Abstract
Background
Linezolid (LZD), an oxazolidinone antibiotic agent, has excellent activity and bioavailability against most methicillin-sensitive and methicillin-resistant gram-positive bacteria. Although LZD is generally well tolerated, several studies have found adverse hematologic effects, of which thrombocytopenia is of most concern.
Objective
To investigate the risk factors for thrombocytopenia in patients who received oral or parenteral LZD therapy between February 1 and November 30, 2010.
Methods
Data were extracted retrospectively from the electronic medical records in our hospital information system. Thrombocytopenia was defined as either a final platelet count of <100 × 109/L (criterion 1) or a 25% reduction from the baseline platelet count (criterion 2). Risk factors were determined using logistic regression analysis, and clinical features were predicted using receiver operating characteristic curves.
Results
The study included 254 patients, with mean (SD) age of 59 (17.66) years. The duration of LZD therapy was 9.43 (5.63) days. Thrombocytopenia developed in 69 patients (27.2%), as defined by criterion 1, and in 127 patients (50%), as defined by criterion 2. At univariate analysis, age, weight, creatinine clearance, serum albumin concentration, baseline platelet count, daily dosage, and concomitant use of caspofungin, levofloxacin, and meropenem were significant risk factors for thrombocytopenia. At multivariate analysis and using ROC curves, daily dose ≥18.75 mg/kg, baseline platelet count ≤181 × 109/L, duration of LZD therapy ≥10 days, and concomitant use of caspofungin and levofloxacin were independent risk factors for thrombocytopenia as defined by criterion 1, whereas creatinine clearance ≤88.39 mL/min/1.73 m2, serum albumin concentration ≤33.5 g/L, daily dose ≥18.46 mg/kg, and caspofungin were independent risk factors for thrombocytopenia as defined by criterion 2.
Conclusions
The incidence of LZD-related thrombocytopenia in the Chinese population is much higher than that suggested by the drug instructions. Low pretreatment platelet count, low body weight, low serum albumin concentration, long-term drug administration, advanced age, renal insufficiency, and concomitant use of caspofungin, levofloxacin, and meropenem have been identified as risk factors. Although predictors have been proposed for use in clinical practice to screen for patients at high risk who require intensified monitoring, further research on the dosage-based pharmacokinetics and pharmacodynamics of LZD are urgently needed.
Key words: baseline platelet count, daily dosage, linezolid, risk factor, serum albumin, thrombocytopenia
Linezolid (LZD), an oxazolidinone antibiotic agent, has excellent activity and bioavailability against most methicillin-sensitive and methicillin-resistant gram-positive bacteria.1–3 Although LZD is generally well tolerated, studies have found associated adverse hematologic effects such as thrombocytopenia, anemia, and hematocytopenia. Of these, thrombocytopenia is of most concern. Linezolid has been available in China since 2007, and many instances of LZD-associated thrombocytopenia have recently been reported in China.4 However, the incidence of and risk factors for LZD-induced thrombocytopenia in a large sample of Chinese have not been reported to date. Moreover, the incidence of LZD-related thrombocytopenia varies between countries such as the United States5,6 and Japan.7,8
The present retrospective cross-sectional study was performed to investigate the risk factors for thrombocytopenia in adult Chinese patients who received LZD therapy, to screen for important clinical features that can be used to predict the incidence of thrombocytopenia, to provide a reasonable assessment of the risk of LZD-related thrombocytopenia in the Chinese population, and to reduce the incidence of serious adverse reactions.
Patients and Methods
This study was conducted at the 4000-bed Chinese PLA General Hospital. Patients who received LZD, 1200 mg (600 mg q12h) via oral, parenteral, or both routes, were examined. Data were collected from the electronic medical records in the hospital central database between February 1 and November 30, 2010.
Exclusion criteria were as follows: (1) age <18 years; (2) blood system disease or platelet decrease as the original disease, for example, acute nonlymphocytic leukemia, lymphocytic leukemia, myelodysplasia, aplastic anemia, acute myeloid leukemia, tumor being treated using chemotherapy, severe pancreatitis, hepatoblastoma, and systemic lupus erythematosus; (3) platelet count not recorded before or after LZD therapy; and (4) abnormal platelet count (<100 × 109/L or >400 × 109/L) before LZD therapy.
