Pneumocystis pneumonia in brain tumor patients: risk factors and clinical features

Journal of Neuro-Oncology 27: 235-240,1996. © 1996 Kluwer Academic Publishers. Printed in the Netherlands.

Clinical Study

Pneumocystis pneumonia in brain tumor patients: risk factors and clinical features

David Schiff Division of Neuro-Oncology, Department of Neurology, Mayo Clinic and Mayo Foundation, USA

Key words." Pneumocystis carinii pneumonia, brain tumor, neoplasm, corticosteroids, dexamethasone, lymphopenia

Summary

We reviewed the clinical features and risk factors for Pneumocystis carinii pneumonia (PCP) in patients with brain tumors (BTs) seen at our institution between 1980 and 1992. Previously rare, this opportunistic infection appears to be increasing among HIV-negative cancer patients receiving immunosuppressive medications. Recent reports have noted PCP among BT patients receiving corticosteroids, and suggested that these patients are particularly likely to develop PCP when corticosteroids are tapered.

Nine BT patients, eight with high-grade gliomas, experienced ten episodes of PCR None were known HIV- positive. All were on dexamethasone (DXM) at PCP onset, and had continuously been receiving it for 47-398 days (median 69). Daily DXM dose at PCP onset ranged from 1-16 mg (median 9). Five episodes occurred in patients receiving a stable DXM dose and five during DXM taper. Nine episodes occurred in patients receiving chemotherapy. All patients had absolute lymphopenia at PCP onset, ranging from 80-900 x 106 lymphocytes/1 (median 222 x 106/1, normal > 1000 x 106). Three episodes were fatal despite appropriate antibiotic therapy.

Unlike others, we did not find that corticosteroid taper predisposed to developing PCR As in HIV, PCP in BT patients appears related to lymphopenia, in these patients attributable to use and duration of corticosteroids and in some cases cytotoxic chemotherapy. Effective prophylaxis exists and should be considered for lymphopenic patients and those requiring DXM for > five weeks.

Introduction

Pneumocystis carinii is an organism colonizing the lower respiratory tract of humans and rodents. Al- though its sensitivity to trimethoprim-sulfamethoxazole (TMP-SMX) and pentamidine suggests clas- sification as a protozoan, ribosomal RNA analysis reveals more homology with fungi [1]. 75% of children have antibodies to Pneumocystis by age four [2]. Pneumocystis generally causes disease only in immunocompromised hosts, and prior to 1980 PCP was a rare disease primarily confined to renal trans- plant recipients and children with acute leukemia.

Although PCP is often fatal, with appropriate antibiotics it is successfully treatable if diagnosed expeditiously. Furthermore, prophylactic antibiotic use is effective in groups at high risk of PCR With the advent of HIV, PCP is now a major infectious cause of death in the United States [2]. While most of this increase has occurred in HIV-positive patients, increasing numbers of cases have been reported in HIV-negative patients with solid tumors, including brain tumors. The reason for this increase is un- clear. Recent reports have suggested that corticosteroid use is the major risk factor in solid tumor patients, and that brain tumor patients are most likely

236

to develop PCP while corticosteroids are being tapered [3-5]. To further our understanding of PCP development in patients with brain tumors, we reviewed our institutional experience.

Methods

Using computerized lists of discharge diagnoses, we identified all patients seen at the Mayo Clinic between January 1, 1980 and December 31, 1992 with diagnoses of primary brain tumor and PCE All primary brain tumors were histologically confirmed. Diagnosis of PCP in all cases was based on an appropriate clinical setting in conjunction with identi- fication of pneumocysts by bronchoalveolar lavage or open lung biopsy utilizing a methenamine silver or calcofluor stain.

Information recorded from the medical record included the following: age, sex, tumor histology and grade, radiation and chemotherapy received, HIV status, duration of corticosteroid use, maximum dose of corticosteroids received, corticoste-

Table 1. Clinical background of patients with PCP

roid dose at diagnosis of PCR recent changes in corticosteroid dose, and use of prophylaxis against PCR We additionally recorded duration of respiratory symptoms prior to diagnosis of PCR nature of symptoms, white blood cell and lymphocyte count at diagnosis, initial chest radiograph results, initial room air arterial blood gas (ABG), means of diagnosis, treatment, outcome, and autopsy results. The alveolar-arterial oxygen gradient (AaDO2) was calculated using standard formulas. Total lymphocyte count was calculated by multiplying the total white blood cell count by the percentage of lymphocytes in the differential.

