Corticosteroids in the Treatment of Tuberculous Pleurisy*

Transcription

1 Corticosteroids in the Treatment of Tuberculous Pleurisy* A Double-blind, Placebo-controlled, Randomized Study Christoph Wyser, MD/ Gerhard Walzl, MD; Jan P. Smedema, MD; Fran9ois Swart, RN; Emmerentia M. van Schalkwyk, MD; and Bernard W. van de Wal, MD Although several studies on tuberculous (TB) pleurisy suggest that the addition of corticosteroids to anti-tb therapy may have beneficial effects, these agents are not used routinely. To assess the effects of short-term oral prednisone therapy in TB pleurisy, 74 patients were randomly assigned in a double-blind fashion to treatment with either placebo or prednisone at a dose of mglkgld for up to 4 weeks with gradual reduction over an additional 2 weeks. All subjects received a standard 3-drug anti-tb chemotherapy regimen for 6 months. TB pleurisy was diagnosed by histologic study and/or culture of pleural biopsy specimens obtained at thoracoscopy. Complete drainage of the effusion was performed simultaneously. Outcome measures were assessed periodically for 24 weeks, including indexes of morbidity and pleural thickening. After randomization, four patients were excluded from the final analysis. Of the 70 patients analyzed, 34 received prednisone and 36 received placebo. Demographic and clinical characteristics of the treatment groups were comparable at the time of hospital admission. Although a statistically significant improvement in symptoms occurred earlier in the prednisone group (8 weeks) than in the placebo group (12 weeks), betweengroup comparison showed no significant differences at any of the follow-up evaluations. The proportion of subjects in the prednisone group (53.1%) with residual pleural thickening at 6 months did not differ significantly from that of the placebo group ( 60%). Pleural effusions did not recur in any of the patients. Initial complete drainage of the effusion was associated with greater symptomatic improvement than any subsequent therapy. We conclude that standard anti-tb therapy and early complete drainage is adequate for the treatment of TB pleurisy. The addition of short-term oral prednisone therapy neither results in clinically relevant earlier symptom relief nor confers a beneficial effect on residual pleural thickening. (CHEST 1996; 110:333-38) Key words: computed tomography; lung function testing; prednisone; therapeutic aspiration; thoracoscopy; tuberculous pleurisy Abbreviations: CXR=chest radiograph; HRCT =high-resolution CT; PFT =pulmonary function testing; TB=tuberculosis or tuberculous; TLC=totallung capacity; VAS=visual analog scale Several studies of tuberculous (TB) pleurisy have suggested that adjuvant corticosteroid therapy may confer beneficial effects. These include a reduced morbidity, a decrease in residual pleural thickening, and more rapid reabsorption of pleural fluid How- *From the Department of Internal Medicine and Stellenbosch University Pulmonary Unit for Research (SUPUR), University of Stellenbosch, Cafle Town, South Africa. 1 Currently at the Department of Internal Medicine, Division of Pulmonoiogy, Kantonspital, Basel, Switzerland. Supported by the Tuberculosis 1992 fund, University of Stellenboscn, Capetown, South Africa. Manuscript received October 9, 1995; revision accepted March 28, Reprint requests: Dr. C. Wyser, Department of Internal Medicine, Division oj Pulrrwnology, Kantonspital, Petersgraben 4, 4031 Basel Switzerland ever, routine corticosteroid therapy in the treatment of TB pleurisy is currently not widely used. One recent study by Lee and coworkers 10 suggested that the addition of oral prednisolone to anti-tb chemotherapy in 40 patients resulted in a more rapid improvement in clinical symptoms and resorption of fluid. The same study could not demonstrate that the occurrence of pleural thickening was decreased by the administration of corticosteroids. However, the number of patients studied was relatively small and only a small amount of pleural fluid (50 ml) was aspirated initially. To our knowledge, no other placebo-controlled studies that examine this form of treatment have been published. Furthermore, in the current literature, no consensus has been reached about the role CHEST /110 I 21 AUGUST,

