F and therapeutic problem in patients with

Transcription

1 MANAGEMENT OF FEVE OF UNKNOWN OIGIN IN PATIENTS WITH NEOPLASMS AND NEUTOPENIA VICTOIO ODIGUEZ, MD, MICHAEL BUGESS, AND GEALD P. BODEY, MD* MD, During 81 -febrile episodes, 76 cancer patients with neutropenia were randomly allocated to continue or discontinue 4 days after initiation of carbenicillin and therapy, if no infection had been demonstrated. During 56 episodes, the patients became afebrile, after initiation of. Infection as a cause of fever was identified in 21% of the episodes. The cause of fever could not be identified in 72% of the episodes. Three of 30 patients randomized to discontinue developed infection which ultimately caused their death. During 25 episodes, the patients remained febrile. Infection was the ultimate cause of fever in 40% of the episodes. The cause of fever could not be identified in 48%. The majority of infections documented in this group responded when gentamicin was added. Antibiotic therapy with carbenicillin and is effective initial therapy for fever due to presumptive infection. If, after 4 days of therapy, no infection is documented and the patient is responding, the should be continued for an additional 3 to 5 days. However, for patients not responding after the initial 4 days of therapy, the addition of gentamicin is indicated. EVE CONTINUES TO E A MAJO DIAGNOSTIC F and therapeutic problem in patients with malignancies, especially in leukemic patients with neutropenia.6.7 Over 80% of febrile episodes in these patients are due to infecti~ns.&~j~ Neutropenic individuals are unable to respond to infections with a normal inflammatory response, and, therefore, these infections may disseminate rapidly and have a fatal outcome if not treated promptly.2.8 Often due to this lack of inflammatory response, the classical clinical signs of infections are absent in these patients. Consequently, fever in neutropenic patients should be regarded as an early sign of infection which justifies the initiation of antimicrobial therapy.3jojl Although infection can be established as the cause of febrile episodes in the majority of From the Department of Developmental Therapeutics, The University of Texas M. D. Anderson Hospital and Tumor Institute at Houston, Houston, Tex. Supported in part, by Grants CA and CA from the National Cancer Institute, National Institutes of Health, U. S. Public Health Service, Bethesda, Md. Scholar of The Leukemia Society of America, Inc. Address for reprints: Dr. Victorio odriguez, MD, M. D. Anderson Hospital, 6723 Bertner, Houston, Tex eceived for publication February 28, neutropenic patients, in a substantial number no etiologic agent can be identified. These fevers of unknown origin may either defervesce promptly after initiation of antibiotic therapy, or fail to respond. The management of both of these types of febrile episodes is a serious problem. Prompt response to antibiotic therapy may indicate that the patient has infection which is responding to appropriate therapy or that the response was coincidental and the patient would have become afebrile without. If the patient is infected, then it would be undesirable to discontinue, but if he has no infection, continuing antibiotic therapy would be undesirable because of the cost and the possible risk of predisposing the patient to infection caused by antibiotic-resistant organisms.1 Patients with fever of unknown origin which persists despite antibiotic therapy pose a different problem. It may be detrimental to discontinue antibiotic therapy even though it appears to be ineffective. On the other hand, the may interfere with isolation of an infectious organism from culture specimens. The management of febrile episodes not

