Hepatosplenic Candidiasis in Patients with Acute Leukemia: Incidence and Prognostic Implications

Transcription

1 375 Hepatosplenic Candidiasis in Patients with Acute Leukemia: Incidence and Prognostic Implications Veli-Jukka Anttila, Erkki Elonen, Stig Nordling, From the Department of Medicine, Divisions of Hematology and Aulikki Sivonen, Tapani Ruutu, and Petri Ruutu Infectious Diseases, Helsinki University Central Hospital, and the Departments of Pathology and Bacteriology and Immunology, University of Helsinki, Helsinki, Finland We studied the incidence among, risk factors for, and survival of adult patients with acute leukemia and hepatosplenic during the period 198 to Of 562 adult patients with acute leukemia, 38 (6.8%) had hepatosplenic. The incidence of infection increased fivefold during the study period. The incidence was higher among patients with acute lymphatic leukemia (11.3%) than among those with acute myeloid leukemia (5.1%) (P =.1). The median survival was 9.5 months, and by the end of follow-up, 74% of patients had died. Patients whose leukemia was in remission before the last cytotoxic treatment preceding hepatosplenic survived longer than did patients with newly diagnosed or refractory or relapsed leukemia (P =.65). Eleven patients died within 3 months after the diagnosis of the infection: 7 of 16 with newly diagnosed leukemia, 4 of 1 with refractory or relapsed leukemia, and of 12 with leukemia in remission (P =.28). In all of the patients who died within 3 months, infection was found at autopsy. In conclusion, the incidence of hepatosplenic has significantly increased since 198, and the outcome for patients with this infection is related to the stage of leukemia. Hepatosplenic was recognized as a distinct form of invasive in patients with acute leukemia in the early 198s [1]. The incidence of this form of seems to have increased [2-5]. Although hepatosplenic is recognized as a major problem in the treatment of acute leukemia, there are little data on its incidence among patients with leukemia [4]. This paucity of information may be due to difficulties in confirming the specific diagnosis [6]. The mortality due to hepatosplenic in patients with leukemia is significant [1, 2, 7, 8]. In most published reports, diagnoses have been based on autopsy findings or there were only a few cases [1, 2, 5, 7, 9, 1]. Little is known about the impact of this infection on patient survival. In a tertiary care hospital, we investigated the incidence among, risk factors for, and survival of adult patients with acute leukemia and hepatosplenic over a period of 14 years. Patients and Methods During the period , 562 patients with acute leukemia were treated at the Department of Medicine, Helsinki University Central Hospital, Helsinki. Reports on fungal cultures, pathological examinations of specimens, and autopsies of Received 7 February 1996; revised 4 September Reprints or correspondence: Dr. Veli-Jukka Anttila, Department of Medicine, Division of Hematology, Helsinki University Central Hospital, Haartmaninkatu 4, FIN-29, Helsinki, Finland. Clinical Infectious Diseases 1997; 24: by The University of Chicago. All rights reserved /97/243 12$2. patients with acute leukemia were screened to find patients with hepatosplenic. Treatment of Acute Myeloid Leukemia (AML) The intensity of induction therapy was similar throughout the whole study period [11, 12]. The treatment cycles consisted of cytarabine and thioguanine for 7 to 1 days and daunorubicin or idarubicin for 3 days. Cycles of consolidation therapy were of low intensity until 1984, and they seldom induced severe long-lasting neutropenia [11]. In later protocols, postremission therapeutic courses were at least as intensive as induction therapy, and all of them induced severe neutropenia for 2 to 4 weeks [12]. Treatment of Acute Lymphatic Leukemia (ALL) Until June 1986, induction therapy consisted of weekly administrations of doxorubicin and vincristine, prednisone for 4 weeks, and asparaginase for 1 days [13]. In 1984, cyclophosphamide was added to this therapy [14]. The intensity of consolidation and maintenance treatments was low. In 1986, more intensive induction and consolidation therapies were commenced [15]. High doses of cytarabine, etoposide, and teniposide were introduced, the infusion time of methotrexate was increased from 4 to 24 hours, and high doses of dexamethasone were given instead of prednisone. In the protocol used during the years , therapies induced severe stomatitis and gastrointestinal toxic effects [15].

