Preventing and Managing Infections in Patients Receiving Nucleoside Analogs

Introduction Nucleoside analogs have marked efficacy in indolent lymphoid malignancies, but the tradeoff is the challenge of preventing and treating infections in these patients, according to Susan OBrien, MD, of M. D. Anderson Cancer Center, Houston. At the American Society of Hematology meeting last December, Dr. OBrien discussed strategies for preventing and treating infections in patients receiving these agents. I would go so far as to say that these drugs have revolutionized the treatment of some of these diseases, including hairy cell leukemia, chronic lymphocytic leukemia, and low-grade lymphoma, said Dr. OBrien. They are also attractive to patients because they have less extramedullary toxicity than other chemotherapy agents. However, there can be major complications related to myelosuppression, immunosuppression, and infection. Im not sure whether the incidence of infections really is increased with nucleoside analogs or whether were simply collecting more data and are more aware of them now, she added, noting the dearth of clinical data on how to deal with them. The highest rate of complications, particularly infection, occurs among the chronic lymphocytic leukemia patients for a number of disease-related reasons. These include hypogammaglobulinemia, decreased cell-mediated immunity, and granulocytopenia (because of heavy involvement of the marrow with leukemia or because of complications of prior therapy). With disease progression and after multiple treatments, the risk for complications grows. This is the group for whom prophylaxis should be targeted, but prophylaxis is a sticky concept, she noted. You will always be treating a number of patients who dont necessarily need treatment in order to benefit the group that will do better with prophylaxis. Targeting High-Risk Groups for Prophylaxis Dr. OBrien advised making a decision for or against prophylaxis based on the following factors: (1) incidence of infection, (2) potential severity of infection, (3) ease of administering prophylaxis, (4) potential side effects of prophylaxis, and (5) cost. There is a need to identify high-risk groups to target for trials of prophylaxis, she said. One such study, recently published by the Leukemia Group at M. D. Anderson,[1] was a retrospective evaluation of more than 400 patients, both previously treated and untreated, who were receiving fludarabine (Fludara)-based therapy, either alone or in combination with prednisone. There was no age restriction, but normal organ function was required for inclusion in this investigation. Major infections in this retrospective analysis were defined as pneumonia, bacteremia, sepsis, meningitis, or fever requiring antibiotics (mostly neutropenic fevers). In the entire cohort of 400, essentially half the patients developed at least one significant infection during therapy with fludarabine. This was fairly evenly divided between fevers of unknown origin, which were usually neutropenic fevers requiring hospitalization, and pneumonia or sepsis (equally gram positive and gram negative). These patients did not have indwelling catheters, but the vast majority of the gram-negative infections were Pseudomonas. Atypical infections accounted for only 5% of the infections that occurred with fludarabine-based therapy, she reported. Most of the infections occurred early on, so that people who got through the first two or three courses generally had fewer problems with subsequent courses, presumably because they were benefitting from a response to treatment, she observed. By univariate analysis, the risk factors that predicted for infection were Rai stage, prior therapy, and baseline neutropenia. Levels of albumin and creatinine were also important (note that the trials did not include patients with renal failure), as was level of beta2-microglobulin; this parameter was not included in the multivariate analysis because data were missing in a significant fraction of patients. There was absolutely no difference in the incidence of infections based on treatment with fludarabine alone or with prednisone. The only difference that emerged was an increase in atypical infections, particularly Listeria and Pneumocystis, in patients who received the combination. Age was not a predictive factor for infection, although the study included few patients over age 75 years, who might have had a higher risk of infection, nor were CD4 counts at baseline a factor. However, there was a higher incidence of herpes zoster reactivation in patients with severe depletion of CD4 cells after therapy. In the multivariate analysis, three factors were again independently predictive of infections: prior therapy and Rai stage, as in the univariate analysis