Results

Table i presents the clinical background of the nine patients with PCR one of whom had two cases of PCR Eight patients had supratentorial high-grade gliomas, while the ninth had an unresectable he- mangioblastoma of the cerebellar vermis and IVth ventricle. Males and females were equally repre-

Pt# Age/Sex Tumor Chemotherapy HIV Dura t ion D X M dose Recent D X M Maximal

D X M a (mg) at PCP taper dose (mg)

(days) onset D X M

1 65M

2 65M

3 45F

4 39F

5 64M

6 47F 6 48F 7 21M

8 58F

9 51M

Grade 3 as t rocytoma B C N U

Grade 4as t rocy toma P C NU

Grade 4 oligoastro- B C N U

cytoma Grade 4 astrocytoma B C N U

Grade 4 astrocytoma B C N U

Grade 4 astrocytoma Grade 4 astrocytoma M O P d Hemangiob las toma of B C N U and vermis dianhydrogalactitol Grade 4 astrocytoma B C N U

Grade 4 astrocytoma B C N U

Neg u 58 12 24-12 previous 4 24

weeks

Unk: 54 8 16-8 previous 2 16

weeks

U n k 91 10 No 40

Unk 134 12 16-12 previous 4 16 weeks

Unk 52 1 8-1 previous 3 16

weeks

Neg 64 10 No 16 Neg 398 4 No 16 Unk 294 6 No 8

Unk 74 2 16-2 previous 5 16

weeks Unk 47 16 No 16

a = dexamethasone b = negative c = unknown d = mechlorethamine, vincristine, and procarbazine.

sented, and ages ranged from 21 to 65 (median 50). Two of the nine were HIV-negative by ELISA; HIV status of the other patients was unknown, but none had known risk factors for HIV or clinical histories suggesting prior immunodeficiency. All patients were receiving dexamethasone (DXM) at the onset of PCR and had continuously been receiving it for 47-398 days (median 69 days). DXM dose at onset of PCP ranged from 1-16 mg (median 9 mg). Five of ten episodes occurred in patients receiving a stable dose of DXM. Four episodes occurred during steroid tapers of _> 2 mg/week, a fifth occurred in a patient tapered four mg over the preceding four weeks. Nine of ten episodes occurred in patients receiving chemotherapy. As recorded in Table 2, all patients had absolute lymphopenia (defined as < 1,000 x 106/1) at onset of PCR ranging from 80- 900 x 106 lymphocytes/1 (median 222 x 106/1). Five episodes occurred in patients who were leukopenic

Table 2. Clinical and laboratory findings in patients with PCP

237

(WBC < 3,500 x 106/1) at the onset of PCP; three other episodes occurred in patients who had documented leukopenia within the preceding three weeks but who had normal WBC counts at onset of PCE

The clinical presentations of PCP in our patients are described in Table 2. All patients were febrile, and 90% either had cough (generally nonproduc- tive), dyspnea, or both. The duration of symptoms prior to diagnosis ranged from 4-28 days (median, 7 days). Only one patient had symptoms for longer than two weeks prior to diagnosis. All patients were outpatients when they developed symptoms, and all were subsequently ]hospitalized. The initial room air ABG revealed a PaO2 between 37-63 (median 46); all patients had a component of respiratory al- kalosis. All patients had an elevated AaDO 2 ranging from 47-77 (median 70, normal < 20). The chest radiograph was invariably abnormal, revealing

Pt# Clinical Duration CXR b Initial room AaDO2 d WBC ~ Absolute Diagnostic

features symptoms air ABG c (x 106/liter) lymphocyte method

before Dx a count*

(days)