2 376 Anttila et al. CID 1997;24 (March) Prophylactic Antimicrobial Regimens No systemic or topical antibacterial or antifungal treatment was used as prophylaxis during the study period. Definition of Hepatosplenic Candidiasis A patient was considered to have hepatosplenic if (1) focal lesions in the liver, spleen, or kidneys were found by imaging methods or at autopsy and (2) a specific diagnosis of yeast infection was established either by the culture of blood or a deep-tissue specimen or by appropriate findings of microscopy of a deep-tissue specimen that was stained according to standard laboratory staining methods. Patients for whom Candida was isolated from blood but who had no evidence of deep-organ involvement were not included in the study. Imaging Methods CT and ultrasonography were used for imaging of visceral organs throughout the study period. Since December 199, MRI was used systematically for imaging of livers of all patients for whom hepatic was clinically suspected [16]. Statistical Analysis The x2 test was used for comparison of differences in rates. An estimation of the probability of survival was made according to the Kaplan-Meier method. Comparisons of the survival between patient groups were assessed with logrank statistics with use of BMDP statistical software (BMDP, West Los Angeles). P values of <.5 were regarded as statistically significant. Results Of the 562 patients with leukemia (411 with AML and 151 with ALL) who were treated during the 14-year study period, 38 (6.8%) had hepatosplenic. The mean age of these patients was 46 years (range, years). There were 2 males and 18 females. Twenty-one patients had AML, and 17 had ALL. The states of the leukemias before the last cytotoxic treatment preceding the diagnosis of hepatosplenic were as follows: newly diagnosed disease, 16 patients; remission, 12; and refractory or relapsed disease, 1. The median duration of the last episode of neutropenia (defined as a blood neutrophil count of X 1 9/L) before the diagnosis of hepatosplenic was 19 days (range, 8-45 days). Focal lesions were seen in the following visceral organs: liver alone, 21 patients; spleen alone, 1; spleen and liver, 6; liver, kidneys, and spleen, 4; liver and kidneys, 2; and spleen and kidneys, 3. One patient had a combination of an exceedingly high alkaline phosphatase level and focal lesions in the kidneys and pancreas. In 13 of these patients, lesions also were seen outside the abdominal or retroperitoneal cavities. Cultures of specimens from a total of 17 patients (45%; blood specimens, six patients; deep-tissue samples, 11 patients) yielded Candida (table 1). In addition, yeast structures were detected by microscopic examination of visceral-organ samples from 14 of these 17 patients. In a further 21 cases, the specific diagnosis was based on microscopic findings of yeast structures in visceral-organ samples. For 11 (29%) of 38 patients, the diagnosis of hepatosplenic was made only at autopsy; cultures of blood specimens from three of these patients yielded Candida shortly before death. Before the detection of hepatosplenic, 35 patients received systemic empirical antifungal treatment. Only one of the 38 patients did not receive systemic antifungal treatment either before or after the diagnosis of the infection (table 1). The mean total dose of amphotericin B was 1,654 mg (range, -7,23 mg). Three patients did not receive treatment with amphotericin B: two received miconazole, and one received ketoconazole followed by flucytosine. Legionella antigen was detected in bronchoalveolar lavage fluid from a patient with pneumonia who did not receive antifungal treatment, but hepatosplenic was confirmed at autopsy. The incidence of hepatosplenic in patients with acute leukemia increased