2 of therapeutic aspiration of the pleural fluid as part of the treatment of TB pleurisy. To examine this further, we conducted a doubleblind, placebo-controlled, randomized study to investigate the beneficial effects of prednisone to reduce clinical symptoms and the occurrence of pleural thickening after complete aspiration of the pleural fluid. Patient Selection MATERIALS AND METHODS Between April 1994 and January 1995, patients with exudative pleural effusions were recruited from the medical admission ward oftygerberg Hospital, a university teaching hospital in Cape Town, South Africa. Thoracoscopy Thoracoscopy followed by bronchoscopy was performed under general anesthesia using stan d ar d me th o d s. u12 B. 10psy spec1mens of the parietal pleura were examined histologically and stained for acid-fast bacilli. Pleural fluid specimens and biopsy specimens suspended in saline solution were cultured for Mycobacterium tuberculosis. Following thoracoscopy, an intercostal drain was inserted to evacuate any remaining air and fluid, and left in situ for up to 48 h. The macroscopic thoracoscopic findings were graded: type 1, nonspecific inflammation of the parietal pleura with no or only a few fibrinous adhesions; type 2, "classic" TB pleurisy 13 with an inflamed reddish parietal pleura and multiple grayish-white nodules; and type 3, fibrous inflammation with a thickened parietal pleura and multiple fibrous adhesions and/or loculations. At bronchoscopy, bronchial lavage fluid was obtained for M tuben.:ulosis culture. Patients with biopsy specimen proven TB pleurisy were eligible for entry into the study. The diagnosis oftb was confirmed by the presence of caseating granulomas with or without acid-fast bacilli on histologic study and/or a positive M tuberculosis culture. Other causes of pleural exudates such as pneumonia or malignancy were excluded. Patients with contraindications to corticosteroid use, such as diabetes mellitus, uncontrolled hypertension, peptic ulcer disease, and empyema were also excluded, as were HIV-seropositive patients and those with neoplastic disease. Written consent was obtained from all subjects and the study was approved by the U m versity of Stellenbosch and Tygerberg Hospital Ethics Committee. Treatment Protocol All eligible patients were randomly assigned in a double-blind fashion to treatment with either prednisone plus standard anti-tb therapy (prednisone group) or placebo plus standard anti-tb therapy (placebo group). All patients were prescribed anti-tb treatment according to the South African National Guidelines, consisting of rifampicin, 10 mglkgld, isoniazid, 8 mglkgld, and pyrazinamide, 25 mglkgld as a fixed combination tablet (Rifater), and pyridoxine, 25 mglkgld for 6 months. In addition, prednisone, 0.75 mglkgld, or identical placebo tablets were administered. After 2 to 4 weeks, depending on the therapeutic response as assessed by a progressive reduction of symptoms and radiologic improvement, this was gradually reduced over a 2-week period by 5 mgld in all patients. Follow-up Visits Patients were evaluated on hospital admission, 2 to 3 days after thoracoscopy (baseline of the study), and at 2, 4, 8, 12, 16, and 24 weeks. The follow-up evaluation included the following: history, physical examination, chest radiograph (CXR) as needed, high-resolution CT (HRCT) of the chest at 24 weeks and pulmonary function tests (PFTs) at baseline, 12 weeks, and 24 weeks. 334 Morbidity On hospital admission, patients were interviewed and information regarding illness duration, weight loss, pleuritic chest pain, and fever was recorded. Dyspnea, cough, night sweats, tiredness, appetite, pleuritic chest pain, and general well-being were each graded from 0 to 100 using a visual analog scale (VAS). A combined index with a maximum possible score of 700 was calculated. 14 The degree of improvement in VAS scores was estimated by subtracting follow-up from baseline scores and expressed as a percentage of baseline values. New symptoms and clinical signs developing during the treatment period were documented as possible side effects. Blood glucose levels, body temperature, weight, and BP were recorded at each visit. CXRIHRCT of the Chest CXRs were performed in the posteroanterior, lateral, and lateral decubitus projection and obtained on hospital admission, after thoracoscopy, and during the outpatient follow-up visits as long as signs of pleural opacification persisted, but in all patients at baseline, 2 weeks, and 24 weeks. Two experienced readers, blinded to the clinical history, assessed the initial size and the recurrence of pleural effusion, the pleural thickness (in millimeters), and the appearance of the costophrenic angle (sharp <90, blunted ~ g oas o ) well as any pulmonary parenchymal involvement (inflltrate, cavity, scarring). A more than 2-mm pleural thickness (at the point of maximal thickness) at 24 weeks was defined as residual pleural thickening. The amount of pleural fluid at the initial presentation was regarded as small (less than one third of one hemithorax), moderate (between one third and half of one hemithorax), or large (more than half of one hemithorax). HRCT of the chest was obtained at three different levels, namely at the aortic arch, the midhilum, and just above the right hemidiaphragm. Maximal pleural thickness was measured adjacent to a rib. 15 Pulmonary Function Tests Spirometry and body plethysmography were performed (using the Master-lab; Jaeger; Wuerzburg, Germany). Pulmonary function measurements were performed according to standard protocols and conformed to American Thoracic Society guidelines 16 The predicted normal values were those proposed by Schoenberg and colleagues17 for spirometry, and by Goldman and Becklake 18 for lung volumes. On completion of treatment, the presence of restriction was defined as decreased total lung capacity (TLC) and FVC of less than 80% of normal predicted. The degree of improvement for each patient was assessed by subtracting baseline from follow-up results and expressed as a percentage of the baseline values. Statistical Analysis Non parametric statistical tests were applied because the variables were not normally distributed and the VAS can be considered an ordinal measurement. The significance of diflerences between baseline and follow-up data within each group was assessed using the Wilcoxon signed rank test (numeric or ordinal data) and the Stuart Maxwell test (ordinal data). For between-group comparisons, we used the Mann-\Vhitney U test (numeric or ordinal data) and a x 2 test with 1 df (dichotomous nominal data). To control for type I error, the p values obtained were multiplied by the number of pairwise comparisons performed (Bonferroni). Significance was defined as p<0.05. Baseline Observations RESULTS A diagnosis of TB pleurisy was made in 7 4 patients. Four patients were excluded from the efficacy analy- Clinical Investigations

3 sis: three because of noncompliance with the treatment, and one because of the detection of an esophageal carcinoma at follow-up. Baseline demographic, clinical, and pathologic features of the remaining 70 patients are summarized in Table 1. Thirty-four patients were randomized to the prednisone group and 36 to the placebo group. There were no significant differences between the treatment groups with regard to sex, age, duration of illness, size of the effusion, histologic features and microbiology results, or the appearance of the pleura at thoracoscopy. Fewer patients in the prednisone group (21.2%) had parenchymal involvement than the placebo group (44.4%; p=0.06). Thoracoscopy At thoracoscopy, most patients demonstrated the classic macroscopic picture (type 2) oftb pleurisy. The distribution of the three different macroscopic morphologic patterns was similar in the two treatment groups after randomization (Table 1). Reexpansion pulmonary edema, confirmed radiologically, occurred after 7 of 68 thoracoscopies (10.3%; 6 patients with large effusions and 1 patient with a hydropneumothorax), but resolved within hours of the initiation of oxygen therapy. Histology and Microbiology In the combined group of 70 patients, caseating granulomas occurred in 64 (92.8%) patients and noncaseating granulomas occurred in 5 (7.2%) patients. Four of the latter five patients had positive cultures for M tuberculosis and in one patient, acid-fast bacilli were seen histologically. Only one patient did not have a pleural biopsy, and the diagnosis of TB pleurisy was based on bilateral pleural lymphocytic exudates and a positive sputum culture form tuberculosis. Positive M tuberculosis cultures of pleural biopsy specimens were obtained in 54 (78.3%) of 69 patients, and Ziehl Neelsen stains of biopsy specimens were positive in 34 patients (49.3%). In an additional two patients, cultures were positive in only the sputum and bronchial lavage fluid, respectively. Sixty-one (87.1 %) patients had TB pleurisy diagnosed by either acid-fast bacilli on histologic study or positive M tuberculosis cultures of pleural tissue. Morbidity A dramatic improvement in the patients' subjective assessment of their general condition was observed immediately after drainage of the pleural fluid as demonstrated by a marked decrease in the median combined VAS index score in both groups (Table 2). Although most patients (77.1 %) had an elevated temperature greater than 37.5 C (mean and SD: 38.3± Table!