Introduction

In addition to lymphoma itself, comorbidities and higher complication rates during the period of therapy contribute to the poor prognosis of elderly patients with non-Hodgkin’s lymphoma (NHL), especially those with poor performance status and age greater than 80 years [1–3] at the time of diagnosis. Infection is among the more frequent complications resulting in death [1, 2].

Rituximab is a chimeric monoclonal antibody directed against CD20 and is now part of standard treatment for CD20-expressing B cell NHL. The addition of first line rituximab to the standard cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) regimen (R-CHOP) in elderly (60–80 years old) patients with diffuse large B cell lymphoma (DLBCL) has increased response rate, disease-free survival, and overall survival [4]. However, no increase in infection rate was noted [4]. Furthermore, several earlier reports, including ours, showed that high-intermediate risk (international prognostic index score ≥2) elderly patients (median age, 75 years old) with aggressive lymphoma appeared to benefit from the anthracycline-based regimen [1, 2]. However, the benefit of adding rituximab treatment in very old DLBCL patients remains controversial. Indeed, patients over 80 are usually not included in trials. Italiano A et al. showed that addition of rituximab to reduced-dose CHOP offers a good compromise between toxicity and efficacy for DLBCL patients older than 80 years [5].

In our study, the DLBCL patients were older and had more advanced disease status (more bone marrow involvement) than in other similar studies [1, 4]. In the setting and patient population of our institution (a tertiary care veteran’s hospital), it was not known whether adding rituximab to standard CHOP chemotherapy is a risk for infection. Our aim was to investigate the incidence of infection (including bacterial, viral, and fungal infection) in elderly DLBCL patients treated with R-CHOP or CHOP and to identify the risk factors for infection.

Material and methods

Patients and study design

Our retrospective, case-control matched study was conducted from July 2002 to July 2005 and included 34 newly diagnosed diffuse large B cell lymphoma patients who were treated with rituximab plus standard CHOP [cyclophosphamide (750 mg/m2), doxorubicin (50 mg/m2), vincristine (1.4 mg/m2, maximal dose of 2 mg), and prednisone (40 mg/m2)] [4] at Taipei Veterans General Hospital. Pathologic diagnosis was confirmed by at least two pathologists, and the histopathologic classification of tumors was based on the 1994 Revised European–American Lymphoma (REAL) classification system. Patients had to be aged >60 years old and have an Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2. But they were not eligible if they had T-cell lymphoma, a history of indolent lymphoma, brain involvement, or active concomitant cancer or serious disease. Patients were also excluded if they had a cardiac contraindication to anthracycline therapy, a neurologic contraindication, e.g., peripheral neuropathy, or a positive serologic test for human immunodeficiency virus or unresolved hepatitis B virus infection.

Rituximab was approved by the Department of Health in Taiwan in January 2002. Our Bureau of National Health Insurance has paid for rituximab treatment since January 2004. Thus, the 35 control patients treated with standard CHOP were chosen between July 2002 and January 2004. Case patients and controls were matched for age, performance status, disease stage, international prognostic index score, and duration of surveillance. The duration of surveillance was defined as the period between the first chemotherapy session and 6 months after the last chemotherapy session. Clinical charts and microbiological data of all patients and controls were reviewed throughout the surveillance period. The control patients with early lethal infections were not excluded in this study. All patients had central venous catheter (Port-a-Cath, Bard Systems, Covington, GA, USA) inserted for chemotherapy.

Clinical evaluation and dose modification

Evaluation included a complete history and physical examination, determination of performance status, complete blood cell count, and serum biochemistries. Computed tomography scan was performed for evaluation of the measurable metastatic lesions. Other examinations were performed only in the presence of a clinical indication. Laboratory tests were repeated before the start of each cycle. Tumor responses were assessed every three cycles of chemotherapy or at the end of treatment and were classified as complete response, partial response, stable disease, or progressive disease according to NCI Sponsored International Workshop criteria [6]. Toxicities were evaluated according to the National Cancer Institute-Common Toxicity Criteria (NCI-CTC), version 3.0. For patients with grade 4 neutropenia or grade 3 or 4 thrombocytopenia, the doses of cyclophosphamide and doxorubicin were decreased by 50%. If the neutrophil count was <1500/mm3 or platelet count <10,000/mm3 before the scheduled chemotherapy, the chemotherapy treatment was stopped and delayed for up to 2 weeks. The doses of rituximab were not modified, but rituximab was discontinued when chemotherapy was stopped. Treatment was stopped if disease progressed or the patient declined to continue.