significantly: between 198 and 1984, there were 4 cases (1.9%) in 213 patients; between 1985 and 1988, 17 cases (9.9%) in 171 patients; and between 1989 and 1993, 17 cases (9.6%) in 178 patients (P =.2; x2 test). The annual incidence of hepatosplenic is shown in table 2. The incidence of hepatosplenic was significantly higher among patients with ALL than among those with AML (11.3% vs. 5.1%, respectively) (P =.1; x2 test). The incidence of hepatosplenic increased significantly among patients with both types of acute leukemia: between 198 and 1984, 3 cases (1.8%) in 165 patients with AML and 1 case (2.1%) in 48 patients with ALL; between 1985 and 1988, 11 cases (8.8%) in 125 patients with AML and 6 cases (13%) in 46 patients with ALL; and between 1989 and 1993, 7 cases (5.8%) in 121 patients with AML and 1 cases (17.5%) in 57 patients with ALL (P =.3 for AML and P =.4 for ALL; x2 test). By April 1995, of the 38 patients with hepatosplenic, 28 (74%) had died, and 1 (26%) were alive. The median survival from diagnosis of yeast infection for all patients was 9.5 months (figure 1A). Logrank analysis showed that those patients whose leukemia was in remission when hepatosplenic manifested survived longer than did those with newly diagnosed or refractory or relapsed leukemia (P =.65) (figure 1B). Logrank analysis of sex, type of leukemia (AML vs. ALL), age (5 years of age or younger vs. older than 5 years of age), duration of last episode of neutropenia (-2 days vs. >2 days), and year of infection ( vs.

3 CID 1997;24 (March) Hepatosplenic Candidiasis in Acute Leukemia 377 Table 1. Characteristics of 38 patients with acute leukemia and hepatosplenic. Antifungal therapy Dose (mg) of AmB Duration (d) before detection of before detection of Total dose Patient Type of leukemia, of Culture results hepatosplenic hepatosplenic (mg) of Clinical status, Survival no. disease status neutropenia (specimen) AmB cause(s) of death (mo) 1 ALL, refractory/relapsed 24 Candida albicans Mic, 5-FC 2,555 Dead, leukemia 12 2 AML, refractory/relapsed 18 C. albicans Ket, 5-FC Dead, hepatosplenic.2 3 AML, newly diagnosed 45 Nonidentified Candida AmB Dead, hepatosplenic.4 4 AML, newly diagnosed 24 C. albicans Mic Dead, hepatosplenic 1 5 ALL, newly diagnosed 43 Negative AmB Dead, leukemia 7 6 ALL, refactory/relapsed 31 Nonidentified Candida AmB 895 1,285 Dead, leukemia 51 7 AML, refractory/relapsed 26 Negative AmB, 5-FC 49 8 Dead, leukemia 2 8 ALL, refractory/relapsed 27 Negative AmB Dead, hepatosplenic.4 9 ALL, remission 9 Negative AmB, 5-FC 15 1,96 Dead, leukemia 4 1 AML, newly diagnosed 34 C. albicans AmB, 5-FC, Mic Dead, hepatosplenic.2 11 AML, remission 31 Negative AmB 745 3,15 Dead, leukemia and 4 hepatosplenic 12 AML, remission 39 Negative AmB, Mic 1, 2,4 Alive, remission >9 13 AML, remission 16 Negative AmB, Mic Dead, leukemia AML, remission 18 Negative AmB, 5-FC 615 1,66 Dead, leukemia ALL, refractory/relapsed 28 Candida parapsilosis AmB, 5-FC 9 1,15 Dead, hepatosplenic.4 16 AML, newly diagnosed 29 Negative 5-FC 3,29 Dead, leukemia AML, newly diagnosed 13 Candida krusei AmB, 5-FC, Mic 72 4,625 Dead, leukemia 9 18 AML, newly diagnosed 15 C. albicans Mic Dead, hepatosplenic.2 19 AML, remission 12 C. parapsilosis 5-FC 135 Alive, remission >75 2 AML, newly diagnosed 24 Nonidentified Candida AmB, 5-FC 86 2,785 Alive, remission >65 21 ALL, newly diagnosed 31 Negative AmB, 5-FC 1,175 3,785 Dead, pneumonia 8 after BMT 22 ALL, newly diagnosed 19 Negative AmB, Mic 765 4,3 Dead, leukemia and 6 hepatosplenic 23 AML, newly diagnosed 14 Candida tropicalis AmB 8 8 Dead, hepatosplenic.2 24 ALL, remission 1 Negative AmB 24 7,32 Dead, leukemia AML, remission 15 Negative AmB, Mic 15 1,35 Alive, remission >69 26 ALL, newly diagnosed 21 Negative AmB 75 1,795 