUS-HOST DISEASE AFTER AUTOLOGOUS BONE MARROW TRANSPLANTATION RICHARD J. JONES ALLAN D. HESS RISA B. MANN STEVEN PIANTADOSI

GEORGIA B. VOGELSANG EVAN R. FARMER ROBERT B. GELLER

GEORGE W. SANTOS

Johns Hopkins Oncology Center, The Johns Hopkins Medical Institutions, Baltimore, Maryland 21205, USA induce graft-versus-host disease (GvHD) in patients undergoing autologous bone marrow transplantation (BMT), five consecutive patients with non-Hodgkin lymphoma or Hodgkin’s disease in resistant relapse were treated with cyclophosphamide and total body irradiation or busulphan and cyclophosphamide, followed by autologous BMT. The patients received cyclosporin 1 mg/kg per day intravenously for 28 days after BMT. Histologically proven grade II acute GvHD of the skin developed in all the patients at a median of 11 days (range 9-13) after BMT. One patient died and the GvHD resolved in 1-3 weeks in the four remaining patients— spontaneously in two and with corticosteroid treatment in two. Lymphocytes from one patient obtained at the time of GvHD were cytotoxic for the patient’s own pretransplant lymphocytes. Cytotoxicity was blocked by antibodies directed against class II histocompatibility (HLA-DR or Ia) antigens. Cyclosporin-induced GvHD after autologous BMT resembles mild GvHD after allogeneic grafting. This syndrome appears to be mediated by autoreactive Ia-

Summary

To

specific lymphocytes. INTRODUCTION

GRAFT-VERSUS-HOST disease (GvHD) develops in about undergoing allogeneic bone-marrow transplantation (BMT), and is the main cause of morbidity and mortality in these patients. However, GvHD after allogeneic BMT also seems to produce a clinically significant anti-tumour effect.12 Although the toxicity of GvHD is avoided with autologous BMT, the graft-versustumour effect is lost, which seems to be a major reason for higher relapse rates after autologous and syngeneic (identical twin) BMT than after allogeneic BMT.1,2 A syndrome similar to GvHD has been reported to occur spontaneously in 8% of patients receiving autologous or syngeneic bone marrow transplants.3 GvHD after autologous or syngeneic BMT clinically and histologically

50-70% of patients

6. Sunderam

G, McDonald RJ, Maniatis T, Oleske J, Kapila R, Reichman LB. Tuberculosis as a manifestation of the acquired immunodeficiency syndrome. JAMA 1986; 256: 362-66. 7. Mackay AD, Cole RB. The problems of tuberculosis in the elderly. Q J Med 1984; 53: 497-510. 8. Stead MW, Lofgren JP, Warren E, Thomas C. Tuberculosis as an endemic and nosocomial infection among the elderly in nursing homes N Engl J Med 1985; 312: 1483-87. 9. Singh MM, Bhargava AN, Jain KP. Tuberculous peritonitis. An evaluation of pathogenenc mechanisms, diagnostic procedures and therapeutic measures. N Engl

J Med 1969; 281: 1091-94. Gilinsky NH, Marks IN, Kottler RE, Pnce SK. Abdominal tuberculosis. A review. S Afr Med J 1983; 64: 849-57. 11. Jorge AD. Pentoneal tuberculosis. Endoscopy 1984; 16: 10-12 10.

10 year

resembles GvHD after allogeneic BMT, but it is a mild, self-limited disease that generally involves only the skin. Whether GvHD could develop after autologous BMT was controversial until the discovery of animal models for the disease.4,5 In these models, GvHD was induced in rats or mice undergoing syngeneic BMT by treatment with cyclosporin. GvHD after syngeneic BMT in rats seems to be mediated by autoreactive lymphocytes directed against class II histocompatibility (HLA-DR or Ia) antigens.5 Because GvHD after autologous BMT is a rare phenomenon, it has not been possible to establish whether it is associated with an anti-tumour effect. It is possible that GvHD after autologous BMT will have beneficial antitumour activity without increased post-transplant toxicity. Our aim was to establish whether GvHD could be induced in patients undergoing autologous BMT. We treated patients, who had non-Hodgkin lymphoma or Hodgkin’s disease and were not responding to conventional salvage chemotherapy (a response termed resistant relapse), with autologous BMT and cyclosporin. The rationale for studying this group of patients was that lymphomas in resistant relapse have an extremely poor disease-free survival following BMT.6 Since non-Hodgkin lymphoma cells and Reed-Sternberg cells of Hodgkin’s disease usually bear the Ia antigen on their surfacesthey could be targets for GvHD that develops after autologous BMT. PATIENTS AND METHODS

Patients

investigated five consecutive patients aged 30-48 years (median 37) who were treated at the Johns Hopkins Hospital Oncology Center between August, 1987, and July, 1988. Table I shows the patient details. All five patients had failed at least two combination chemotherapy regimens and their disease had progressed throughout their last chemotherapy. Both patients with Hodgkin’s disease and one patient with non-Hodgkin lymphoma also failed radiation therapy. The protocol was approved by the joint committee on clinical investigation of The Johns Hopkins Hospital, and informed consent was obtained from all patients before entry into the study. We

Treatment About 3 x 101 nucleated bone marrow cells per kg were harvested from the posterior iliac crest. The marrow buffy coat cells were cryopreserved, unpurged, with 10% dimethylsulfoxide (DMSO) in liquid nitrogen. The BMT preparative regimen was cyclophosphamide, 50 mg/kg per day intravenously for 4 days, followed by total body irradiation, 3 Gy per day for 4 days (Cy-TBI) in two patients. For the three patients who had had irradiation, the preparative regimen was busulphan, 4 mg/kg per day orally for 4 days, followed by cyclophosphamide as before (Bu-Cy). The autologous bone marrow grafts were infused 1 day after Cy-TBI and 2 days after Bu-Cy. Cyclosporin was started on the day of marrow infusion and was continued for twenty-eight days after

12. Levine H. Needle biopsy diagnosis of tuberculous peritonitis. Am Rev Resp Dis 1968; 98: 519. 13. Karney WW, O’Donoghue JM, Ostrow JH, Holmes KK, Beaty HN. The spectrum of

tuberculous peritonitis. Chest 1977; 72: 310-15. PR, Malan C, Van Der Walt A, Schoeman JF. The simultaneous determination of cerebrospinal fluid and plasma adenosine deaminase activity as a diagnostic aid in tuberculous meningitis. S Afr Med J 1986; 69: 505-07. 15. Maritz FJ, Malan C, Le Roux I. Adenosine deaminase estimations in the differentiation of pleural effusions. S Afr Med J 1982, 62: 556-58. 16. Martinez-Vazquez JM, Ocana I, Ribera E, Segura RM, Pascual C Adenosine deaminase activity in the diagnosis of tuberculous peritonitis. Gut 1986, 27: 1049-53. 17. Giusti G. Adenosine deaminase. In: Bergmeyer HU, ed. Methods of enzymatic analysis, vol 2, 2nd ed. New York: Academic Press, 1974: 1092-99. 14. Donald