2 1008 CANCE October 1973 Vol. 32 TABLE 1. andomization of Patients with Neutropenia and Fever At onset of fever (lolf or greater) Carbenicillin plus Cephalothin X 4 davs Day 4 No fever ( Continue ontinue With fever<kscontinue proven to be due to infection has not been systematically studied. The present study was designed to determine whether the appropriate approach should be to continue or discontinue. The study indicates that should be continued in patients whose fever responds because some of these patients are infected. Patients whose fever continues are also likely to be infected but require additional. MATEIALS AND METHODS Febrile patients with neutropenia (less than 1000 polymorphonuclear leukocytes/mm3 of blood) admitted to the Department of Developmental Therapeutics at the University of Texas M. D. Anderson Hospital and Tumor Institute, from July 1970 to December 1971, were eligible for this study when they developed fever. Fever was defined as temperature of lolf or greater not related to transfusion of blood products. Patients with fever which persisted for more than 4 hours after transfusion were also eligible for this study. Cultures of throat, urine, blood, and other appropriate sites were obtained from all patients before initiation of antibiotic therapy. Blood cultures were repeated daily so long as fever persisted. X-ray examination of the chest and urinalysis were obtained at the onset of fever, on day 4 of the study and subsequently as clinically indicated. White blood counts TABLE 2. Fever of Ilnknown Origin in Neutropenic Patients esponding to Antibiotics Continue Total Episodes Cause of fever: Infection 4(15) 8(27) 12(21) Neoplasia 2 (8) 1 (3) 3 ( 5) Drug fever 1(4) 0 1(2) FUO 19 (73) 21 (70) 40(72) Numbers in parentheses are percentages. were obtained daily in the majority of the patients. Initial antibiotic therapy consisted of the combination of carbenicillin and. Carbenicillin was given intravenously in doses of 5 g in 100 ml of 5% glucose solution in 1- to 2-hour infusions every 4 hours. Cephalothin was administered intravenously in doses of 3 g in 50 ml of 5% glucose solution over a onehour period every 6 hours. Patients in whom infection was documented within 4 days were treated with appropriate and were not included in this study. Only patients in whom no infection was documented after 4 days of carbenicillin and therapy were entered on the study and were randomly allocated to continue or discontinue antibiotic therapy. Separate randomizations were made for those patients who had become afebrile after the initiation of and for those patients who had persistent fever. The randomization schedule of this study is shown in Table 1. The duration of additional therapy was 10 days, or 5 days after becoming afebrile, which ever was longer. Whenever an infection was documented with an organism not sensitive to either carbenicillin or, the antibiotic regimen was modifed to include the appropriate antibiotic. ESULTS Eighty-one episodes of fever of unknown origin (FUO) occurred in 76 patients. There were 47 males and 29 females whose ages ranged between 15 and 80 years (median 33 years). Sixty-two had acute leukemia (52 with acute myelogenous leukemia and 10 with acute lymphoblastic leukemia), 4 had blastic transformation of chronic myelogenous leukemia, 6 had malignant lymphoma, and 4 had metastatic carcinoma. All patients were receiving intensive antineoplastic chemotherapy and had neutropenia. During 56 episodes (69y0), the patients became afebrile after the initiation of antibiotic therapy with carbenicillin plus (Table 2). None of these patients had evidence of infection before day 4, and they were randomized to continue or discontinue. During 26 episodes, the patients were randomized to continue receiving the antibiotic combination. In this group, the ultimate cause of fever was found to be infection in four episodes (15 %), neoplasia in two (8yob), and drug fever in one (4y0). The cause of

3 No. 4 FEVE OF UNKNOWN OIGIN - odriguez et al fever could not be identified in 73y0 of 26 episodes. During the remaining 30 episodes, the patients were randomized to discontinue antibiotic therapy. In this group, the fever was ultimately proven to be due to infection in eight episodes (27y0) and neoplasia in one (3y0). The cause of fever could not be ascertained in 70y0 of these 30 episodes. The types of infections subsequently diagnosed in the group of patients who became afebrile during therapy with carbenicillin and are shown in Table 3. These infections were diagnosed after the randomization had taken effect based on repeated diagnostic studies. Eight of the 12 infections were caused by organisms of the -Enterobarter group. Two of the eight patients who were randomized to discontinue developed overwhelming infection which ultimately caused their death. E.B. was a patient whose bronchopneumonia became clinically evident 1 day after discontinuation of the antibiotic therapy and did not respond to prompt reinitiation of the same. In patient K.H., fever recurred 48 hours after dis- continuation of due to overwhelming which caused her death 6 days later despite reinstitution of the same. The organisms causing fatal infection in these two patients were sensitive to. In patients M.Gl., A.F., and G.B., the antibiotic regimen was modified even though they had become afebrile because the sensitivity tests revealed that the organisms causing infection were resistant to carbenicillin and. Patient M.G. had two separate episodes of pneumonia. The first time, the microorganism was sensitive to and the patient responded to carbenicillin and ; another episode of pneumonia occurred one month later. This time the microorganism was resistant to, but the patient responded to gentamicin. Patient M.Gu. had Candid pneumonitis which failed to respond to amphotericin B therapy. During 25 episodes, the patients remained febrile after 4 days of therapy with carbenicillin and (Table 4). None of these patients had evidence of infection before day TABLE 3. Typeof Infections* in Patients with FUO esponding to Carbenicillin and Cephalothin Pa tien t Fourth day randomization Infection Sensitivity to Carb. and Ceph. Outcome M.GI. M.G. Continue Continue Enterobacter pneumonitis S D.F. Continue Staph. aureus - S U..I.6 F.D. Continue Pneumonitist - - F.. Pneumonitist - - M.G. pneumonitis M.Gu. Candida - - pneumonitis: E.B. A.F. pneumonitis S K.H. S c.r. S U.T.I.11 G.B. En terobacter SePSlS * Diagnosed after the randomization had taken effect. t Diagnosis confirmed by x-rays. * Diagnosis confirmed at autopsy. 8 U..I. = Upper respiratory infection. I LJ.T.1. = Urinary tract infection. esponded to chloramphenicol + gentamicin esponded to carbenicillin + esponded to carbenicillin + esponded to carbenicillin + No further ; continued to improve esponded to gentamicin Failed to respond to antifungal therapy; died Failed to respond to carbenicillin + ; died esponded to gentamicin -t chloramphenicol Failed to respond to + gentamicin; died esponded to carbenicillin + esponded to gentamicin + chloramphenicol