Treatment Outcome

1 PC f, D g, F h, 6 Bilateral alveolar 42/7.52/27 74 5,800 900 BAU TMP-SMX k Recovered CP i infiltrates

2 NC, D, F 6 Bilateral interstitial 46/7.47/35 60 5,800 580 BAL TMP-SMX Recovered infiltrates

3 NC ~, E D 28 Bilateral interstitial 63/7.50/27 53 1,400 266 BAL Pentamidine Recovered infiltrates

4 E D, NC 7 Bilateral interstitial 37/7.51/31 74 3,200 192 BAL Pentamidine Died and alveolar

infiltrates

5 F, NC 7 Bilateral perihilar 46/7.54/26 71 1,200 252 BAL TMP-SMX Recovered infiltrates

6 F, NC 11 Bilateral interstitial 52/7.46/28 63 3,600 180 BAL TMP-SMX Recovered infiltrates

6 E D 14 Bilateral alveolar 39/7.51/27 77 2,300 184 BAL TMP-SMX Died 6 days

infiltrates later 7 E D 7 Bilateral alveolar 50/7.53/24 70 5,600 140 OLB m TMP-SMX Improved:

infiltrates died 21 days

later 8 F 4 Bilateral alveolar 55/7.45/38 47 8,500 255 BAL TMP-SMX Recovered

infiltrates

9 E NC 4 Right alveolar 44/7.45/24 76 500 80 BAL TMP-SMX Died 2 days infiltrate after dx

= diagnosis; b chest x-ray; c arterial blood gas; d alveolar-arterial oxygen gradient; * (x 106/liter; e white blood cell count;

cough; g dyspnea; h fever; ~ chest pain; J bronchoalveolar lavage; k t r imethoprim/sulfamethoxazole; ~ nonproduct ive cough; biopsy.

productive m open lung

238

either an alveolar or interstitial pattern of infiltrates. Nine episodes were diagnosed with bronchoalveolar lavage, and one on the basis of open lung biopsy. Two patients with sulfa allergies were treated with intravenous pentamidine; the others all received intravenous TMP-SMX followed by oral TMP-SMX when stable. Three patients died as a direct result of PCP; a fourth died of tumor pro- gression three weeks after the onset of PCR No patient had documented extrapulmonary Pneumocys - tis.

Discussion

The clinical features of our patients are similar to previous reports of HIV-negative patients with PCP [3, 4, 6-8]. In general, these patients have symptoms lasting a few days to a week prior to hos- pitalization, in contrast to patients with HIV, who often have a much more gradual onset over a few weeks [7]. As in the other series of BT patients, our patients were invariably hypoxic with an elevated AaDO 2 gradient [3, 5].

All of our patients were receiving corticosteroids at the time PCP developed. In general, series of HIV-negative PCP report corticosteroids as a risk factor in 87-100% of patients [3-6, 8, 9]. Occasional patients not receiving corticosteroids but iatrogen- ically immunosuppressed with cytotoxic drugs have been reported with PCP as well [1, 4,10]. In one series, only 5/17 patients received corticosteroids, but most patients had hematologic malignancies and 15/ 17 received cytoreductive chemotherapy [1].

Prior series have suggested that PCP is likely to occur in BT patients when corticosteroids are tapered. 80% of cases reported in two prior series occurred during DXM taper [3, 5]. These authors speculated that corticosteroid taper unmasked pulmonary inflammation from PCP that had developed during corticosteroid therapy, a phenomenon observed in animal studies [11,12]. Conversely, corticosteroids are useful adjunctive treatment in HIV- positive PCP patients with hypoxia, presumably through a pulmonary anti-inflammatory mechanism [13]. In the present series, only half the cases occurred during steroid taper. A tendency for cases

to present during taper may be an artifact of the physician's tendency with BT patients to increase corticosteroids dramatically and then gradually to taper as tolerated. Clearly corticosteroid taper is not necessary for PCP to develop.

Corticosteroids are likely the most important risk factor for PCP in BT patients. Of 25 reported cases of PCP in BT patients (including the current series), 100% occurred in patients receiving DXM, and all had been receiving DXM for > five weeks [3, 5]. DXM doses at PCP onset have ranged from 1- 60 mg (median 8 mg). Maximum previous dose of DXM has generally been _ 16 mg, although one case developed after a maximum dose as low as 8 mg. 60% of reported cases developed in patients who were also concurrently receiving chemotherapy.