Thrombocytopenia was defined as platelet count <100 × 109/L (criterion 1) or 25% reduction from baseline platelet count (criterion 2) because of current lack of uniform diagnostic criteria for LZD-induced thrombocytopenia. Criterion 1 was considered standard for severe thrombocytopenia, and criterion 2 for small decrease in platelet count. Therefore, the defined diseases per both criteria can be better compared with other published data.7,9 Moreover, a more inclusive standard can better prevent severe adverse events. Variables including sex, age, body weight, daily dosage of LZD, and duration of LZD therapy, and laboratory data including alanine aminotransferase, total bilirubin, creatinine, and albumin concentrations, and baseline platelet count before LZD therapy were analyzed as candidate risk factors for thrombocytopenia.
Statistical analysis was performed using commercially available software (SAS version 9.1.3; SAS Institute, Inc, Cary, North Carolina). The Student t-test (normal distribution) or Mann-Whitney U test (non-normal distribution) was used to compare continuous variables, and the χ2 test to compare dichotomous variables. P < 0.05 was considered statistically significant. Univariate and multiple logistic regression analyses were performed for each variable. Receiver operating characteristic curves (ROCs) were used to estimate the sensitivity, specificity, accuracy, and cutoff values for continuous variables that closely related to thrombocytopenia in multiple logistic regression analysis.
Results
Patient Data
During the study, 512 patients were screened. According to the exclusion criteria, 254 patients (169 men and 85 women; mean [SD] age, 59.0 [17.7] years [range, 18–95 years]) were included in the study. Duration of LZD therapy was 9.43 (5.63) days (range, 1.5–35.5 days). Linezolid-induced thrombocytopenia was determined in 69 patients (27.2%) on the basis of criteria 1, and 127 patients (50%) according to criteria 2. In 27 patients (10.6%), thrombocytopenia was grade III or IV according to the World Health Organization assessment of acute and subacute adverse performance and indexing standards (grade III, 26–49 × 109/L; and grade IV, ≤25 × 109/L). Seventeen patients (6.7%) required platelet transfusions.
Risk of Thrombocytopenia
Sixteen variables potentially associated with development of thrombocytopenia were tested using univariate analysis. For thrombocytopenia defined using criterion 1, variables including age, weight, creatinine clearance, serum albumin concentration, daily dosage, baseline platelet count, duration of medication, and levofloxacin, meropenem, and caspofungin therapy exhibited a statistically significant difference between patients with or without thrombocytopenia. For thrombocytopenia defined using criterion 2, similar results were obtained except for the baseline platelet count, which was not statistically significant different between patients with or without thrombocytopenia (Table I).
Table I.
Demographic and clinical data.
| Variable | 25% Reduction from Baseline Platelet Count |
Final Platelet Count <100 × 109/L |
||||
|---|---|---|---|---|---|---|
| Patients With Thrombocytopenia (n = 127) | Patients Without Thrombocytopenia (n = 127) | P | Patients With Thrombocytopenia (n = 69) | Patients Without Thrombocytopenia (n = 185) | P | |
| Demographic data | ||||||
| Sex, No. (%) | 0.35⁎ | 0.57⁎ | ||||
| Male | 81 (63.78) | 88 (69.29) | 44 (63.77) | 125 (67.57) | ||
| Female | 46 (36.22) | 39 (30.71) | 25 (36.23) | 60 (32.43) | ||
| Age, y | 61.36 (17.11) | 56.83 (18.03) | 0.04† | 63.49 (15.97) | 57.45 (18.06) | 0.02† |
