Thrombocytopenia with Tedizolid and Linezolid

ABSTRACT

Several studies have suggested the risk of thrombocytopenia with tedizolid, a second-in-class oxazolidinone antibiotic (approved June 2014), is less than that observed with linezolid (first-in-class oxazolidinone). Using data from the Food and Drug Administration adverse event reporting system (July 2014 through December 2016), we observed significantly increased risks of thrombocytopenia of similar magnitudes with both antibiotics: linezolid reporting odds ratio [ROR], 37.9 (95% confidence interval [CI], 20.78 to 69.17); tedizolid ROR, 34.0 (95% CI, 4.67 to 247.30).

TEXT

Linezolid is an oxazolidinone antibiotic that was approved by the Food and Drug Administration (FDA) in April 2000. Tedizolid, a second-in-class oxazolidinone antibiotic, was approved in June 2014. As tedizolid has only been on the market for 3 years, adverse events during real-world use are still being discovered and studied. Though thrombocytopenia is a known side effect with linezolid as noted in the warnings and precautions section of the label, the package insert for tedizolid does not include such a warning (1, 2). In a pooled analysis of two phase 3, double-blind, randomized comparator-controlled trials in patients with acute bacterial skin and skin structure infection (ABSSSI), rates of thrombocytopenia were found to be lower with tedizolid than with linezolid (3). Therefore, we sought to determine whether the rates of thrombocytopenia adverse event reporting were lower with tedizolid than with linezolid.

We reviewed adverse event reports from the FDA adverse event reporting system ([FAERS] July 2014 through December 2016) (4). Follow-up reports and reports missing all three categories of event date, sex, and age were excluded. Thrombocytopenia was assessed from the medical dictionary for regulatory activities (MedDRA) adverse event terms (“thrombocytopenia” and “thrombocytopenic”), and subsequent listings of adverse events were reviewed for inclusion. To evaluate adverse events reports for thrombocytopenia with linezolid before tedizolid was approved, we used AERSMine with a restricted time period (January 2004 to June 2014) (5). Using a case-noncase design, reporting odds ratios (RORs), proportional reporting ratios (PRRs), and corresponding 95% confidence intervals (CIs) were calculated with OpenEpi (6–8).

We included 1,995,573 adverse event reports from the tedizolid postapproval period (July 2014 to December 2016) (Table 1). Of these adverse events, 0.07% (n = 1,468) were thrombocytopenia. Among all adverse events, 0.02% (n = 408) were from linezolid and 0.002% (n = 41) were from tedizolid. Thrombocytopenia represented 2.70% (n = 11) of the adverse events for linezolid and 2.44% (n = 1) for tedizolid. The ROR for thrombocytopenia with linezolid was 37.9 (95% CI, 20.78 to 69.17) and with tedizolid was 34.0 (95% CI, 4.67 to 247.30).

TABLE 1.

Thrombocytopenia from adverse event reports

Medication	No. of adverse events	No. (%) with thrombocytopenia	Reporting odds ratio (95% CI)	Proportional reporting ratio (95% CI)
July 2014 to December 2016
All medications	1,995,573	1,468 (0.07)
Linezolid	408	11 (2.70)	37.9 (20.78–69.17)	36.9 (20.56–66.28)
Tedizolid	41	1 (2.44)	34.0 (4.67–247.30)	33.2 (4.79–230.10)
January 2004 to June 2014
All medications	5,772,958	45,701 (0.79)
Linezolid	9,267	795 (8.58)	12.1 (11.19–12.96)	11.1 (10.38–11.87)

From AERSmine (January 2004 to June 2014), we included 5,818,659 adverse event reports. Of these adverse events, 0.79% (n = 45,701) were thrombocytopenia, and among all adverse events, 0.16% (n = 9,267) included linezolid. There were 795 cases of thrombocytopenia reported with linezolid, resulting in an ROR of 12.1 (95% CI, 11.19 to 12.96).

Based on our analysis of adverse event reports from the real-world clinical use of linezolid and tedizolid, both antibiotics were associated with significantly increased risks of thrombocytopenia and these risks were of similar magnitude. Since it is recognized that adverse event reporting to FAERS is higher in the years following drug approval, we assessed two time periods, namely, the period since tedizolid approval and a 10-year period prior to the approval of tedizolid (9). In the tedizolid postapproval period, the risk of thrombocytopenia was more than 30 times higher with both tedizolid and linezolid than that with adverse event reports from other medications. In the period prior to tedizolid approval, the risk of thrombocytopenia was 12 times higher with linezolid (3, 10).