All episodes in the current series occurred in patients with absolute lymphopenia. Three prior series of PCP in HIV-negative patients noted an asso- ciated with lymphopenia. Godeau et al. noted that 91% of PCP cases in patients with connective tissue disorders occurred in lymphopenic patients, and 50% had total lymphocyte counts < 400 x 106/1 [6]. Kovacs et al. reported a median lymphocyte count of 396 x 106/1 in HIV-negative PCP cases [7]. The third series noted that, 'peripheral lymphopenia was present in all for whom this information was available' [3]. The extensive experience with PCP in HIV-positive patients also supports an association with lymphpenia. The relative risk of PCP in HIV- positive patients increases to 4.9 when the CD4+ count falls below 200 x 106/1, leading to recommendations that patients receive PCP prophylaxis at or below that level [14, 15]. In immunocompetent humans, T cells comprise approximately 75 % of circulating lymphocytes, and about 1/3 of these are CD4+. Animal and human studies suggest that corticosteroids produce their lymphopenic effect primarily by decreasing T cell counts, and that CD4+ cells are particularly susceptible to this effect [16- 20]. Cytotoxic chemotherapy presumably also pre- disposes to PCP by effects on T lymphocytes, as both animal models and human cases suggest that cell-mediated immunity is the main defense against PCP and that antibodies and granulocytes play little role [17, 19]. Since the maximum lymphocyte count

in our patients was 900 x 106/1, it is likely that most, if not all, of our patients had CD4+ counts <200 x 106/1. Thus, decreased CD4+ counts may repre- sent a common link between PCP in HIV-positive patients and BT patients. Future studies should ex- amine lymphopenia and T cell subset counts at risk factors for PCP in BT patients.

A case series such as this has obvious limitations. It is likely that other cases went undiagnosed, and many patients with an acute respiratory illness may have sought local care rather than return to a dis- tant tertiary care hospital. Consequently, we have not attempted to calculate the incidence of PCP in BT patients (elsewhere estimated at 2-6%) [3, 5]. Nonetheless, it is clear that BT patients requiring prolonged courses of corticosteroids are at risk of developing PCE and that this illness is often fatal. BT patients with unexplained fever or respiratory symptoms, especially if on DXM for > five weeks, with lymphopenia, and an elevated AaDO 2, should be evaluated expeditiously for PCP (and probably presumptively treated until PCP is excluded). Ef- fective prophylaxis for PCP exists; TMP-SMX 1 DS b.i.d, two or three days weekly is extremely effective at preventing PCP in HIV-negative patients [13, 21-23]. TMP-SMX is very well tolerated in the general population, with 4% of patients reporting mi- nor gastrointestinal symptoms and 3.5% dermato- logic reactions [24, 25]. TMP-SMX rarely causes leukopenia and thrombocytopenia; it also has an antifolate effect that could theoretically cause meg- aloblastic bone marrow failure [26]. However, even in patients receiving concomitant chemotherapy these complications are rare; in the extensive pedi- atric experience with acute lymphoblastic leukemia patients, only 2% of patients were unable to toler- ate TMP-SMX because of persistent rash or bone marrow suppression [23]. One month of prophylaxis with generic TMP-SMX costs only $ 8.83 at our institution's outpatient pharmacy. Dapsone and aerosolized pentamidine are reasonably effective alternative that have been primarily utilized in HIV-positive patients, who are much more prone to adverse reactions from TMP-SMX [27, 28]. Al- though it is debatable whether the incidence of PCP in this population warrants prophylactic antibiotics in all patients receiving corticosteroids, considera-

239

tion should be given to prophylaxis in chronically corticosteroid-dependent BT patients, and perhaps particularly those with lymphopenia.

References

1. Varthalitis I, Aoun M, Daneau D, Meunier F: Pneumocystis carinii pneumonia in patients with cancer. Cancer 71: 481- 485, 1993

2. Hughes WT: Pneumocystis carinii pneumonia. In: Gorbach SL, Bartlett JG, Blacklow NR (eds) Infectious Diseases, W.B. Saunders, Philadelphia 494-497, 1992

3. Henson JW, Jalaj JK, Walker RW et aL: Pneumocystis carinii pneumonia in patients with primary brain tumor. Arch Neu- ro148: 406-409,199l

4. Sepkowitz KA, Brown AE, Telzak EE et aL: Pneumocystis

carinii pneumonia among patients without AIDS at a cancer hospital. JAMA 267: 832-837, 1992

5. Slivka A, Wen PY, Shea WM, Loeffler JS: Pneumocystis carinii pneumonia during steroid taper in patients with primary brain tumors. Am J Med 94: 216-219, 1993

6. Godeau B, Coutante-Perronne V, Huong DLT et aL: Pneu- mocystis carinii pneumonia in the course of connective tissue disease: Report of 34 cases. J Rheumatol 21: 246-251, 1994