Supportive care

Pneumocystis jirovecii pneumonia prophylaxis consisted of trimethoprim, 160 mg and sulfamethoxazole, 800 mg orally thrice per week until the completion of chemotherapy. Oral prophylactic antifungal drugs including nonabsorbable antifungal drugs and antibiotics were not used while patients were neutropenic but afebrile. Prophylactic granulocyte-colony stimulation factor (G-CSF) was given subcutaneously at a dose of 5 μg/kg/d after chemotherapy until the absolute neutrophil count (ANC) was >1,000/μl for three consecutive days. Guidelines for the use of antibiotics in neutropenic fever in Taiwan [7] are similar to Infectious Diseases Society of America (IDSA) guidelines and based on local epidemiology and susceptibility patterns.

Definition of infection

A fever episode was defined as a body temperature above 38.3°C once or above 38°C on three consecutive occasions. An infection episode was defined by two or more of the following clinical symptoms or signs [8]: a fever episode or body temperature <36°C, elevated heart rate above 90 beats/min, tachypnea (respiratory rate >20 breaths/min) or hypoventilation (partial pressure of CO2 [PaCO2]<32 mmHg), white blood count (WBC) of >12,000/μl or <500/μl, and presence >10% immature neutrophils (bands) [8]. Infections leading to sepsis and multiorgan failure, adult respiratory distress syndrome, and death were also considered as one infection episode. Complications due to bacterial or viral infection were defined by both the occurrence of an infection episode and the results of laboratory culture (blood, sputum, wound, fluid, and tissue). The definition of fungal infection is an infection of the bloodstream and/or organs of the body (skin or oral lesion) caused by the presence of fungus (skin or tissue biopsy) or at last two isolated episodes of urine fungus culture. Pulmonary fungal infections were diagnosed by clinical symptoms, CT scan of chest, sputum cultures, and response to antifungal treatment. The positive cultures of Port-a cath mean that patients had positive cultures of blood taken through the catheters and/or positive cultures of catheters but negative cultures of blood taken from peripheral veins.

Statistical analysis

Associations between categorical variables were examined using the Chi-square or Fisher exact test. Univariate and multivariate logistic regression analyses were performed to assess risk factors for the development of infection, including age, bone marrow involvement, response rate, age-adjusted international prognostic index score, previous administration of broad-spectrum antibiotics >1 week, and neutropenia (ANC < 500/μl) >3 days. Most of these factors were selected from previous reports [9–11]. The risk factors of infections with p values<0.05 in the univariate analysis were placed in a multivariate analysis. All analyses were performed using the SPSS 12.0 software package for Windows. Statistical significance was defined as P < 0.05.

Results

Patient characteristics

Patient characteristics are listed in Table 1. There were no statistically significant differences in age, performance status, stage, bone marrow involvement, frequency of elevated lactate dehydrogenase (LDH), bulky mass >10 cm, age-adjusted international prognostic index score, and response rate distribution between the case and control groups. The median age was 76 years old, and the percentage of patients with bone marrow involvement was 52.9 and 48.6% in the case and control groups, respectively. The complete response rate was 55.8 and 45.7% in the case and control groups, respectively.

Table 1 Patient characteristics
Full size table

Infection during and after chemotherapy

A between-group comparison of infection complications is presented in Table 2. The percentage of fever episodes and one or more infectious episodes was 78.3 and 69.6%, respectively, in all 69 elderly DLBCL patients. Most elderly patients had met the criteria of fever episodes at the time of infection. Only five patients (two skin fungal infections and three viral infections) did not meet the definition of fever episode at the time of infection. There was no statistically significant between-group differences in the percentage of patients with <1 fever episodes, <1 infectious episodes, <2 infectious episodes, <1 bacterial infections, <1 viral infections, and infection-related deaths. The bacterial infection rate was 70.5% (for cases) and 57.1% (for controls). However, the fungal infection rate was significantly different (41.7 and 17.1%; P = 0.03), respectively, in the R-CHOP and CHOP groups. Most identified fungal organisms were Candida spp. (18 cases), which were isolated 2–8 months after the start of chemotherapy. Most patients (16 cases, 80%) developed fungal infections during chemotherapy. Only four patients (20%) developed fungal infections after six cycles of chemotherapy. Candida albicans was isolated in 14 cases. Candida tropicalis, Candida krusei, and Candida parapsilosis were isolated in 2, 1, and 1 case, respectively. The other fungal organisms were not further characterized. The Candida spp. were cultured from urine (four cases), blood (four cases), Port-a-Cath catheters (four cases), sputum (three cases), skin (three cases), and oropharynx (two cases) (Table 3). Amphotericin B (0.5–1.5 mg/kg per day) was given to six patients and fluconazole (400 mg per day) to 14 patients. Two patients died of invasive fungal infections (Candida tropicalis, which was isolated from the bloodstream) in rituximab plus CHOP group.