and creatinine level. The creatinine is interesting because there were not many patients who had a high creatinine level (> 1.4 mg/dL, the upper limit of normal at M. D. Anderson). My hypothesis would be that in these patients, who for the most part are older, even mild increases in creatinine probably represent very marked decreases in creatinine clearance. Fludarabine is excreted largely via the kidneys, so what I would propose is that in patients with mild renal insufficiency, were in fact giving them what is, in effect, much higher doses of fludarabine, and this is translating into more myelosuppression and infection. Dr. OBrien recommended reducing the dose of fludarabine in patients with even mild increases in creatinine by giving it for 3 days rather than 5. M. D. Anderson investigators have found that 3-day therapy maintains efficacy while reducing the incidence of infections, she said. Potential Approaches to Minimize Infections Dr. OBrien then made recommendations, based on available data, regarding the four potential approaches to minimize infection: (1) prophylactic antibiotics, (2) growth factors, (3) intravenous immunoglobulin (IVIG), and (4) immune stimulation. There are no prospective data for the use of prophylactic antibiotics with nucleoside analogs in chronic lymphocytic leukemia (CLL) patients, she said. There are, however, retrospective data from one study and a randomized trial in multiple myeloma. (While nucleoside analogs are not effective in myeloma, it is reasonable to extrapolate from this disease in terms of risk for infection during therapy, she said). The multicenter study evaluated the activity of fludarabine in patients with CLL and found that about half had received prophylaxis with co-trimoxazole (trimethoprim/sulfamethoxazole), but this measure did not protect them from hospitalization. Its important to point out that there are no data on the comparability of these patients in this retrospective evaluation, she added. In fact, one could postulate that the patients who had received prophylaxis were considered at higher risk for infection and would have had a higher rate of hospitalization without prophylaxis; this would mean that the outcome was more positive in this group of patients than it might seem, she suggested. Another trial in myeloma patients starting chemotherapy with melphalan (Alkeran) and prednisone, randomized patients to receive trimethoprim/sulfamethoxazole twice daily for 2 months or no prophylaxis at all. During the 2 months on prophylaxis and the following month, the study assessed the rate of infections and looked for a potential rebound in infections in the prophylaxis group during the third month. There was a dramatic reduction in bacterial infections during months 1 and 2 and some degree of benefit as well in the third month, when the patients were off prophylaxis; there was even a trend for a decreased mortality rate in the prophylaxis group. This is an interesting and provocative trial, and it would be reasonable to conduct this kind of trial in high-risk patients with CLL, she commented. Based on very limited date, Dr. OBrien said she would make the following recommendations: (1) When steroids are warranted (such as in autoimmune complications), Pneumocystis carinii pneumonia (PCP) prophylaxis is mandatory. (2) It is reasonable to use prophylaxis in patients who have had a documented infection early in their treatment course and are going to require further therapy. (3) Based on experience in other disease states, it may be reasonable to use prophylaxis during neutropenia, but data are lacking in the setting of nucleoside analog use. Herpes Zoster: A Thorny Issue Herpes zoster represents another thorny issue in patients with CLL receiving nucleoside analogs, she said. These infections are rarely life-threatening, but they carry significant morbidity in the form of postherpetic neuralgia, which increases in frequency and severity with age. In an article in The New England Journal of Medicine,[2] the authors report that more than half of patients over the age of 70 years developed postherpetic neuralgia lasting for over a year. The problem is how to predict which patients will develop herpes zoster, she noted. The only clear predictive factor is based on very limited data from fewer than 150 patients at M. D. Anderson, again suggesting that very, very low CD4 counts (< 50/µL) may predict for reactivation, she said. Outside of depleted CD4 counts, there are no other means of identifying a high-risk group. Unlike bacterial infections, herpes zoster infections may not occur during therapy but months after completion, raising the question of duration of prophylaxis. Based on very limited data, it would be reasonable to treat prophylactically patients with CD4 