| No. of pts (missing) | 127 (0) | 127 (0) | NA | 69 (0) | 185 (0) | NA |
| Weight, kg | 62.15 (12.48) | 67.15 (13.45) | 0.002† | 61.70 (13.18) | 65.75 (13.06) | 0.03† |
| No. of pts (missing) | 127 (0) | 127 (0) | NA | 69 (0) | 185 (0) | NA |
| Laboratory data | ||||||
| Alanine aminotransferase (U/L) | 33.69 (42.93) | 34.86 (42.87) | 0.26‡ | 30.24 (27.55) | 35.79 (47.27) | 0.31‡ |
| No. of pts (missing) | 127 (0) | 126 (1) | NA | 69 (0) | 184 (1) | NA |
| Total bilirubin (μmol/L) | 20.12 (31.82) | 15.48 (14.75) | 0.63‡ | 18.01 (17.52) | 17.76 (27.30) | 0.38‡ |
| No. of pts (missing) | 127 (0) | 124 (3) | NA | 69 (0) | 182 (3) | NA |
| Creatinine clearance, mL/min | 87.67 (59.00) | 106.90 (58.28) | 0.004‡ | 84.37 (63.12) | 102.07 (57.26) | 0.009‡ |
| No. of pts (missing) | 127 (0) | 126 (1) | NA | 69 (0) | 184 (1) | NA |
| Albumin, g/L | 31.89 (5.12) | 34.45 (10.08) | 0.01‡ | 31.34 (5.27) | 33.84 (8.79) | 0.005‡ |
| No. of pts (missing) | 126 (1) | 123 (4) | NA | 68 (1) | 181 (4) | NA |
| Daily dose, mg/kg | 19.99 (4.18) | 17.83 (4.39) | <0.001‡ | 20.11 (4.14) | 18.46 (4.44) | 0.002‡ |
| No. of pts (missing) | 127 (0) | 127 (0) | NA | 69 (0) | 185 (0) | NA |
| Baseline platelet count, ×109/L | 212.50 (76.96) | 224.71 (72.85) | 0.12‡ | 176.96 (61.97) | 234.14 (73.68) | <0.001‡ |
| No. of pts (missing) | 127 (0) | 127 (0) | NA | 69 (0) | 185 (0) | NA |
| Treatment duration, d | 10.31 (5.64) | 8.55 (5.54) | 0.002‡ | 10.52 (5.01) | 9.02 (5.83) | 0.007‡ |
| No. of pts (missing) | 127 (0) | 127 (0) | NA | 69 (0) | 185 (0) | NA |
| Anti-infective agent coadministered with linezolid, No. (%) | ||||||
| Nuoroquinolone | 47 (37.01) | 37 (29.13) | 0.18⁎ | 29 (42.03) | 55 (29.73) | 0.06⁎ |
| Moxifloxacin | 14 (11.02) | 14 (11.02) | >0.99 | 9 (13.04) | 19 (10.27) | 0.53⁎ |
| Levofloxacin | 33 (25.98) | 20 (15.75) | 0.04⁎ | 20 (28.99) | 33 (17.84) | 0.05⁎ |
| β-Lactams | 63 (49.61) | 61 (48.03) | 0.80⁎ | 34 (49.28) | 90 (48.65) | 0.93⁎ |
| Piperacillin and tazobactam sodium | 15 (11.81) | 14 (11.02) | 0.84⁎ | 5 (7.25) | 24 (12.97) | 0.20⁎ |
| Cefoperazone and sulbactam sodium | 35 (27.56) | 26 (20.47) | 0.19⁎ | 22 (31.88) | 39 (21.08) | 0.07⁎ |
| Carbapenems | 75 (59.06) | 55 (43.31) | 0.01⁎ | 42 (60.87) | 88 (47.57) | 0.06⁎ |
| Imipenem and cilastatin sodium | 21 (16.54) | 19 (14.96) | 0.73⁎ | 10 (14.49) | 30 (16.22) | 0.74⁎ |
| Meropenem | 55 (43.31) | 55 (43.31) | 0.009 | 31 (44.93) | 59 (31.89) | 0.05⁎ |
| Antifungal agents | 62 (48.82) | 48 (37.80 | 0.08⁎ | 35 (50.72) | 75 (40.54) | 0.15⁎ |
| Voriconazole | 14 (11.02) | 12 (9.45%) | 0.68⁎ | 9 (13.04) | 17 (9.19) | 0.37⁎ |
| Caspofungin | 26 (20.47) | 11 (8.66) | 0.008⁎ | 16 (23.19) | 21 (11.35) | 0.02⁎ |
| Fluconazole | 29 (22.83) | 25 (19.69) | 0.54⁎ | 16 (23.19) | 38 (20.54) | 0.65⁎ |
Unless otherwise noted, data are given as mean (SD).
NA = not applicable; pts = patients.
χ2 test.
Student t test.
Mann-Whitney U test.
At univariate logistic regression 7 variables including age, weight, creatinine clearance, serum albumin concentration, daily dosage, baseline platelet count, and caspofungin therapy were significant risk factors for LZD-associated thrombocytopenia as defined using criterion 1, whereas 9 variables including age, weight, creatinine clearance, serum albumin concentration, daily dosage, treatment duration, and levofloxacin, meropenem, and caspofungin therapy were significantly correlated with thrombocytopenia as defined using criterion 2 (Table II). According to regression coefficient analysis, whereas creatinine clearance, serum albumin concentration, baseline platelet count, and weight were protective factors, age and daily dosage were risk factors.
Table II.