Dead, leukemia AML, newly diagnosed 26 C. tropicalis AmB, Flu 8 11 Alive, remission >61 28 ALL, newly diagnosed 17 Torulopsis glabrata None Dead, hepatosplenic 29 AML, newly diagnosed 14 Negative AmB 81 3,3 Alive, remission >37 3 AML, refractory/relapsed 35 Negative AmB, 5-FC 75 1,65 Dead, leukemia 9 31 ALL, remission 15 C. albicans AmB, Flu 17 1,12 Alive, remission >46 32 ALL, refractory/relapsed 17 C. albicans AmB Dead, hepatosplenic 33 AML, refractory/relapsed 16 Negative AmB 595 2,295 Dead, CMV 1 infection after BMT 34 ALL, newly diagnosed 25 Negative AmB Dead, hepatosplenic.2 35 AML, remission 14 Negative None 6 Alive, remission >38 36 ALL, refractory/relapsed 19 C. albicans (deep AmB 715 3,858 Dead, leukemia 5 tissue) 37 ALL, remission 17 Negative AmB, Itr 32 1,14 Alive, remission >23 38 ALL, remission 8 Negative None 66 Alive, remission > NOTE. ALL = acute lymphatic leukemia; AmB = amphotericin B; AML = acute myeloid leukemia; BMT = bone marrow transplantation; CMV = cytomegalovirus; 5-FC = flucytosine; Flu = fluconazole; Itr = itraconazole; Ket = ketoconazole; Mic = miconazole.

4 378 Anttila et al. CID 1997;24 (March) Table 2. The annual incidence of hepatosplenic in patients with acute leukemia between 198 and Year No. of patients with hepatosplenic and new acute leukemia/total no. with new acute leukemia (%) 198 1/43 (2.3) /47 (2.1) /42 (2.4) 1983 / /39 (2.6) /43 (7.) /52 (3.8) /36 (13.9) /4 (17.5) /4 (1.) 199 2/35 (5.7) /46 (15.2) /34 (5.9) /23 (8.7) Total 38/562 (6.8) ) identified no other variables of prognostic significance. Eleven patients died within 3 months after the diagnosis of hepatosplenic (table 1). In all of these patients, the infection was also evident at autopsy. Rates of short-term mortality due to infection were lower among younger patients (5 years of age or younger) and those whose leukemia was in remission when hepatosplenic manifested than among patients with newly diagnosed or refractory or relapsed A leukemia (table 3). Another 1 patients with hepatosplenic died between >3 and < 12 months after the diagnosis of the infection. At autopsy, infection was still evident in two but not in five of these patients. Autopsies were not performed on three patients with previously diagnosed hepatosplenic. One year after the diagnosis of hepatosplenic, only two of the 12 patients with leukemia in remission had died, whereas 11 of the 16 patients with newly diagnosed leukemia and eight of the 1 patients with refractory or relapsed leukemia had died (P =.3; x2 test). Cytotoxic treatment was given to 23 patients after the diagnosis of hepatosplenic. The median delay of cytotoxic treatment due to the infection, calculated as the time from the detection of hepatosplenic until the next cytotoxic treatment, was 73 days (range, days). None of these patients had reactivation of infection. Twenty of these 23 patients received systemic antifungal treatment during episodes of neutropenia. Three patients later underwent allogeneic bone marrow transplantation and did not have any fungal infections. Discussion Our study of 562 adult patients with acute leukemia indicated a significant increase in the incidence of hepatosplenic between 198 and The mortality rate among patients with acute leukemia and hepatosplenic was high, although only part of the deaths were directly attributable to the infection. nn = Cn C 1.= CL rn Newly diagnosed G 1 B Remission I No. of months ) a I Ref/rel No. of months Figure 1. Kaplan-Meier survival curves for 38 patients with acute leukemia and hepatosplenic. A: all patients. B: patients with leukemia in remission before the last cytotoxic treatment preceding hepatosplenic (12), patients with newly diagnosed leukemia (16), and patients with refractory or relapsed (Ref/rel) leukemia (1).