Rates of thrombocytopenia from previous clinical trials, though numerically higher with linezolid, were not statistically significantly higher than that with tedizolid (ESTABLISH-1 and ESTABLISH-2, NCT01170221 and NCT01421511 as registered at ClinicalTrials.gov) (11, 12). In the safety analysis set from ESTABLISH-1, substantially low platelet counts (<112,500 cells/mm³) were observed in 2.3% of patients in the tedizolid group (n = 331) and in 4.9% of patients in the linezolid group (n = 335) (11). It was noted that half of the patients with thrombocytopenia (11/22) also had hepatitis C. Though there were more patients in the linezolid group with hepatitis C (tedizolid, 101/329 [30.7%]; linezolid, 116/327 [35.5%]; P = 0.19), the distribution of hepatitis C plus thrombocytopenia by treatment group was not presented. Event rates were higher in ESTABLISH-2, as a higher threshold was used (<150,000 cells/mm³), but were still not significantly different (tedizolid, 27/314 [9%]; linezolid, 41/305 [13%]; P = 0.07). Again, hepatitis C was slightly more common in the linezolid group (tedizolid, 65/322 [20%]; linezolid, 80/321 [25%]; P = 0.15); however, the distribution of hepatitis C plus thrombocytopenia by treatment group was not presented.

A significant difference in thrombocytopenia was not observed until the data from ESTABLISH-1 and ESTABLISH-2 were pooled. Rates of thrombocytopenia (defined as both <150,000 cells/mm³ and <100,000 cells/mm³) were higher with linezolid during study days 11 to 13 but not during study days 7 to 9 (3). A limitation of this pooled analysis was that tedizolid therapy ended on day 6, while linezolid therapy ended on day 10. Additionally, this was a pooled analysis of clinical trial data, which does not confer the same benefits as a meta-analysis of randomized data (13). In the real-world clinical setting, the duration of exposure will be important to consider as comparative safety is assessed, particularly if the duration for either or both antibiotics is shorter than the duration used in clinical trials.

A possible explanation for the observed differences in thrombocytopenia from clinical trials may relate to the differences in metabolism and excretion with these two antibiotics (14). The majority of tedizolid is metabolized through the liver (82%), and less so through the kidneys (18%), while 30% of linezolid is excreted through urine (1, 2). Studies suggest that for linezolid, the risk of thrombocytopenia might arise from its increased exposure in renal-insufficient patients (15–17). Although, the involvement of linezolid's metabolites in thrombocytopenia is still unclear (16).

There are some limitations with our study. First, the true incidences of adverse events with tedizolid and linezolid are not known. Due the nature of FAERS reporting, which is not mandatory in all cases, the incidence cannot be estimated. Second, there were few events in both groups. As there was only one report of thrombocytopenia with tedizolid, the confidence intervals of the RORs and PRRs were large and though statistically significant, the magnitude of the lower ends of the confidence intervals did vary between linezolid (∼20 times higher risk) and tedizolid (∼4 times higher risk). Third, FAERS data are subject to different sources of bias, such as overreporting, underreporting, and missing information (18). Fourth, MedDRA terms were used to define thrombocytopenia, since platelet counts/changes were not available, and reports with other adverse event terms which did not specifically mention thrombocytopenia were not included. Lastly, FDA FAERS data were not available for the first several years after linezolid approval (April 2000 through December 2003); therefore, this initial time period after approval could not be assessed for linezolid.

Though several publications have suggested a lower risk of thrombocytopenia with tedizolid, using FDA FAERS data, we observed significantly increased risks of thrombocytopenia of similar magnitude with both linezolid and tedizolid. The incidence of thrombocytopenia in linezolid clinical trials was low, affecting only 2.2% of patients (19). Much higher rates were observed in real-world studies in the 2 years following drug approval, which ranged from 19 to 32%, with 47 to 48% of patients experiencing greater than a 30% reduction in platelet count (10, 18). Based on this previous experience, thrombocytopenia with tedizolid should be monitored and event rates should be assessed with real-world comparative safety studies as more patients are treated with tedizolid.