Risk factors for thrombocytopenia selected using logistic regression.⁎
| Risk Factor | 25% Reduction From Baseline Platelet Count |
Final Platelet Count <100 × 109/L |
||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P | OR | 95% CI | P | |
| Analysis by univariate logistic regression | ||||||
| Age, y | 1.01 | 1.00–1.03 | 0.04 | 1.02 | 1.00–1.04 | 0.02 |
| Weight, kg | 0.97 | 0.95–0.99 | 0.003 | 0.98 | 0.95–1.00 | 0.03 |
| Creatinine clearance, mL/min | 0.99 | 0.99–1.00 | 0.01 | 0.99 | 0.99–1.00 | 0.04 |
| Albumin, g/L | 0.94 | 0.90–0.99 | 0.01 | 0.94 | 0.89–0.99 | 0.01 |
| Daily dose, mg/kg | 1.13 | 1.06–1.20 | 0.002 | 1.09 | 1.02–1.16 | 0.009 |
| Baseline platelet count, ×109/L | 1.00 | 0.99–1.00 | 0.20 | 0.99 | 0.98–0.99 | <0.001 |
| Treatment duration, d | 1.06 | 1.01–1.11 | 0.02 | 1.05 | 1.00–1.10 | 0.06 |
| Moxifloxacin | 1.00 | 0.46–2.19 | >0.99 | 1.31 | 0.56–3.05 | 0.53 |
| Levofloxacin | 1.88 | 1.01–3.49 | 0.046 | 1.88 | 0.99–3.57 | 0.05 |
| Piperacillin and tazobactam sodium | 1.08 | 0.5–2.34 | 0.84 | 0.52 | 0.19–1.43 | 0.21 |
| Cefoperazone and sulbactam sodium | 1.48 | 0.83–2.64 | 0.19 | 1.75 | 0.95–3.25 | 0.07 |
| Imipenem and cilastatin sodium | 1.13 | 0.57–2.21 | 0.73 | 0.88 | 0.40–1.90 | 0.74 |
| Meropenem | 2.01 | 1.19–3.39 | 0.009 | 1.74 | 0.99–3.07 | 0.054 |
| Voriconazole | 1.19 | 0.53–2.68 | 0.68 | 1.48 | 0.63–3.50 | 0.37 |
| Caspofungin | 2.71 | 1.28–5.77 | 0.009 | 2.36 | 1.15–4.85 | 0.02 |
| Fluconazole | 1.21 | 0.66–2.20 | 0.54 | 1.17 | 0.60–2.27 | 0.65 |
| Selected by multivariate logistic analysis | ||||||
| Creatinine clearance, mL/min | 0.10 | 0.99–1.00 | 0.04 | NA | NA | NA |
| Albumin, g/L | 0.95 | 0.90–0.19 | 0.03 | NA | NA | NA |
| Daily dose, mg/kg | 1.12 | 1.05–1.20 | 0.001 | 1.08 | 1.01–1.16 | 0.03 |
| Baseline platelet count, ×109/L | NA | NA | NA | 0.99 | 0.98–0.99 | 0.001 |
| Treatment duration, d | NA | NA | NA | 1.06 | 1.01–1.12 | 0.03 |
| Caspofungin | 2.81 | 1.28–6.20 | 0.01 | 2.41 | 1.09–5.3 | 0.04 |
| Levofloxacin | NA | NA | NA | 2.00 | 0.97–4.11 | 0.03 |
CI = confidence interval; NA = not available; OR = odds ratio.
P < 0.05.
At multivariate logistic regression analysis, baseline platelet count, daily dosage, duration of medication, and caspofungin and levofloxacin therapy were independent risk factors for LZD-induced thrombocytopenia as defined using criterion 1, whereas creatinine clearance, serum albumin concentration, daily dosage, and caspofungin therapy were significantly associated with LZD-induced thrombocytopenia as defined using criterion 2 (Table II).
Critical Features of Thrombocytopenia as Predicated Using ROC Curves
At logistic regression multivariate analysis, baseline platelet count, daily dosage, and duration of medication, which related to thrombocytopenia as defined using criterion 1, were included to calculate the area under the ROC curve (AUC). The maximum Youden index was 0.44, and the AUC was 0.76. According to the cutoff point, the daily dose was 18.75 mg/kg, the baseline platelet count was 181 × 109/L, and the duration of medication was 10 days.
Creatinine clearance, serum albumin concentration, and daily dosage, as related to thrombocytopenia defined using criterion 2, were used to calculate the AUC. The maximum Youden index was 0.37, and the AUC was 0.71. According to the cutoff point, creatinine clearance was 88.39 mL/min/1.73 m2, the serum albumin concentration was 33.5 g/L, and the daily dose was 18.46 mg/kg.
