Leukopenia and Severe Infection in Patients With SLE
Leukopenia and Severe Infection in Patients With SLE
Leukopenia is a common manifestation in SLE. This study demonstrated that 50% of patients had leukopenia at the diagnosis of SLE and increased to nearly 60% by the end of the study. The prevalence of leukopenia is comparable to its prevalence in previous studies, 36% to 66%. In this study, no patient had leukopenia with less than 2000/mL. Like most studies, WBC count less than 2000/mL is not common in SLE patients, but 1 study found 14% of untreated SLE patients had WBC count of less than 2000/mL.
Persistent lymphopenia was a common finding in SLE patients, whereas leukopenia and neutropenia were usually transient conditions. In our cohort, only 1 patient had persistent leukopenia without any immunosuppressive agent, and no patient had persistent neutropenia. On the other hand, 41% of patients had persistent lymphopenia. Similarly, in the Coimbra Lupus Cohort, persistent lymphopenia was found in 41%, and only 5% had persistent neutropenia.
In this study, all leukopenia was counted, no matter what the causes of leukopenia were. It should be noted that, in this study, 99% of all patients received corticosteroids, and 74% to 83% of patients who had leukopenia had been exposed to immunosuppressive agents. In fact, corticosteroids cause lymphopenia but neutrophilia, whereas immunosuppressive drugs can cause both lymphopenia and neutropenia. As known, from SLE disease activity, lymphopenia is more common than neutropenia. Therefore, some fraction of lymphopenia and neutropenia in SLE patients may be modified by medication. A previous study demonstrated that higher exposure to immunosuppressive agents and other drugs was an independent risk factor for neutropenia in SLE, (OR, 4.8 and 16.5, respectively).
Severe infection is common in SLE patients. In this study, the incidence of a severe infection rate as defined primarily by need for intravenous antibiotic therapy was 12.4 cases per 100 patient-years. This incidence was comparable to the study from the New York and Spanish populations, 13.9 and 17.0 cases per 100 patient-years, respectively. However, our severe infection rate was much higher than was observed in the Canadian and Swedish populations, 6.9 cases per 100 patient-years. Severe infections were responsible for 37% of admissions in this cohort and caused longer hospital stays. In a parallel direction, 35% of admissions were associated with severe infections in the cohort from Hong Kong, whereas only 11% and 14% of admissions from the cohort from Sweden and Hopkins Lupus Cohort, respectively. This diversity may be due to the differences in each cohort, such as duration of disease, disease activity, treatment, and environment.
The urinary tract was the most common site of severe infection in both the CAI and the HAI groups. Bacteria were the leading causative pathogens. These findings were consistent with other cohorts. Only 2 patients had opportunistic infections from the viral pathogens CMV and herpes zoster. There was no mycobacterium or systemic fungal infection in this cohort. It is possible that the patients in this cohort were in an early stage of disease and had been exposed to immunosuppressive agents only for a short period. As previously known, severe opportunistic infections were strongly associated with extensive use of immunosuppressive agents.
There was no association of ever having leukopenia and infection in SLE in this observation. White blood cell count in SLE usually vary with time, disease activity, and treatment, which were changing throughout the disease course. As mentioned previously, the association between leukopenia and infection is a time-dependent relationship. The WBC count at diagnosis or at only one time point may not predict an infection in a subsequent time. In this cohort, the WBC count was collected in prospective fashion every 2 to 4 months. There were 14 (33%) from 43 episodes associated with leukopenia within 3 months before severe infection. In analysis of association, when comparing the proportion of patients who ever had leukopenia between patients with and without severe infection, there was no difference of proportion of all 3 leukopenia conditions between both groups. For survival analysis, surprisingly, those patients who had leukopenia had a trend of longer SIFSR than patients who did not have leukopenia. The reason for this result may be because, in patients with leukopenia, the physicians may avoid using immunosuppressive agents or use them in a lower dosage.
Many previous studies tried to view the association between leukopenia and infection. The results in those studies were not consistent. Two studies found leukopenia was not associated with severe infection. However, a retrospective study from an early SLE cohort demonstrated that lymphopenia at presentation increased risk of infection. One cross-sectional study showed neutropenia in at least 1 determination during the follow-up period associated with infection. Another prospective cohort study showed adjusted neutropenia (the accumulated area under the curve before severe infection divided by interval follow-up days) less than 2500/mL was associated with severe infection. Therefore, not only a decrease in number of WBCs but also leukocyte component was considered to be the cause of infection.