4 ~ ~~ 1010 CANCE October 1973 Vol. 32 TABLE 4. Fever of Unknown Origin in Neutropenic Patients Falling to espond to Antibiotics Continue Total Episodes Cause of fever: Infection 6 (43) 4(36) ll(40) Neoplasia 2 (14) 1(9) 3 (12) Drug fever FUO 6 (43) 6(55) 12(48) Numbers in parenthesis are percentages. 4, and they were randomized to continue or discontinue. During 14 episodes, the patients were randomized to continue. In this group, the cause of fever was ultimately proven to be infection in six episodes (43y0) and neoplasia in two (14%). The cause of fever could not be identified in the remaining six (43y0) episodes. During 11 episodes, the patients were randomized to discontinue. In this group, the ultimate cause of fever was infection in four episodes (36y0) and neoplasia in one (9%). The cause of fever could not be established in the remaining six (55%) episodes. The types of infections diagnosed in these patients who failed to respond to carbenicillin plus cephal- othin are shown in Table 5. No microorganisms were recovered during three pneumonitis episodes and one cellulitis episode, which occurred in four patients who had been randomized to continue carbenicillin plus. Carbenicillin was substituted for gentamicin when the infection was identified in those four patients, but only two of them responded. A microorganism causing infection was recovered from the remaining two patients. Patients D.. and P.E. had infections due to sp. which was sensitive to. D.. responded to carbenicillin plus, but P.E. only responded when gentamicin was added. Four febrile patients who were randomized to discontinue subsequently developed with microorganisms resistant to carbenicillin and. They all responded when therapy which included gentamicin was reinstituted. The cause of persistent fever could not be identified in 12 patients. Fever in one of these patients defervesced after a course of therapy with amphotericin B. Another patient became afebfile after a transfusion of granulocytes but 12 days later developed which responded to antibiotic therapy. In nine patients, the fever subsided while receiving anti- TABLE 5. Types of Infections* in Patients with FUO Not esponding to Carbenicillin and Cephalothin Fourth day Sensitivity to Patient randomization Infection Carb. and Ceph. Outcome D.. Continue esoonded to carbenicillin + pneumonitis P.E. Continue esponded when gentamicin added cellulitis E.T. Continue Pneumonitist - Gentamicin added; did not respond and died W.C. Continue Pneumonitist - esponded when gentamicin added E.. Continue Cellulitis' - esponded when gentamicin added.h. Continue Pneumonitist -.S. L.G. J.H. L.B. Enterobacter 5 seps1s Pseudomonas semis * Diagnosed after the randomization had taken effect. + Diagnosis confirmed by x-rays. * No microorganisms recovered. Sensitivity not tested. Gentamicin added: did not respond and died cultured from heart blood esponded to carbenicillin + gentamicin esponded to carbenicillin + gentamicin esponded to gentamicin + esponded to gentamicin