5 CID 1997;24 (March) Hepatosplenic Candidiasis in Acute Leukemia 379 Table 3. Characteristics of patients with acute leukemia and hepatosplenic : surviving patients vs. patients who died within the first 3 months after the diagnosis of hepatosplenic. Characteristic Surviving patients (n = 27) No. of patients Patients who died (n = 11) P value* Sex Male 14 6 NS Female 13 5 Type of leukemia ALL 12 5 NS AML 15 6 Age 5 y or younger Older than 5 y 9 8 Duration of neutropenia -2 d 15 5 NS >2 d 12 6 Year of diagnosis NS Stage of leukemia Newly diagnosed Remission 12 Refractory or relapsed 6 4 NOTE. ALL = acute lymphatic leukemia; AML = acute myeloid leukemia; NS = not significant. * Determined by the x2 test. Hepatosplenic is difficult to diagnose. Cultures of blood and tissue samples often remain negative [6, 7]. A strong inflammatory reaction appears around the fungal organism after the recovery of neutrophils, and the number of fungal organisms inside the foci may be small. Although imaging methods reveal typical foci in the liver or spleen, in many cases both cultures and microscopic examinations of biopsy specimens remain negative for fungi [16]. The new more-sensitive imaging methods such as MRI can detect foci in the liver of patients for whom hepatosplenic is clinically suspected more frequently than CT or ultrasonography [16]. Ultrasonography often cannot detect foci, thus preventing directed invasive procedures. For some patients, invasive procedures must be avoided because of the risk of bleeding. In this study, we included only patients with microbiologically or microscopically proven hepatosplenic yeast infection. Therefore, the incidence of hepatosplenic in our study can be considered the minimum estimate of the true incidence of hepatosplenic in this patient population. MRI was introduced systematically in 199. Since a significantly increased incidence was already seen during the preceding period of 1985 to 1988, the increase is not explained by the introduction of a more sensitive imaging method. Fungal infections have become more common as cytotoxic treatment of leukemia has become more intensive [4, 5]. At our hospital, the incidence of hepatosplenic increased over fivefold during the study period. We found that the incidence of hepatosplenic in patients with ALL was significantly higher than that in patients with AML. Antileukemic therapy plays an important role in the risk of invasive fungal infection [4]. It seems plausible to suggest that the increased intensity of our treatment protocols for ALL, including high doses of steroid components, could be the major factor influencing the dissemination of Candida [15]. There is little information about the outcome for patients with hepatosplenic [1, 8]. In most previous studies, diagnoses were based on autopsy findings, the study population was not always homogeneous, and the type of fungal infection varied [1-3]. To avoid these confounding factors, we focused only on patients with acute leukemia who were at high risk of contracting this infection. We studied the survival of patients with acute leukemia and hepatosplenic yeast infection between 198 and All patients whose fungal infections were diagnosed while they were alive received systemic antifungal treatment (9% received intravenous amphotericin B). Despite the antifungal treatment, the short-term mortality rate at 3 months after the diagnosis of the infection was almost 3% and was strongly dependent on the stage of leukemia when the infection manifested; none of the patients whose leukemia was in remission died within 3 months after the diagnosis of the infection, whereas almost one-half of the patients with newly diagnosed leukemia and over one-half of the patients with refractory or relapsed leukemia had died. In all 11 patients who died within 3 months after the diagnosis of hepatosplenic, the infection was the major cause of death. The treatment of hepatosplenic often delays the treatment of leukemia, thereby exposing the patients to an increased risk of relapse. For those 23 patients who subsequently received cytotoxic therapy, the median delay of cytotoxic treatment caused by the infection was >2 months. In this retrospective study, it was difficult to assess exactly the effect of hepatosplenic on long-term mortality. It seems that the infection can usually be successfully treated if the leukemia itself is in control, evidenced by the fact that 8% of the patients who had the infection when leukemia was in remission were alive 1 year after the infection was diagnosed. In conclusion, the incidence of hepatosplenic in patients with acute leukemia has significantly increased since the early 198s. The survival of patients with acute leukemia and hepatosplenic is related to the stage of leukemia. The short- and long-term mortality rates among patients whose leukemia was in remission when they contracted the infection were significantly lower than those among other patients. To improve the prognosis of patients with acute leukemia and hepatosplenic, more effective prophylactic and therapeutic antifungal treatments are needed.