-Demographic and Clinical Features of Treatment Groups at Baseline Placebo Prednisone p Value No. of patients * Male,% Race Mixed race, % Black,% Age, yr, mean±sd 32.8± ±13 Duration of illness, wk, prior 3.7± ±2.7 to hospital admission, mean±sd Pleuritis only, % Pleuritis and pulmonary TB, % Initial size of pleural effusion (CXR), % Small Moderate Large TB culture positive, % Pleural fluid Pleural biopsy specimen Bronchial lavage Histology, % Caseating granulomata Noncaseating granulomata Ziehl-Neelsen positive Appearance on thoracoscopy, % Type Type Type *=not significant C), only 1 patient had an elevated temperature 3 days after drainage of the pleural fluid and before the initiation of treatment with the trial medication. Improvement in the individual VAS and combined VAS index scores became statistically significantly different from baseline for the first time at 8 weeks in the prednisone group and at 12 weeks in the placebo group. However, differences in the degree ofimprovement in VAS scores between the 2 groups were not statistically significant at any of the follow-up evaluations (Table 2). Furthermore, the most dramatic Table 2-Morbidity-VAS* Placebo Prednisone p Value f No. of patients Admission 434 ( ) 449 ( ) Baseline 57 (14-150) 1 58 (18-133) 1 2wk 55 (0-115) 51 (11-85) 4wk 40 (9-97) 11 (0-105) Swk 20 (0-65) 3 (0-43)1 12 wk 2 (0-65)1 0 (0-15)1 16 wk 0 (0-12)1 0 (0-20)1 24wk 0 (0-0)1 0(0-0)I *Visual analog scale (median combined index score of well-being, appetite, night sweats, pleuritic chest pain, tiredness, dyspnea, and cough) at different evaluation time points. =not significant; f =comparison between placebo and prednisone group. 1 p<0.000l compared to admission. lp<0.05 compared to baseline. CHEST I 110 I 2 I AUGUST,

4 Table 3-Pleural Thickening Placebo Prednisone p Value No. of patients Residual pleural thickening (%) (CXR) Residual pleural thickening (%) (CT scan) Blunted costophrenic angle(%) (CXR) Pleural thickness, mm, mean:t:sd (CXR) At baseline At 2 wk At 4 wk At 8 wk At 24 wk Pleural thickness at 24 wk, mm (mean:!:sd) on CT scan (50) 21 (60) 10 (27.8) 10.4::!:: ::!:: ::!::7.0 (33)* 8.4::!::7.6 (19)* 2.5::!:: ::!:: (50) 17 (53.1) 14 (41.1) 9.4::!:: ::!:: :!:4.3 (23)* 5.1 ::!::4.3 (17)* 2.1::!:: ::!::3.7 *Number of patients in whom CXRs were taken. improvements in the VAS scores were observed after complete drainage of the effusion, before randomization. Body weight at 4, 8, and 24 weeks after initiation of treatment showed a significant but comparable increase in the 2 groups (mean weight gain at 24 weeks: 5.3 kg in the placebo group and 5.8 kg in the prednisone group). At 6 months, all patients were cured according to clinical and radiologic data and treatment was subsequently stopped. Drug resistance necessitating a change in the anti-tb treatment was not observed in any patient. Radiography and HRCT of the Chest At baseline, only one patient had bilateral pleural effusions. In 59 cases ( 84.3%), more than one half the volume of the hemithorax was involved, and in 24 (34.3%) cases coexisting pulmonary lesions could be seen. The pulmonary involvement included patchy infiltrates with or without small cavitation or scarring. One patient had a hydropneumothorax on hospital admission. CXR at 24 weeks revealed residual pleural thickening on the lateral chest wall in 17 (50%) of the 34 prednisone-treated patients and 18 (50%) of the 36 patients treated with placebo (Table 3). HRCT of the chest showed residual pleural thickening in 17 ( 53.1%) of 32 patients in the prednisone group and 21 (60%) of 35 patients in the placebo group. Three patients did not have a HRCT at 24 weeks. There was no statistically significant difference in the degree of residual pleural thickness on CXR and HRCT (p=0.52). In patients with thoracoscopic findings of acute inflammation (type 1 and 2), residual pleural thickening occurred in 10 of21 patients ( 42.8%) in the prednisone group and in 14 of24 placebo-treated patients (58.8%). Seven of 10 patients (70%) in the prednisone group and 6 of 10 patients (60%) in the placebo group with thoracoscopic evidence of fibrous inflammation (type 336 3) showed residual pleural thickening. These differences were not statistically significant. After complete aspiration, no patient experienced a recurrence of pleural fluid in either treatment group, as assessed by CXR (posteroanterior, lateral, and lateral decubitus projections). The intrapulmonary lesions, when present, appeared to resolve more slowly than the pleural abnormalities. Lung Function Testing At baseline, both groups had mild to moderate lung function impairment (Fig 1); mean±sd TLC was 76.7±12% of predicted normal values in the prednisone group and 77.8±22.9% of predicted normal values in the placebo group; FVC was 59.1± 16.7% and 57.9±12.8%, respectively. A significant improvement of TLC in the respective groups was noted after 3 months (p<o.oool prednisone group; p=0.009 placebo group) and 6 months (p< for both groups). A similar response was observed in FVC. The groups did not differ significantly with respect to the degree of