Discussion
The mechanisms underlying LZD-related hematologic toxicity remain unclear, and the reported incidence of LZD-induced thrombocytopenia varies. Attassi et al5 and Orrick et al6 have reported the incidences of LZD-induced thrombocytopenia in western countries. When thrombocytopenia was defined as a ≥30% decrease in platelet count, similar incidences were observed: 47% (9 of 19 patients)5 vs 48% (23 of 48 patients).6 When thrombocytopenia was defined as platelet count <100 × 109/L, thrombocytopenia was 32% (6 of 19 patients)5 vs19% (9 of 48 patients).6 Thrombocytopenia has been reported in 7 of 42 patients (16.7%) by Niwa et al7 in Japan. Although criterion 2 in the present study is more inclusive, all of these studies included a relatively small sample size. Recently, another report from Japan reported drug-induced thrombocytopenia (defined as <100 × 109/L or a 30% reduction from baseline) in 128 of 331 included patients (38.7%).8
The present study reports LZD-induced thrombocytopenia in the Chinese population on the basis of a survey of 254 patients. Incidence of thrombocytopenia defined using criterion 2 was much higher than that defined using criterion 1: 50% vs 27.2%. The results indicated that a small decrease in platelet count is common, and the independent risk factors are different for a small decrease in platelet count vs severe thrombocytopenia. In addition, physicians should be concerned about hemorrhagic tendency. The results of the present study are consistent with those of previous reports; however, the incidence of thrombocytopenia reported herein is much higher than that in LZD-treated adult patients with substantially low platelet counts in phase III comparator-controlled trials.10 Such differences may be attributed in part to inclusion of patients with comorbid conditions, who were excluded from the phase III clinical trials. Inasmuch as different criteria result in a significantly different incidence, it is necessary to unify the standard in this field to facilitate comparisons between various studies.
In general, the cutoff point for risk factors is determined by the well-known human pathologic classification, such as age ≥65 years (elderly group), creatinine clearance ≤30 or 50 mL/min/1.73 m2 (grading the severity of renal dysfunction), and baseline platelet count ≤200 × 109/L (normal), rather than analysis of clinical data. The present study considered thrombocytopenia as a drug-induced disease, and used the independent risk factors as predictors. The ROC curve was used to predict the cutoff point with good sensitivity and specificity for development of thrombocytopenia. The AUC >0.5 under both defining criteria is a good initial predictive (diagnostic) value.11 With a daily dose of ≥18.75 mg/kg (P = 0.005; odds ratio [OR], 2.70; 95% confidence interval [CI], 1.54–4.74), baseline platelet count ≤181 × 109/L (P < 0.001; OR, 3.93; 95% CI, 2.24–6.90), or number of treatment days ≥10 (P = 0.04; OR, 1.77; 95% CI, 1.02–3.09), patients are susceptible to thrombocytopenia or bleeding risk, and blood monitoring should be more frequent. With creatinine clearance ≤88.39 mL/min/1.73 m2 (P = 0.005; OR, 2.06; 95% CI, 1.25–3.39), serum albumin concentration ≤33.5 g/L (P = 0.049; OR, 1.67; 95% CI, 1.00–2.77), daily dose of LZD ≥18.46 mg/kg (P < 0.001; OR, 2.72; 95% CI, 1.64–4.50), patients are at risk of a small decrease in platelet count. Because these clinical data are easy to obtain, they are useful in clinical practice to screen high-risk patients for intensified monitoring.
In the present study, creatinine concentration was chosen as the indicator of renal function, and its clearance rate was calculated using the Cockcroft-Gault formula. The results indicated that thrombocytopenia may develop when creatinine clearance is at its lower limit, while there are yet no signs or symptoms of renal insufficiency. Moreover, we observed a higher incidence of thrombocytopenia in the group with severe renal impairment. With creatinine clearance ≤30 mL/min/1.73 m2, the OR is 2.92. This finding is similar to the results of a retrospective case-control study that found a higher incidence in the group with end-stage renal disease.12 Matsumoto et al13,14 have reported that the clearance rate of LZD depends on creatinine clearance and that thrombocytopenia is concentration dependent, which suggests that the drug concentration may be elevated in patients with renal insufficiency, and the probability of thrombocytopenia is hence increased. Creatinine clearance <60 mL/min and/or urea nitrogen concentration >23 mg/dL indicate insufficient renal function due to drug-induced thrombocytopenia. Brier et al15 have reported that the levels of 2 in vivo metabolites of LZD, PNU-142586 and PMU-142300, in patients with renal insufficiency are 10-fold higher than in patients with normal renal function. However, the clinical significance of accumulation of these metabolites needs to be further studied.