6 38 Anttila et al. CID 1997; 24 (March) References 1. DeGregorio MW, Lee WMF, Linker CA, Jacobs RA, Ries CA. Fungal infections in patients with acute leukemia. Am J Med 1982; 73: Thaler M, Pastakia B, Shawker TH, O'Leary T, Pizzo PA. Hepatic in cancer patients: the evolving picture of the syndrome. Ann Intern Med 1988; 18: Bodey G, Bueltmann B, Duguid W, et al. Fungal infections in cancer patients: an international autopsy survey. Eur J Clin Microbiol Infect Dis 1992; 11: Bow EJ, Loewen R, Cheang MS, Schacter B. Invasive fungal disease in adults undergoing remission-induction therapy for acute myeloid leukemia: the pathogenetic role of the antileukemic regimen. Clin Infect Dis 1995;21: Haron E, Feld R, Tuffnell P, Patterson B, Hasselback R, Matlow A. Hepatic : an increasing problem in immunocompromised patients. Am J Med 1987; 83: Anttila V-J, Ruutu P, Bondestam S, et al. Hepatosplenic yeast infection in patients with acute leukemia: a diagnostic problem. Clin Infect Dis 1994; 18: Maksymiuk AW, Thongprasert S, Hopfer R, Luna M, Fainstein V, Bodey GP. Systemic in cancer patients. Am J Med 1984; 77(suppl 4D): Guiot HFL, Fibbe WE, van't Wout JW. Risk factors for fungal infection in patients with malignant hematologic disorders: implications for empirical therapy and prophylaxis. Clin Infect Dis 1994; 18: Loeliger A, van Leeuwen M, Rozenberg-Arska M, Dekker AW. Hepatosplenic, a fatal disease? Infection 1992; 2: Tashjian LS, Abramson JS, Peacock JE Jr. Focal hepatic : a distinct clinical variant of in immunocompromised patients. Rev Infect Dis 1984; 6: Palva IP, Almqvist A, Elonen E, et al. Value of maintenance therapy with chemotherapy or interferon during remission of acute myeloid leukaemia. Eur J Haematol 1991; 47: Elonen E, Almqvist A, Hanninen A, et al. Intensive chemotherapy of adult AML: a randomized trial between 4 and 8 cycles [abstract no 1297]. Blood 1993,82(1)(suppl 1):328a. 13. Lister TA, Whitehouse JMA, Beard MEJ, et al. Combination chemotherapy for acute lymphoblastic leukaemia in adults. Br Med J 1978; 1: Finnish Leukaemia Group. Long-term survival in acute lymphoblastic leukaemia in adults: a prospective study of 51 patients. Eur J Haematol 1992; 48: Elonen E, Almqvist A, Hanninen A, et al. Intensive cyclical chemotherapy in the treatment of adults with acute lymphatic leukaemia [abstract no 244]. In: Proceedings of the 24th International Congress of Haematology. London: International Society of Haematology, 1992: Anttila V-J, Lamminen EA, Bondestam S, et al. Magnetic resonance imaging is superior to computed tomography and ultrasonography in imaging infectious liver foci in acute leukaemia. Eur J Haematol 1996; 56:82-7.