5 No. 4 FEVE OF UNKNOWN OIGIN odriguez et al neoplastic chemotherapy which contained corticosteroids. DISCUSSION Ideally, antibiotic therapy should be withheld until the cause of fever is identified, a position long held by most infectious disease experts. However, many neutropenic patients will die if they are not treated promptly with effective. Approximately 33% of patients with Pseudomonas septicemia and 22% with septicemia at our institution died within 24 hours after onset of fe~er.1~8'4 Consequently, broad-spectrum anti- biotics should be instituted promptly before the diagnosis is established. This study was designed to determine what is the best course of action to take when no infection is identified after 4 days of broad-spectrum antibiotic therapy. During the period of this study, no infection could be identified at the onset of 40% of all febrile episodes occurring in neutropenic patients. Ultimately, 28% of these 81 episodes of fever of unknown origin were determined to be due to infection, and 7y0 were due to the underlying neoplastic disease. In one patient who developed fever with the administi ation of daunorubicin, it was considered to be drug fever. During 56 of the 81 episodes, the patients became afebrile after institution of antibiotic tlierapy. None of these patients had evidence of infection before day 4. Although only 12 of these episodes were finally proven to be due to infection, it is likely that many more were due to infections which promptly responded to antibiotic therapy. None of the patients randomized to receive further antibiotic therapy died. One patient (M.Gl.) had an infection caused by Enterobacter sp. which was resistant in vitro to carbenicillin and. Although she became afebrile on this regimen, the antibiotic therapy was replaced by chloramphenicol and gentamicin when the sensitivities became available. The remaining patients responded to carbenicillin and cephal- othin therapy. However, three patients among the group randomized to discontinue died. Two of these patients had infections which were caused by organisms sensitive to and were clearly responding befdre the were discontinued. Hence, the premature discontinuation of led to an unnecessary 70/, mortality rate. Therefore, those patients who promptly respond to initial antibiotic therapy should continue to receive this therapy for 7 to 10 days. The patients with persistent fever present a more difficult problem. In this study, 40y0 of these patients ultimately were found to have bacterial infections. The ability to eventually diagnose the infection was the same whether were continued or discontinued at the end of 4 days. The only fatalities occurred in two patients who continued to receive the original. Both patients also received gentamicin. Organisms were eventually isolated from six patients, and four were resistant to carbenicillin and. Gentamicin was administered during 10 of these 11 episodes of infection. Eight of the 10 patients were cured by the addition of gentamicin. This suggests that for patients failing to respond to carbenicillin and after 4 days of therapy, gentamicin should be added. Although only one patient in this study had a fungal infection, it has been our experience over the years that approximately 20% of patients with persistent fever unresponsive to broad-spectrum have this type of infection. Unfortunately, less than 25y0 of systemic fungal infections are diagnosed antemortem. This, and previous studies, suggest that for neutropenic cancer patients with fever of unknown origin the prompt initiation of therapy with carbenicillin and is advantageous.4.9 For those patients responding after 4 days, the should be continued for an additional 3 to 5 days. For those patients not responding after 4 days of therapy, the addition of gentamicin appears indicated. 1. Adler, J. L., Burke, J. P., and Finland, M.: Infection and antibiotic usage at Boston City Hospital. January Arch. Intern. Mcd. 127:460-46!5, Bodey, G. P., Buckley, M., Sathe, Y. S., and Freir- EFEENCES eich, E. J.: Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann. Intern. Med. 64:32&340, Bodey, G. P., odriguez, V., and Whitecar, J. P.:

6 1012 CANCE October 1973 Vol. 32 Severe infections in leukemic patientean approach to antibiotic therapy. In Advances and Management of Pseudomonas and Proteus Infections. Proceedings of the Symposium, July 12, 1969, F. Hoffman, Ed. New York, Exerpta Medica Foundation Publishers, 1970 pp Bodey. G. P., Whitecar, J. P., Jr., Middleman, E., and odriguez, V.: Carbenicillin therapy of Pseudomonas infections. JAMA 218:62-66, Boggs. D.., and Frei, E. 111: Clinical studies of fever and infection in cancer. Cancer 13: , Browder, A. A., Huff, J. W., and Petersdorf,. G.: The significance of fever in neoplastic disease. Ann. Intern. Med , Frei, E., 111, Levin,. H., Bodey, G. P., Morse, E. E., and Freireich, E. J.: The nature and control of infections in patients with acute leukemia. Cancer es. 25~ , Hersh, E. M., Bodey, G. P., Nies, B. A., and Freireich, E. J.: Causes of death in acute leukemia: A tenyear study of 414 patients from JAMA 193~ , Middleman, E. A., Watanabe, A., Kaizer, H., and Bodey, G. P.: Antibiotic combinations for infections in neutropenic patients. Evaluation of carbenicillin plus either or kanamycin. Cancer , odriguez, V., Gutterman, J. U., McMullan, G. K., and Heckman, A. A.: The spectrum of infections in patients with acute and malignant lymphoma in a military hospital. Milit. Med. 137:199, Schimpff, S., Satterlee, W., Young, V. M., and Serpick, A.: Empiric therapy with carbenicillin and gentamicin for febrile patients with cancer and granulocytopenia. N. Engl. J. Med. 284: , Silver,. T., Utx, J. P., Frei, E., and Mc- Cullough, N. B.: Fever infection and host resistance in acute leukemia. Am. J. Med. 24:25-39, Umsawasdi, T., Middleman, E. A., Luna, M., and Bodey, G. P.: bacteremia in cancer patients. Am. J. Med. Sci. (In Press). 14. Whitecar. 1. P.. Tr.. Bodev. G. P.. and Luna. M.: Pseudomonas bicterexka in cihcer patients. Am. J. Med. Sci &223, 1970.