counts that are severely depleted or patients with a prior episode of zoster, Dr. OBrien suggested. Prophylaxis With Growth Factors Historical control data suggests that with single-agent fludarabine, neutropenia, infection, and therapy delays are common in patients with CLL who have been previously treated and have advanced-stage disease. Therefore, an M. D. Anderson trial was designed to give patients a routine dose of fludarabine (30 mg/m²/d) for 5 days, then G-CSF (Neupogen) at the standard dose subcutaneously starting on day 6, with dose adjustments based on neutrophil counts. The study found that neutrophil counts less than 1,000 mg/µL were seen much less frequently with the use of G-CSF, and severe neutropenia (< 500 mg/µL) was almost completely abrogated. This was an interesting observation because these were advanced-stage patients with heavily involved bone marrows, yet even after the first chemotherapy cycle, there was a very brisk response to G-CSF. When compared to data from historical controls, this translated into a benefit in terms of a dramatic reduction in pneumonia, and a trend for a reduction in fever of unknown origin or neutropenic fever, which was reduced by almost half by the addition of G-CSF. I think that this is a lead that needs to be explored further. While there were limitations to this small study, these preliminary data suggest that growth factors may be very useful in CLL, she concluded. Immunoglobulin The data are clear that the use of IVIG prevents bacterial infections in some patients, but there are a number of controversial issues. The original randomized trial of patients with low levels of IgG or a history of infection showed a significant reduction in both bacterial infections and pneumonia; however, subsequent analysis found these benefits not to be cost-effective. A more recent trial used a dose of 18 g (roughly half the original dose) in patients who had both hypogammaglobulinemia and a history of two or more recent severe infections. Even in this higher-risk group, lower doses of IVIG produced a significant benefit in terms of reduced sepsis, pneumonia, and number of infections. In addition, several crossover trials have been conducted in which patients served as their own controls, randomized to receive either IVIG or no prophylaxis for the first 12 months, then crossed over to the opposite treatment for the second 12 months. Again, there was a significant reduction in bacterial infections, including patients on lower doses, she reported. But there are issues with IVIG treatmentspecifically, cost, scarcity of the product, and adequate dose (which has not been established). What we have been doing empirically is treating patients with the standard dose from The New England Journal article (400 mg/kg). If they have low IgG levels and recurrent bacterial sinopulmonary infectionswhere they are constantly on one antibiotic after another without resolution of their symptomswe have found they can benefit significantly from IVIG, she observed. Stimulating the Immune System There are even fewer data on the benefit of boosting the immune system as a method of prophylaxis in patients with CLL. Several European trials are assessing this strategy, and there is some indication that interleukin-2 (Proleukin) can boost CD4 and CD8 counts significantly. This is an interesting strategy that deserves further pursuit, she said. In summary, Dr. OBrien commented on the paucity of controlled trials evaluating infection prophylaxis in CLL patients receiving nucleoside analogs, but made some recommendations. She said steroids clearly should be avoided with these agents but, if they are used, PCP prophylaxis is essential. Useful preventive strategies include growth factors in high-risk patients, antibiotics during neutropenia or for prior serious infection on treatment, herpes zoster prophylaxis if CD4 counts are markedly depleted, and IVIG for patients with very low IgG levels and recurrent sino-pulmonary infections. Questions From the Audience Question: You commented on the use of PCP prophylaxis when you need to use steroids with the purine analogs. What is your current recommendation for PCP prophylaxis in that setting? Also, what type of prophylaxis would you recommend against herpes zoster? Dr. OBrien: We use trimethoprim-sulfamethoxazole twice a week for PCP, just like in the transplant setting. For zoster, we use valacyclovir (Valtrex), once a day. References: 1. Anaissie EJ, Kontoyiannis DP, OBrien S, et al: Infections in patients with chronic lymphocytic leukemia treated with fludarabine. Ann Intern Med 129:559-566, 1998. 2. Gilden DH, Kleinschmidt-DeMasters BK, LaGuardia JJ, et al: Neurologic complications of the reactivation of varicella-zoster virus. N Engl J Med 342(9):635-645, 2000.