Leukopenia and Severe Infection in Patients With SLE
There were 100 patients in the lupus cohort from May 2007 to June 2011. We excluded 10 patients who had a follow-up period of less than 1 year and 1 patient who had incomplete data. Eighty-nine patients were included in an analysis. Eighty-two patients (92%) were female, and the female-to-male ratio was 12:1. The mean age at cohort entry was 31.66 (SD, 12.22) years. The mean duration from diagnosis to cohort entry was 2.40 (SD, 2.93) months. At cohort entry, most patients had high disease activity and a low damage index. The mean M-SLEDAI-2K was 10.66 (SD, 7.18), and the mean SLICC damage index was 0.67 (SD, 1.16). The 3 most common clinical features at diagnosis were hematologic abnormality (75.28%), arthritis (56.18%), and nephritis (50.56%). For the hematologic abnormality category, there were autoimmune hemolytic anemia, 30.34%; leukopenia, 51.68%; and thrombocytopenia, 7.86%.
From all 89 patients, 51.68% of patients had leukopenia at presentation. During the mean observational period 34.95 (SD, 14.67) months, the cumulative leukopenia, lymphopenia, and neutropenia were 57.30%, 96.63%, and 60.67%. Only 1 patient had persistent leukopenia. This patient had not received any immunosuppressive agent. Persistent lymphopenia was 41.57%, whereas no patient had persistent neutropenia (Table 1).
Most patients received corticosteroids (98.88%) and antimalarial drugs (83.15%). Seventy-one patients (79.78%) received immunosuppressive drugs, cyclophosphamide both oral (34.62%) and intravenous (65.38%) route (mean dose [min, max] 50 [25, 75] mg/d in oral route and 800 [400, 1200] mg/mo in intravenous route), azathioprine (50 [25, 100] mg/d), mycophenolate mofetil (MMF) (1500 [250, 3000] mg/d), and/or methotrexate (7.5 [5, 15] mg/wk), in some periods during the observation. Only 18 patients (20.22%) did not have any immunosuppressive agent. Thirty-eight (74.51%) from 51 patients with leukopenia, 40 (74.07%) from 54 patients with neutropenia, 1 (50.00%) in 2 patients with severe neutropenia, and 20 (83.33%) from 24 patients with severe lymphopenia received at least 1 immunosuppressive agent.
There were 43 episodes of severe infection from 27 patients (30.34%). The mean observational period in this study was 34.95 (SD, 14.67) months, and the cumulative follow-up was 217 patient-years. The incidence of severe infection rate was 12.44 cases per 100 patient-years. From 43 severe infection episodes, 42 episodes were hospitalized, and one other episode of herpes zoster was treated in the outpatient department. Forty-two (36.84%) from a total of 114 admissions were associated with severe infection. Severe infection was the cause of admission in 30 (26.32%) from 114 admissions. Severe infection developed during hospitalization in 12 (10.53%) episodes from 114 admissions. The mean of hospital stay for the admissions that was associated with severe infection was longer than the admissions that were not associated with severe infection, 13.46 (SD, 12.44) versus 6.47 (SD, 4.14) days (P = 0.003).
For those reasons, 43 severe infection episodes were categorized into 2 groups, community-acquired infection (CAI) and hospital-acquired infection (HAI). Thirty-one (72.09%) of 43 episodes of severe infection were CAI. The most common site of infection in both the CAI and HAI groups was urinary tract infection, 48.39% in the CAI group and 33.33% in the HAI group. Escherichia coli was the most common organism of urinary tract infection in CAI group (40.00%) (Table 2). Two patients had an opportunistic infection: 1 patient had CMV pneumonitis, and another had herpes zoster.
Fourteen (32.56%) from 43 episodes of severe infection were associated with leukopenia. Nine episodes had leukopenia within 3 months before a severe infection but not at the diagnosis of severe infection, 3 episodes had leukopenia within 3 months before and at the diagnosis of severe infection, and 2 episodes had leukopenia only at the diagnosis of severe infection.
The proportion of patients who ever had leukopenia, neutropenia, severe lymphopenia in at least 1 determination during the observation, and persistent lymphopenia was not different between patients with and without severe infection, as shown in Table 3. Because 96.63% of patients had lymphopenia, the proportion of patients with lymphopenia between patients with and without severe infection was not compared.
Using the Fisher exact test for comparing patients with and without severe infection, there was no difference of age at SLE diagnosis, marital status, household income per year, and level of education. Also, there was no difference of SLICC damage index and M-SLEDAI-2K at cohort entry and clinical presentations between the 2 groups.