Serum albumin concentration has been listed as one of the indicators because we have noticed that drug-induced thrombocytopenia rarely develops in patients who received human serum albumin during LZD therapy. That the serum albumin concentration is significantly negatively related to development of thrombocytopenia suggests that serum albumin may be a protective factor. The ROC results also indicated that serum albumin concentration ≤33.5 g/L is associated with a high incidence of thrombocytopenia. However, the underlying mechanism of specific protection by serum albumin needs further study.
The daily dose in the present study was determined on the basis of body weight. Because the total daily dosage of LZD in most patients was 1200 mg, the differences in daily dosage resulted from the variance in patients' body weight. Daily dosage is an independent risk factor closely associated with the incidence of thrombocytopenia, which suggests that LZD-related thrombocytopenia depends on drug concentration. Univariate analysis indicating that body weight is a protective factor against thrombocytopenia also supports this conclusion. The results of the ROC cutoff point suggest that thrombocytopenia is more likely to develop in patients receiving a daily dose of LZD ≥18.46 mg/kg and with body weight ≤65 kg. Takashi et al8 have also reported that a daily dose ≥22 mg/kg is an important risk factor for LZD-related thrombocytopenia. Abe et al16 have also indicated that body weight was an influential covariate on creatinine clearance and that the distribution volume was affected by body weight.
The recommended duration of LZD treatment is 10 to 14 days. One study has reported that prolonged use of LZD (>14 days) increased the risk of thrombocytopenia from 2.9% (36 of 1243 patients) to 4.1% (19 of 461 patients).17 Birmingham et al18 have also reported that LZD-associated thrombocytopenia is closely related to the duration of treatment. The incidence of thrombocytopenia was 1.9% when treatment duration was <14 days, 5.1% when treatment duration was between 15 and 28 days, and 7.4% when treatment duration was >28 days. The present study further confirmed the findings of these reports that medication duration is closely related to development of thrombocytopenia, and is an independent risk factor, as revealed at multivariate analysis. The ROC curve indicates that treatment duration >10 days is more likely to cause thrombocytopenia. Therefore, the frequency of clinical blood tests should be increased in patients receiving longer treatment, and prolonged and high-dose drug use should be avoided in patients at risk of bleeding.
Baseline platelet count has also been reported as a risk factor for thrombocytopenia. Whereas Grau et al9 reported that thrombocytopenia is more likely to develop in patients with baseline platelet count ≤240.7 × 109/L (8 of 18 patients [44.4%]) compared with that in other patients (4 of 31 [12.9%]) (P = 0.02), baseline platelet count was considered the only thrombocytopenia-related risk factor in another study.17 The present study has confirmed that baseline platelet count is an independent risk factor for severe thrombocytopenia, which indicates that baseline platelet count is closely related to LZD-induced severe thrombocytopenia, although it has little connection with the small decrease in platelet count. Similar to the results of previous studies, on the basis of ROC curve prediction, we found that patients with baseline platelet count ≤181 × 109/L are more likely than others to develop thrombocytopenia (45.9% [39 of 85] vs 17.8% [30 of 169]; P < 0.001; OR, 3.93; 95% CI, 2.24–6.90).
In the present study, concomitant use of anti-infective drugs and LZD in the hospital was investigated. These concomitantly administered medications included carbapenem (51.18%), β-lactam (48.82%), antifungal (43.31%), and fluoroquinolone (33.07%) agents. Nine anti-infective drugs, 7 of which may affect platelet count according to official recommendation, were analyzed as potential risk factors. Data indicated that levofloxacin, meropenem, and caspofungin could increase the risk of a small reduction in platelet count. In particular, levofloxacin and caspofungin are independent risk factors for LZD-induced thrombocytopenia. These findings suggest that concomitant use of levofloxacin, meropenem, and caspofungin with LZD may aggregate and enhance thrombocytopenia. Takahashi et al8 also indicated that concomitant use of fluoroquinolone is a risk factor for LZD-induced thrombocytopenia (P = 0.009; relative risk ratio, 1.48; 95% CI, 1.05–2.08). Therefore, more attention should be paid to the development of thrombocytopenia when LZD is used with other anti-infective medications, and selection of concomitantly used drugs should be dependent on the state of disease and drug characteristics.