improvement in PFT results (p=0.39 for TLC and p=0.65 for FVC). At 24 weeks, the difference in the proportions of patients with restrictive PFT results (33.3% prednisone group; 39.4% placebo group) was not significant (p=0.72). Side Effects No serious side effects were observed except for epigastric pain (four patients in the prednisone group and three patients in the placebo group). None of the patients had evidence of impaired glucose tolerance. DISCUSSION Our study shows that oral prednisone does not have a beneficial effect on residual pleural thickening in the treatment oftb pleurisy. This is in accordance with the study by Lee and coworkers, 10 who had conducted a similar study, with the exception of the complete Clinical Investigations

5 ~ " ~ 100 iii ~ 90 c:.., 0 ~ 80 ~ ~ 70 ~ 60 g... j100 ~ ~ 90 ~ c: ~ 80 '6! 70 a. 0 ~ 60 0 it 50 BASELINE CHANGE IN TLC WITH TREATMENT J j WEEKS CHANGE IN FVC WITH TREATMENT PREDNISONE --PLACEBO 24 WEEKS L-----J L PLACEBO BASELINE 12 WEEKS 24 WEEKS FIGURE 1. Changes in TLC and FVC with treatment in 70 patients with TB pleurisy. Results are expressed as mean::'::sd. drainage of pleural fluid on hospital admission. In that study, the addition of corticosteroids resulted in a more rapid improvement in clinical symptoms. We also observed a statistically significant improvement in the VAS scores compared with baseline occurring earlier in the prednisone group. However, there was no significant difference between the two groups at any of the follow-up evaluations. We therefore believe that the statistically earlier symptom relief in our prednisone group was clinically irrelevant. The main factor responsible for symptomatic improvement in all patients was the complete early drainage of the effusion at the time of the thoracoscopy. Although a "placebo" effect attributable to the diagnostic procedure per se could contribute to the symptomatic improvement, we believe that the evacuation of the effusion was the principal reason. This is supported by the observation that the body temperature normalized in most patients before initiation of therapy. The reabsorption of pleural fluid in TB pleurisy can take up to a year 10 if complete drainage is not performed. In the study by Lee and coworkers, 10 the average reabsorption rates of pleural effusions were 54.5 days in prednisolone-treated patients and days in placebo-treated patients. Better assessment of the underlying lung parenchyma after drainage is a further consideration in favor of the complete removal of pleural fluid. Taken together, these data suggest that a TB pleural effusion is best managed by complete drainage. In our opinion, thoracoscopy presents the best approach for both diagnosis and simultaneous complete drainage of the effusion of TB pleurisy. In the absence of thoracoscopy facilities, drainage should be attempted as completely as possible by needle thoracentesis. The incidence of residual pleural thickening in our sample was comparable to that of previous reports In contrast to research by Hulnick and coworkers,20 HRCT was not found to be more sensitive than CXR in diagnosing pleural thickening, since the difference in diagnostic yield of the two methods was not statistically significant. Assessment of residual pleural thickening by HRCT or CXR failed to reveal any beneficial effect of additional prednisone in preventing this complication. The final radiologic outcome of patients with the various macroscopic pleural appearances at thoracoscopy showed no significant difference between prednisone- and placebo-treated patients either for the acute inflammatory (type 1 and 2) or fibrous inflammatory (type 3) types. Clinically and functionally significant sequelae of pleural thickening are rare when TB pleurisy is treated with standard anti-tb chemotherapy and drained completely. It is of note that neither of our treatment groups experienced any recurrence of their pleural effusions. Lung function impairment as a result of pleural thickening is an occasional sequela of TB pleural effusion. In keeping with research on children by Filler and Porter, 21 our study found that the addition of prednisone did not influence the final PFT result outcome of the disease in an adult sample. Both groups demonstrated a significant improvement in all PFT parameters after 3 and 6 months, but group differences were not significant. Both groups demonstrated a significant number of patients with a restrictive pattern in PFT results at 6 months. Whether this reflects residual pleural disease or other factors is not clear. A continued follow-up of all patients is being conducted at 6-monthly intervals to detect the development of late fibrothorax or relapse of active TB. Side effects ascribed to prednisone are well recognized and occurred in few patients in both treatment groups, the most important being epigastric pain. Body weight increased significantly in both groups during treatment. Withdrawal of prednisone therapy due to adverse effects was not necessaty in any patient. A potential concern of the addition of corticosteroid therapy might be the dissemination or progression of TB. This was not observed in any of our patients, which is in accordance with earlier studies. 22,2 3 We conclude that standard anti-tb therapy and early complete drainage is adequate for the treatment CHEST /110/2/ AUGUST,