For treatment factors, the proportion of patients who received cyclophosphamide was higher in patients with severe infection than in patients without severe infection (77.78% vs 50.00%, P = 0.015), and the odds ratio (OR) was 2.50 (95% confidence interval [CI], 1.11–5.56). There was no difference of cumulative corticosteroid dose and proportion of MMF, azathioprine, methotrexate, or antimalarial drug between patients with and without severe infection (Table 4).
From 89 patients, the 12-month severe infection–free survival rate (12-month SIFSR) was 78.65% (SD, 4.34%) (Fig. 1). Using the log-rank test, patients who ever had leukopenia had a higher SIFSR than patients who never had leukopenia. The probability of 12-month SIFSR between patients who ever and never had leukopenia was 82.35% (SD, 5.34%) versus 71.05% (SD, 7.36%) (P = 0.060) (Fig. 2). However, there was no difference of SIFSR between patients who ever and never had neutropenia, combined neutropenia and lymphopenia, persistent lymphopenia, and severe lymphopenia.
(Enlarge Image)
Figure 1.
The cumulative probability of SIFSR (n = 89).
(Enlarge Image)
Figure 2.
The cumulative probability of SIFSR between patients with and without white leukopenia (n = 89).
For treatment factors, patients treated without cyclophosphamide had a higher SIFSR than patients treated with cyclophosphamide. The probability of 12-month SIFSR between patients treated without and with cyclophosphamide was 83.78% (SD, 6.06%) vs 73.08% (SD, 6.15%) (P = 0.020) (Fig. 3). There was no difference of SIFSR between patients with and without MMF, azathioprine or methotrexate.
(Enlarge Image)
Figure 3.
The cumulative probability of SIFSR between patients with and without cyclophosphamide (n = 89). P value was from log-rank test.
On multivariate analysis by using Cox regression analysis, only patients treated with cyclophosphamide had an increased risk of severe infection: hazard ratio 2.73 (95% CI, 1.10–6.77; P = 0.031). After adjustment for cyclophosphamide, leukopenia was not associated with severe infection: hazard ratio 0.51 (95% CI, 0.234–1.089; P = 0.081).
There were 6 deaths (6.74%) during this observation period. Severe pneumonia was the cause of death in 3 cases. One of them had leukopenia and neutropenia at the diagnosis of pneumonia, but the other 2 cases had no leukopenia. Three other cases died of massive pneumothorax, lung hemorrhage, and liver hemorrhage. The latter 3 cases had no evidence of infection.
Results
There were 100 patients in the lupus cohort from May 2007 to June 2011. We excluded 10 patients who had a follow-up period of less than 1 year and 1 patient who had incomplete data. Eighty-nine patients were included in an analysis. Eighty-two patients (92%) were female, and the female-to-male ratio was 12:1. The mean age at cohort entry was 31.66 (SD, 12.22) years. The mean duration from diagnosis to cohort entry was 2.40 (SD, 2.93) months. At cohort entry, most patients had high disease activity and a low damage index. The mean M-SLEDAI-2K was 10.66 (SD, 7.18), and the mean SLICC damage index was 0.67 (SD, 1.16). The 3 most common clinical features at diagnosis were hematologic abnormality (75.28%), arthritis (56.18%), and nephritis (50.56%). For the hematologic abnormality category, there were autoimmune hemolytic anemia, 30.34%; leukopenia, 51.68%; and thrombocytopenia, 7.86%.
Leukopenia
From all 89 patients, 51.68% of patients had leukopenia at presentation. During the mean observational period 34.95 (SD, 14.67) months, the cumulative leukopenia, lymphopenia, and neutropenia were 57.30%, 96.63%, and 60.67%. Only 1 patient had persistent leukopenia. This patient had not received any immunosuppressive agent. Persistent lymphopenia was 41.57%, whereas no patient had persistent neutropenia (Table 1).
Most patients received corticosteroids (98.88%) and antimalarial drugs (83.15%). Seventy-one patients (79.78%) received immunosuppressive drugs, cyclophosphamide both oral (34.62%) and intravenous (65.38%) route (mean dose [min, max] 50 [25, 75] mg/d in oral route and 800 [400, 1200] mg/mo in intravenous route), azathioprine (50 [25, 100] mg/d), mycophenolate mofetil (MMF) (1500 [250, 3000] mg/d), and/or methotrexate (7.5 [5, 15] mg/wk), in some periods during the observation. Only 18 patients (20.22%) did not have any immunosuppressive agent. Thirty-eight (74.51%) from 51 patients with leukopenia, 40 (74.07%) from 54 patients with neutropenia, 1 (50.00%) in 2 patients with severe neutropenia, and 20 (83.33%) from 24 patients with severe lymphopenia received at least 1 immunosuppressive agent.