Table 2 Comparison of infectious complications in patients treated with R-CHOP or CHOP
Full size table
Table 3 Comparison of fungal infection in patients treated with R-CHOP or CHOP
Full size table

Factors affecting fungal infections

Rituximab plus CHOP (P = 0.03), age greater than 80 years old (P = 0.04), and bone marrow involvement (P = 0.04) were risk factors for development of fungal infection by univariate logistic regression analyses. Age-adjusted international prognostic index score >2, complete response rate, previous administration of broad-spectrum antibiotic >1 week, and neutropenia (ANC < 500/μl) >3 days were not significant risks (Table 4). In multivariate logistic regression models, only rituximab plus CHOP (P = 0.041) and age greater than 80 years (P = 0.046) were risk factors for development of fungal infections.

Discussion

Our retrospective study showed a higher risk of fungal infection in elderly DLBCL patients, especially those older than 80 years, who were treated with rituximab plus CHOP. Although the overall infection rates associated with R-CHOP in our study and a previous R-CHOP study [4] were similar, the fungal infection rate in our R-CHOP recipients was higher. The difference might be due to differences in age distribution between the two studies. In our study, the median age was older (76 years old vs 69 years old) [4] and disease status was more advanced (the percentage of bone marrow involvement was 52.9 vs 28%) [4]. In this setting, adding rituximab to the CHOP regimen might cause more severe fungal infections and increase the fungal infection rate.

Clinically, several reports also indicated that old age, prolonged leucopenia, previous administration of broad-spectrum antibiotics, corticosteroids and indwelling central venous catheters in cancer patients could increase the risk of opportunistic fungal infections [9, 10, 12, 13]. In our study, all patients received the same steroid-contained chemotherapy regimen and had the same Port-a-Cath types/placement before chemotherapy. Most patients (16 cases, 80%) developed fungal infections during chemotherapy. The incidence of fungal infections did not increase with Port-a-Cath placement duration.

Most opportunistic fungal infections occur in individuals with defective innate, cellular and humoral immune responses, which are considered to be the principal defenses against fungal infections in humans [14]. With age, the function of sweat glands decreases, the skin thins, its ambient moisture content and integrity decline, and the immune system is less responsive [15]. Moreover, the rate of opportunistic fungal infections is increased in patients undergoing immunosuppressive treatment [16, 17]. Defective innate immunity in old age might explain the role of old age ≥80 years in the development of fungal infections observed in our study. Clinically, rituximab increases susceptibility to infection by a variety of opportunistic organisms including virus and fungus pathogens [18]. The basis for rituximab’s contribution to infection risk might be its ability to interfere with humoral immunity (i.e., lower serum immunoglobulin G [IgG] and M [IgM] levels) [19]. Humoral immune deficiencies increase the risk for opportunistic fungal infection. Several clinical and experimental findings point to the critical role of antigen-specific cellular immunity as good in host protection against fungi [20–23]. Nonetheless, recent data strongly support the existence of protective antibodies against Candida and Cryptococcus species, two of the major opportunistic fungal infections [23]. Similar to previous reports, our study found Candida spp. were the main fungal organisms isolated from urine, blood, and Port-a-Cath catheters [9–11]. In rituximab plus CHOP group, most Candida spp. were cultured from blood, Port-a-Cath, and skin.

Bishop JF et al. showed that bone marrow involvement is not a risk factor for the development of infection during intensive NHL chemotherapy [24]. Bone marrow involvement was a risk factor for infection by univariate but not multivariate logistic regression analyses in our study Table 4.

Table 4 Univariate analysis of risk factors for fungal infection
Full size table

In conclusion, adding rituximab to the standard CHOP regimen in elderly lymphoma patients might increase the incidence of fungal infection, especially in patients older than 80 years. Administration of antifungal drugs (although their usefulness has not been proven) should be considered before clinical symptoms appear.