6 of TB pleurisy. The addition of short-term oral prednisone therapy neither results in clinically relevant earlier symptom relief nor confers a beneficial effect on residual pleural thickening. The question whether achievement of early drainage is superior to no drainage for the long-term outcome has not been answered by our study and would necessitate further controlled studies. ACKNOWLEDGMENTS: We acknowledge the help and encoura_gement of Prof. T.R. Joubert and Drs. H. Welke and R. Gie in the development of tbis project. \Ne are indebted to Z. Buttner, M. Plaatjies, and J. Mouton tor their tireless work as research assistants, to M. Engelbrecht for randomization and drug preparation, to Prof. P.G. Bardin and Drs. C. Corbett, J.J. Jansen, and C.T. Bolliger for their critical review of this manuscnpt and to Dr. N. Smuts Torreporting the HRCT. REFERENCES 1 Aspin J, O'Hara H. Steroid-treated tuberculous pleural effusions. Br J Tuber 1958; 52: Grewal KS, Dixit RP. A comparative study of therapeutic regimens with and vvithout corticosteroids in the treatment of tuberculous pleural effusion. J Indian Med Assoc 1969; 52: Damany SJ, Shah KT. Treatment of pleural effusion vvith and \Vithout triamcinolone in addition to usual antituberculosis chemotherapy. J Indian Med Assoc 1968; 51: Mathur KS, Prasad R, Mathur JS. Intrapleural hydrocortisone in tuberculous pleural effusion. Tubercle 1960; 40: Mathur KS, Mathur JS, Sapru RP. Treatment of tuberculous pleural effusion with local instillation ofhydrocortisone. Dis Chest 1965; 47: Menon NK. Steroid therapy in tuberculous effusion. Tubercle 1964; 45: Paley SS, Mihaly JP, Mais EL, eta!. Prednisone in the treatment of tuberculous pleural effusions. Am Rev Tuber 1959; 79: Smith MHD, Matsaniotis N. Treatment of tuberculous pleural effusions with particular reference to adrenal corticosteroids. Pediatrics 1958; 22: Tani P, Poppius H, Makipaja J. Cortisone therapy for exudative tuberculous pleurisy in the light of a follow-up study. Acta Tuber Scand 1964; 44: lo LeeCH, Wang \VJ, Lan RS, eta!. Corticosteroids in the treatment of tuberculous pleurisy: a double-blind, placebo-controlled randomized study. Chest 1988; 94: Boutin C, Loddenkemper R, Astoul P. Diagnostic and therapeutic thoracoscopy: techniques and indications in pulmonary medicine. Tuber Lung Dis 1993; 74: American Thoracic Society. Clinical role of bronchoalveolar lavage in adults with pulmonary disease. Am Rev Respir Dis 1990; 142: Bergqvist S, Nordenstam H. Thoracoscopy and pleural biopsy in the diagnosis of pleurisy. Scand J Respir Dis 1966; 47: McCormack HM, Home DJ, Sheather S. Clinical applications of visual analogue scales: a critical review. Psycho] Med 1988; 18: Im JG, Webb \VH, Rosen A, eta!. Costal pleura: appearances at high-resolution CT. Radiology 1989; 171: American Thoracic Society Statement. Snowbird workshop on standardization of spirometry. Am Rev Respir Dis 1979; 119: Schoenberg JB, Beck GJ, Bouhuys A. Grov.th and decay of pulmonary function in healthy blacks and whites. Respir Physiol 1978; 33: Goldman HI, Becklake MR. Hespiratory function tests: normal values at median attitudes and the prediction of normal results. Am Rev Tuber 1959; 79: Barbas CSV, Cukier A, de Varvalho CRH, eta!. The relationship between pleural fluid findings and the development of pleural thickening in patients with pleural tuberculosis. Chest 1991; 100: Hulnick DH, Naidich DP, McCauley DI. Pleural tuberculosis evaluated by computed tomography. Radiology 1983; 149: Filler J, Porter M. Physiologic studies of the sequelae of tuberculous pleural effusion in children treated with antimicrobial drugs and prednisone. Am Rev Respir Dis 1963; 88: Houghton LE. Combined corticotrophin therapy and chemotherapy in pulmonary tuberculosis with special reference to hypersensitive reactions. Lancet 1954; 1: Muthuswamy P, Hu TC, Carasso B, et a!. Prednisone as adjunctive therapy in the management of pulmonary tuberculosis: report of 12 cases and review of the literature. Chest 1995; 107: Clinical Investigations