In Japanese patients with renal dysfunction, LZD could cause adverse effects of severe thrombocytopenia.19 Moreover, Sasaki et al20 have reported that platelet profiles are not determined only by exposure to LZD and that there is substantial variability in individual sensitivity to LZD. In a study from the United States, the overall rate of thrombocytopenia in patients receiving LZD therapy was found to be higher than that in phase III clinical trials.21 In contrast to our study, the incidence of thrombocytopenia was low in patients receiving LZD in a matched cohort study performed at the Upstate New York Veterans Affairs Healthcare Network.22 Other reports have also indicated that the incidence of LZD-induced thrombocytopenia varies from 2.4% to 64.7%.20 Considered together, the differences between Chinese patients and patients of other races/ethnicities may contribute to the diverse effects of LZD in different patients.
Conclusion
The present study is the first analysis of risk factors for LZD-induced thrombocytopenia in a large Chinese population sample, and is the first to attempt to predict thrombocytopenia risk using the ROC curve. The incidence of LZD-related thrombocytopenia in our Chinese population was much higher than that suggested by the drug instructions. Low pretreatment platelet count, low body weight, low serum albumin concentration, long-term drug administration, advanced age, renal insufficiency, and concomitant use of caspofungin, levofloxacin, and meropenem are risk factors. Although there are predictors with good prospects for use in clinical practice to screen high-risk patients for intensified monitoring, further research on the dosage-based pharmacokinetics and pharmacodynamics of LZD are urgently needed. Recent studies have found that thrombocytopenia is positively correlated with plasma concentration of LZD,23 and we have established a method to determine the LZD concentration in human plasma. Therefore, according to pharmacokinetics, LZD dosage can be adjusted according to body weight to reduce the incidence of thrombocytopenia. Moreover, testing the effects of LZD metabolites on thrombocytopoiesis may contribute to understanding the mechanisms involved in thrombocytopenia.
Conflicts of interest
The authors have indicated that they have no conflicts of interest regarding the content of this article.
Acknowledgments
This work was supported by the Innovation Fund (10KMM41) of Chinese PLA General Hospital. Ms. Chen and Ms. Guo contributed to the study design, data colletion and analysis, and writing of this study. Mr. Cao contributed to the data intepretation. Dr. Cai contributed to the writing and revision. Mr. Xu, Ms. Zhu, and Mr. Ma contributed to the data collection.
References
- 1.Kaya O., Akcam F.Z., Temel E.N. In vitro activities of linezolid and tigecycline against methicillin-resistant Staphylococcus aureus strains. Microb Drug Resist. 2008;14:151–153. doi: 10.1089/mdr.2008.0801. [DOI] [PubMed] [Google Scholar]
- 2.Fung H.B., Kirschenbaum H.L., Ojofeitimi B.O. Linezolid: an oxazolidinone antimicrobial agent. Clin Ther. 2001;23:356–391. doi: 10.1016/s0149-2918(01)80043-6. [DOI] [PubMed] [Google Scholar]
- 3.Perry C.M., Jarvis B. Linezolid: a review of its use in the management of serious gram-positive infections. Drugs. 2001;61:525–551. doi: 10.2165/00003495-200161040-00008. [DOI] [PubMed] [Google Scholar]
- 4.Xu H.B., Jiang R.H., Li L., Xiao H.P. Linezolid in the treatment of MDR-TB: a retrospective clinical study. Int J Tuberc Lung Dis. 2012;16:358–363. doi: 10.5588/ijtld.11.0493. [DOI] [PubMed] [Google Scholar]
- 5.Attassi K., Hershberger E., Alam R., Zervos M.J. Thrombocytopenia associated with linezolid therapy. Clin Infect Dis. 2002;34:695–698. doi: 10.1086/338403. [DOI] [PubMed] [Google Scholar]
- 6.Orrick J.J., Johns T., Janelle J., Ramphal R. Thrombocytopenia secondary to linezolid administration: what is the risk? Clin Infect Dis. 2002;35:348–349. doi: 10.1086/341310. [DOI] [PubMed] [Google Scholar]
- 7.Niwa T., Suzuki A., Sakakibara S. Retrospective cohort chart review study of factors associated with the development of thrombocytopenia in adult Japanese patients who received intravenous linezolid therapy. Clin Ther. 2009;31:2126–2133. doi: 10.1016/j.clinthera.2009.10.017. [DOI] [PubMed] [Google Scholar]