7. Kovacs JA, Hiemenz JW, Macher AM et aL: Pneumocystis carinii pneumonia: A ,comparison between patients with the acquired immunodeficiency syndrome and patients with other immunodeficiencies. Ann Int Med 100: 663-671, 1984

8. Peters SG, Prakash UBS: Pneumocystis carinii pneumonia: Review of 53 cases. Mayo Clin Proc 82: 73-78, 1987

9. Chechani V, Bridges A: Pneumocystis carinii pneumonia in patients with connective tissue disease. Chest 101: 375-378, 1992

10. Porter DR, Marshall DAS, Madhok R et aL: Pneumocystis carinii infection complicating cytotoxic therapy in two patients with lymphopenia, but a normal total white cell count. Br J Rheumatol 31: 71-72,1992

11. Frenkel JK, Good JY, Shultz JA: Latent Pneumocystis infection of rates, relapse, and chemotherapy. Lab Invest 15: 1559-1577, 1966

12. Walzer PD, Powell RD Jr, Yoneda K et al.: Growth charac- teristics and pathogenesis of experimental Pneumocystis carinii pneumonia. Infect Immuno127: 928-937,1980

13. Masur H: Prevention and treatment of Pneumocystis pneumonia. N Engl J Med 327: 1853-1860,1992

14. Phair R Munoz A, Detels R et aL: The risk of Pneurnocystis carinii pneumonia among men infected with human immunodeficiency virus type 1. N Engl J Med 332: 161-165,1990

15. Task Force Recommendations (Recommendations for prophylaxis against Pneumocystis carinii pneumonia for adults and adolescents with Human Immunodeficiency Virus). MMWR 41: 1-11,1992

16. Fauci AS, Dale DC, Balow JE: Glucocorticoid therapy:

240

Mechanism of action and clinical considerations. Ann Int Med 84: 304-315, 1976

17. Masur H, Lane HC, Kovacs JA et al.: Pneumocystis pneumonia: From bench to chair. Ann Int Med 111: 813-826, 1989

18. Slade JD, Hepburn B: Prednisone-induced alterations of circulating human lymphocyte populations. J Lab Clin Med 101: 479-487, 1983

19. Walzer PD, Labine M, Redington TJ, Cushion MT: Lym- phocyte changes during chronic administration of and with- drawal from corticosteroids: Relation to Pneumocystis carinii pneumonia. J Immuno1133: 2502-2508, 1984

20. Zweiman B, Atkins PC, Bedard P-M et al.: Corticosteroid effects on circulating lymphocyte subset levels in normal humans. J Clin Immunol 4: 151-155, 1984

21. Hughes WT, Kuhn S, Chaudhary S et al.: Successful prophylaxis for Pneumocystis carinii pneumonitis. N Engl J Med 297: 1419-1426, 1977

22. Hughes WT, Rivera GK, Schell MJ et al.: Successful inter- mittent prophylaxis for Pneumocystis carinii pneumonia. N Engl J Med 316: 1627-1632, 1987

23. Wilber RB, Feldman S, Malone WJ et al.: Chemoprophylax-

is for Pneumocystis carinii pneumonia: Outcome of unstruc- tured delivery. Am J Dis Child 134: 643-648, 1980

24. Lawson DH, Jick H: Adverse reactions to co-trioxazole in hospitalized medical patients. Am J Med Sci 275: 53-57, 1978

25. Lawson DH, Paice B J: Adverse reactions to trimethoprim- sulfamethoxazole. Rev Inf Dis 4: 429, 1982

26. Rubin RH, Swartz MN: Trimethoprim-sulfamethoxazole. N Engl J Med 303: 426-432, 1980

27. Bozzette SA, Finkelstein DM, Spector SA et al.: A rando- mized trial of three antipneumocystis agents in patients with advanced human immunodeficiency virus infection. N Engl J Med 332: 693-699, 1995

28. Hardy WD, Feinberg J, Finkelstein DM et aL: A controlled trial of trimethoprim-sulfamethoxazole or aerosolized pentamidine for secondary prophylaxis of Pneumocystis carinii pneumonia in patients with the acquired immunodefiecien- cy syndrome. N Engl J Meal 327: 1842-1848,1992

Address for offprints: D. Schiff, Mayo Clinic, 200 First Street SW, Rochester MN 55905, USA

Documents

Pneumocystis pneumonia in brain tumor patients: risk factors and clinical features