From our cohort, in subsequent analysis, to see an association between WBC component deletion and severe infection, there was no difference of SIFSR between patients who ever and never had neutropenia, severe lymphopenia, concomitant neutropenia and lymphopenia in at least 1 determination during the observation, and persistent lymphopenia. Generally, susceptibility to bacterial infection increases in patients with neutropenia, and the rate of infection is inversely related to neutrophil count, whereas very severe lymphopenia (<200/mL) increases the risk of opportunistic infection in HIV patients. Although bacteria were the most common pathogen of patients in our study, there was no association of neutropenia and severe infection. There were 2 patients who developed transient, very severe lymphopenia. One of them did not have a severe infection, and another patient developed streptococcus pneumonia. However, both of them did not have an opportunistic infection. It may be because neutropenia and severe lymphopenia in these patients were transient conditions. Therefore, a longer duration of cytopenia may be necessary for the mechanism of infection.
Use of cyclophosphamide was an important risk factor for severe infection. This study did not demonstrate any differences in age, sex, educational level, economic status, marital status, disease activity and SLICC Damage Index at cohort entry, and clinical manifestation between patients with and without severe infections. There were also no differences with a cumulative dose of corticosteroids, use of azathioprine, MMF, methotrexate, or antimalarial agents. However, patients with severe infection received a higher proportion of cyclophosphamide. To confirm this result, subsequent survival analysis was done. The results also showed patients who received cyclophosphamide had a lower SIFSR than patients who did not receive cyclophosphamide. From Cox regression analysis, only the use of cyclophosphamide was the major risk factor of severe infection, which was independent from leukopenia or neutropenia. In contrast, a previous study that observed SLE patients treated with cyclophosphamide found cyclophosphamide to be responsible for more infections because it induced leukopenia. Therefore, the mechanisms of cyclophosphamide that increase the risk of severe infection may induce not only cytopenia, but also disruption of WBC function, which needs further study.
Limitations of this study include, first, a relatively small sample size. Second, patients in this study were at early SLE with approximately 3 years' duration of observation. Another, there were only 2 patients with severe viral infection and no patient with fungal and mycobacterium infection in our cohort. For that reason, these results may not represent what is seen in longstanding SLE patients and cannot generalize any role of leukopenia in opportunistic infection. However, our cohort measured WBCs by standard prospective protocol. Therefore, this study provided an opportunity to observe the accumulated status of low total WBCs and WBC component, and persistent cytopenia as well as the association between an abnormal WBC condition and severe infection, particularly common bacterial infection, over a period of time.
Discussion
Leukopenia is a common manifestation in SLE. This study demonstrated that 50% of patients had leukopenia at the diagnosis of SLE and increased to nearly 60% by the end of the study. The prevalence of leukopenia is comparable to its prevalence in previous studies, 36% to 66%. In this study, no patient had leukopenia with less than 2000/mL. Like most studies, WBC count less than 2000/mL is not common in SLE patients, but 1 study found 14% of untreated SLE patients had WBC count of less than 2000/mL.
Persistent lymphopenia was a common finding in SLE patients, whereas leukopenia and neutropenia were usually transient conditions. In our cohort, only 1 patient had persistent leukopenia without any immunosuppressive agent, and no patient had persistent neutropenia. On the other hand, 41% of patients had persistent lymphopenia. Similarly, in the Coimbra Lupus Cohort, persistent lymphopenia was found in 41%, and only 5% had persistent neutropenia.
In this study, all leukopenia was counted, no matter what the causes of leukopenia were. It should be noted that, in this study, 99% of all patients received corticosteroids, and 74% to 83% of patients who had leukopenia had been exposed to immunosuppressive agents. In fact, corticosteroids cause lymphopenia but neutrophilia, whereas immunosuppressive drugs can cause both lymphopenia and neutropenia. As known, from SLE disease activity, lymphopenia is more common than neutropenia. Therefore, some fraction of lymphopenia and neutropenia in SLE patients may be modified by medication. A previous study demonstrated that higher exposure to immunosuppressive agents and other drugs was an independent risk factor for neutropenia in SLE, (OR, 4.8 and 16.5, respectively).
Severe infection is common in SLE patients. In this study, the incidence of a severe infection rate as defined primarily by need for intravenous antibiotic therapy was 12.4 cases per 100 patient-years. This incidence was comparable to the study from the New York and Spanish populations, 13.9 and 17.0 cases per 100 patient-years, respectively. However, our severe infection rate was much higher than was observed in the Canadian and Swedish populations, 6.9 cases per 100 patient-years. Severe infections were responsible for 37% of admissions in this cohort and caused longer hospital stays. In a parallel direction, 35% of admissions were associated with severe infections in the cohort from Hong Kong, whereas only 11% and 14% of admissions from the cohort from Sweden and Hopkins Lupus Cohort, respectively. This diversity may be due to the differences in each cohort, such as duration of disease, disease activity, treatment, and environment.