- 8.Takahashi Y., Takesue Y., Nakajima K. Risk factors associated with the development of thrombocytopenia in patients who received linezolid therapy. J Infect Chemother. 2011;17:382–387. doi: 10.1007/s10156-010-0182-1. [DOI] [PubMed] [Google Scholar]
- 9.Grau S., Morales-Molina J.A., Mateu-de Antonio J. Linezolid: low pre-treatment platelet values could increase the risk of thrombocytopenia. J Antimicrob Chemother. 2005;56:440–441. doi: 10.1093/jac/dki202. [DOI] [PubMed] [Google Scholar]
- 10.Rubinstein E., Isturiz R., Standiford H.C. Worldwide assessment of linezolid's clinical safety and tolerability: comparator-controlled phase III studies. Antimicrob Agents Chemother. 2003;47:1824–1831. doi: 10.1128/AAC.47.6.1824-1831.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Topcuoglu M.A., Oguz K.K., Buyukserbetci G., Bulet E. Prognostic value of magnetic resonance imaging in post-resuscitation encephalopathy. Intern Med. 2009;48:1635–1645. doi: 10.2169/internalmedicine.48.2091. [DOI] [PubMed] [Google Scholar]
- 12.Wu V.C., Wang Y.T., Wang C.Y. High frequency of linezolid-associated thrombocytopenia and anemia among patients with end-stage renal disease. Clin Infect Dis. 2006;42:66–72. doi: 10.1086/498509. [DOI] [PubMed] [Google Scholar]
- 13.Matsumoto K., Takeda Y., Takeshita A. Renal function as a predictor of linezolid-induced thrombocytopenia. Int J Antimicrob Agents. 2009;33:98–99. doi: 10.1016/j.ijantimicag.2008.07.002. [DOI] [PubMed] [Google Scholar]
- 14.Matsumoto K., Takeshita A., Ikawa K. Higher linezolid exposure and higher frequency of thrombocytopenia in patients with renal dysfunction. Int J Antimicrob Agents. 2010;36:179–181. doi: 10.1016/j.ijantimicag.2010.02.019. [DOI] [PubMed] [Google Scholar]
- 15.Brier M.E., Stalker D.J., Aronoff G.R. Pharmacokinetics of linezolid in subjects with renal dysfunction. Antimicrob Agents Chemother. 2003;47:2775–2780. doi: 10.1128/AAC.47.9.2775-2780.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Abe S., Chiba K., Cirincione B. Population pharmacokinetic analysis of linezolid in patients with infectious disease: application to lower body weight and elderly patients. J Clin Pharmacol. 2009;49:1071–1078. doi: 10.1177/0091270009337947. [DOI] [PubMed] [Google Scholar]
- 17.Gerson S.L., Kaplan S.L., Bruss J.B. Hematologic effects of linezolid: summary of clinical experience. Antimicrob Agents Chemother. 2002;46:2723–2726. doi: 10.1128/AAC.46.8.2723-2726.2002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Birmingham M.C., Rayner C.R., Meagher A.K. Linezolid for the treatment of multidrug-resistant, gram-positive infections: experience from a compassionate-use program. Clin Infect Dis. 2003;36:159–168. doi: 10.1086/345744. [DOI] [PubMed] [Google Scholar]
- 19.Tsuji Y., Hiraki Y., Matsumoto K. Thrombocytopenia and anemia caused by a persistent high linezolid concentration in patients with renal dysfunction. J Infect Chemother. 2011;17:70–75. doi: 10.1007/s10156-010-0080-6. [DOI] [PubMed] [Google Scholar]
- 20.Sasaki T., Takane H., Ogawa K. Population pharmacokinetic and pharmacodynamic analysis of linezolid and a hematologic side effect, thrombocytopenia, in Japanese patients. Antimicrob Agents Chemother. 2011;55:1867–1873. doi: 10.1128/AAC.01185-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Minson Q., Gentry C.A. Analysis of linezolid-associated hematologic toxicities in a large Veterans Affairs medical center. Pharmacotherapy. 2010;30:895–903. doi: 10.1592/phco.30.9.895. [DOI] [PubMed] [Google Scholar]
- 22.Patel N., VanDeWall H., Tristani L. A comparative evaluation of adverse platelet outcomes among Veterans' Affairs patients receiving linezolid or vancomycin. J Antimicrob Chemother. 2012;67:727–735. doi: 10.1093/jac/dkr522. [DOI] [PubMed] [Google Scholar]
- 23.Hiraki Y., Tsuji Y., Hiraike M. Correlation between serum linezolid concentration and the development of thrombocytopenia. Scand J Infect Dis. 2012;44:60–64. doi: 10.3109/00365548.2011.608712. [DOI] [PubMed] [Google Scholar]