Severe Infection
There were 43 episodes of severe infection from 27 patients (30.34%). The mean observational period in this study was 34.95 (SD, 14.67) months, and the cumulative follow-up was 217 patient-years. The incidence of severe infection rate was 12.44 cases per 100 patient-years. From 43 severe infection episodes, 42 episodes were hospitalized, and one other episode of herpes zoster was treated in the outpatient department. Forty-two (36.84%) from a total of 114 admissions were associated with severe infection. Severe infection was the cause of admission in 30 (26.32%) from 114 admissions. Severe infection developed during hospitalization in 12 (10.53%) episodes from 114 admissions. The mean of hospital stay for the admissions that was associated with severe infection was longer than the admissions that were not associated with severe infection, 13.46 (SD, 12.44) versus 6.47 (SD, 4.14) days (P = 0.003).
For those reasons, 43 severe infection episodes were categorized into 2 groups, community-acquired infection (CAI) and hospital-acquired infection (HAI). Thirty-one (72.09%) of 43 episodes of severe infection were CAI. The most common site of infection in both the CAI and HAI groups was urinary tract infection, 48.39% in the CAI group and 33.33% in the HAI group. Escherichia coli was the most common organism of urinary tract infection in CAI group (40.00%) (Table 2). Two patients had an opportunistic infection: 1 patient had CMV pneumonitis, and another had herpes zoster.
Leukopenia and Severe Infection
Fourteen (32.56%) from 43 episodes of severe infection were associated with leukopenia. Nine episodes had leukopenia within 3 months before a severe infection but not at the diagnosis of severe infection, 3 episodes had leukopenia within 3 months before and at the diagnosis of severe infection, and 2 episodes had leukopenia only at the diagnosis of severe infection.
The proportion of patients who ever had leukopenia, neutropenia, severe lymphopenia in at least 1 determination during the observation, and persistent lymphopenia was not different between patients with and without severe infection, as shown in Table 3. Because 96.63% of patients had lymphopenia, the proportion of patients with lymphopenia between patients with and without severe infection was not compared.
The Risk Factors of Severe Infection
Using the Fisher exact test for comparing patients with and without severe infection, there was no difference of age at SLE diagnosis, marital status, household income per year, and level of education. Also, there was no difference of SLICC damage index and M-SLEDAI-2K at cohort entry and clinical presentations between the 2 groups.
For treatment factors, the proportion of patients who received cyclophosphamide was higher in patients with severe infection than in patients without severe infection (77.78% vs 50.00%, P = 0.015), and the odds ratio (OR) was 2.50 (95% confidence interval [CI], 1.11–5.56). There was no difference of cumulative corticosteroid dose and proportion of MMF, azathioprine, methotrexate, or antimalarial drug between patients with and without severe infection (Table 4).
Severe Infection–free Survival Rate
From 89 patients, the 12-month severe infection–free survival rate (12-month SIFSR) was 78.65% (SD, 4.34%) (Fig. 1). Using the log-rank test, patients who ever had leukopenia had a higher SIFSR than patients who never had leukopenia. The probability of 12-month SIFSR between patients who ever and never had leukopenia was 82.35% (SD, 5.34%) versus 71.05% (SD, 7.36%) (P = 0.060) (Fig. 2). However, there was no difference of SIFSR between patients who ever and never had neutropenia, combined neutropenia and lymphopenia, persistent lymphopenia, and severe lymphopenia.
(Enlarge Image)
Figure 1.
The cumulative probability of SIFSR (n = 89).
(Enlarge Image)
Figure 2.
The cumulative probability of SIFSR between patients with and without white leukopenia (n = 89).
For treatment factors, patients treated without cyclophosphamide had a higher SIFSR than patients treated with cyclophosphamide. The probability of 12-month SIFSR between patients treated without and with cyclophosphamide was 83.78% (SD, 6.06%) vs 73.08% (SD, 6.15%) (P = 0.020) (Fig. 3). There was no difference of SIFSR between patients with and without MMF, azathioprine or methotrexate.
(Enlarge Image)
Figure 3.
The cumulative probability of SIFSR between patients with and without cyclophosphamide (n = 89). P value was from log-rank test.
On multivariate analysis by using Cox regression analysis, only patients treated with cyclophosphamide had an increased risk of severe infection: hazard ratio 2.73 (95% CI, 1.10–6.77; P = 0.031). After adjustment for cyclophosphamide, leukopenia was not associated with severe infection: hazard ratio 0.51 (95% CI, 0.234–1.089; P = 0.081).
Death
There were 6 deaths (6.74%) during this observation period. Severe pneumonia was the cause of death in 3 cases. One of them had leukopenia and neutropenia at the diagnosis of pneumonia, but the other 2 cases had no leukopenia. Three other cases died of massive pneumothorax, lung hemorrhage, and liver hemorrhage. The latter 3 cases had no evidence of infection.