The urinary tract was the most common site of severe infection in both the CAI and the HAI groups. Bacteria were the leading causative pathogens. These findings were consistent with other cohorts. Only 2 patients had opportunistic infections from the viral pathogens CMV and herpes zoster. There was no mycobacterium or systemic fungal infection in this cohort. It is possible that the patients in this cohort were in an early stage of disease and had been exposed to immunosuppressive agents only for a short period. As previously known, severe opportunistic infections were strongly associated with extensive use of immunosuppressive agents.
There was no association of ever having leukopenia and infection in SLE in this observation. White blood cell count in SLE usually vary with time, disease activity, and treatment, which were changing throughout the disease course. As mentioned previously, the association between leukopenia and infection is a time-dependent relationship. The WBC count at diagnosis or at only one time point may not predict an infection in a subsequent time. In this cohort, the WBC count was collected in prospective fashion every 2 to 4 months. There were 14 (33%) from 43 episodes associated with leukopenia within 3 months before severe infection. In analysis of association, when comparing the proportion of patients who ever had leukopenia between patients with and without severe infection, there was no difference of proportion of all 3 leukopenia conditions between both groups. For survival analysis, surprisingly, those patients who had leukopenia had a trend of longer SIFSR than patients who did not have leukopenia. The reason for this result may be because, in patients with leukopenia, the physicians may avoid using immunosuppressive agents or use them in a lower dosage.
Many previous studies tried to view the association between leukopenia and infection. The results in those studies were not consistent. Two studies found leukopenia was not associated with severe infection. However, a retrospective study from an early SLE cohort demonstrated that lymphopenia at presentation increased risk of infection. One cross-sectional study showed neutropenia in at least 1 determination during the follow-up period associated with infection. Another prospective cohort study showed adjusted neutropenia (the accumulated area under the curve before severe infection divided by interval follow-up days) less than 2500/mL was associated with severe infection. Therefore, not only a decrease in number of WBCs but also leukocyte component was considered to be the cause of infection.
From our cohort, in subsequent analysis, to see an association between WBC component deletion and severe infection, there was no difference of SIFSR between patients who ever and never had neutropenia, severe lymphopenia, concomitant neutropenia and lymphopenia in at least 1 determination during the observation, and persistent lymphopenia. Generally, susceptibility to bacterial infection increases in patients with neutropenia, and the rate of infection is inversely related to neutrophil count, whereas very severe lymphopenia (<200/mL) increases the risk of opportunistic infection in HIV patients. Although bacteria were the most common pathogen of patients in our study, there was no association of neutropenia and severe infection. There were 2 patients who developed transient, very severe lymphopenia. One of them did not have a severe infection, and another patient developed streptococcus pneumonia. However, both of them did not have an opportunistic infection. It may be because neutropenia and severe lymphopenia in these patients were transient conditions. Therefore, a longer duration of cytopenia may be necessary for the mechanism of infection.
Use of cyclophosphamide was an important risk factor for severe infection. This study did not demonstrate any differences in age, sex, educational level, economic status, marital status, disease activity and SLICC Damage Index at cohort entry, and clinical manifestation between patients with and without severe infections. There were also no differences with a cumulative dose of corticosteroids, use of azathioprine, MMF, methotrexate, or antimalarial agents. However, patients with severe infection received a higher proportion of cyclophosphamide. To confirm this result, subsequent survival analysis was done. The results also showed patients who received cyclophosphamide had a lower SIFSR than patients who did not receive cyclophosphamide. From Cox regression analysis, only the use of cyclophosphamide was the major risk factor of severe infection, which was independent from leukopenia or neutropenia. In contrast, a previous study that observed SLE patients treated with cyclophosphamide found cyclophosphamide to be responsible for more infections because it induced leukopenia. Therefore, the mechanisms of cyclophosphamide that increase the risk of severe infection may induce not only cytopenia, but also disruption of WBC function, which needs further study.
Limitations of this study include, first, a relatively small sample size. Second, patients in this study were at early SLE with approximately 3 years' duration of observation. Another, there were only 2 patients with severe viral infection and no patient with fungal and mycobacterium infection in our cohort. For that reason, these results may not represent what is seen in longstanding SLE patients and cannot generalize any role of leukopenia in opportunistic infection. However, our cohort measured WBCs by standard prospective protocol. Therefore, this study provided an opportunity to observe the accumulated status of low total WBCs and WBC component, and persistent cytopenia as well as the association between an abnormal WBC condition and severe infection, particularly common bacterial infection, over a period of time.
Source...