No Effect of Pyridoxine on the Incidence of Myelosuppression during Prolonged Linezolid Treatment

Linezolid, the first antimicrobial in the class of oxazolidinones, is active against methicillin-resistant Staphylococcus aureus (MRSA), methicillin-resistant Staphylococcus epidermidis (MRSE), and Enterococcus species, and it is available in both intravenous and oral forms [1]. However, linezolid treatment can be complicated by anemia and thrombocytopenia [2, 3], especially when administered for >2 weeks. Nosocomial bone and prosthetic joint infections due to MRSA pose a significant management problem because of the lack of availability of active oral antimicrobials. The glycopeptides that are universally active against MRSA—vancomycin and teicoplanin—are only available in parenteral form and are poorly tolerated for the long duration necessary for the treatment of these infections.

Linezolid achieves adequate concentrations in bone and surrounding soft tissues [4] and is an attractive alternative for the treatment of these infections [5], but administration for >2 weeks is commonly complicated by anemia [2] and thrombocytopenia [3], both of which are reversible on discontinuation of linezolid treatment. Therefore, use of oral linezolid is limited in cases of bone infection, where it would be most needed.

Mitochondrial protein synthesis inhibition has been implicated in the pathogenesis of anemia caused by linezolid use [6]. Ringed sideroblasts have been observed in cases of linezolid-induced anemia, suggesting a possible pyridoxine (also called vitamin B6) responsive pathophysiological mechanism. On the basis of these observations, pyridoxine treatment was administered in 2 cases of mycobacterial infection being managed by long-term linezolid treatment; the result was reversal of anemia and thrombocytopenia [7]. We report the results of a study regarding the effect of pyridoxine supplementation on the incidence of myelosuppression in patients treated with linezolid for bone and prosthetic joint infection for a prolonged period of time.

Patients and methods. This retrospective study of patients' records included records of patients treated at the infectious diseases outpatient clinic of “Attikon” University General Hospital, a tertiary care center in Athens, Greece, from December 2004 to October 2005. The study was approved by the ethics committee of the hospital, and informed consent was obtained from each patient. Patients with chronic osteomyelitis or prosthetic joint infection due to MRSA, MRSE, or Enterococcus species, as well as patients in need of empirical antimicrobial treatment in cases where no pathogen was identified (in the instance that the patient either did not accept or was not tolerant of parenteral treatment with glycopeptides) received linezolid (Zyvoxid; Pfizer) at a dosage of 600 mg twice daily. On the basis of limited published findings [7], patients were offered supplemental treatment with pyridoxine (Besix; Cana; 125 mg daily). The planned duration of treatment was 6–12 weeks, depending on the type and severity of infection. Complete blood cell counts were recorded weekly in agreement with current recommendations, and linezolid treatment was discontinued if the hematocrit decreased to <30% or the platelet count decreased to <140 × 10⁹ platelets/L, according to local practice. Rates of discontinuation due to myelosuppression were compared among groups of patients using Fisher's exact test on the basis of a pretreatment hematocrit of <35%, underlying renal failure, and prior glycopeptide administration.

Results. A total of 24 patients received linezolid with pyridoxine. Baseline data are summarized in table 1. Mean duration of treatment was 4.59 weeks (median, 3.5 weeks; range, 2–14 weeks). Fifteen patients (65%) discontinued treatment because of adverse effects. Eleven patients (45.8%) presented with thrombocytopenia, and 6 (25%) presented with anemia. No effect of treatment on WBC counts was observed in our patients. Trends of platelet counts and hematocrit are presented in figure 1. The time to development of thrombocytopenia was 2–13 weeks (mean, 4 weeks). Time to observance of anemia was 2–8 weeks (mean, 4.16 weeks). A hematocrit of <35% at the time treatment began was significantly associated with discontinuation of treatment because of anemia (P = .006). One patient discontinued treatment at week 13 because of peripheral neuropathy. Treatment was discontinued because of myelosuppression by 4 patients with renal failure, compared with 7 of 19 patients without renal failure (P = .03). Only 2 patients had previously received vancomycin, and 3 patients had received teicoplanin in the month prior to linezolid administration; no correlation was observed between previous administration of these agents and discontinuation of linezolid because of myelosuppression. Regarding other potentially myelosuppressive medication, 4 patients received combination antimicrobial treatment with ciprofloxacin, 1 patient received a low–molecular weight heparin preparation, and 1 patient received a coumarin derivative. No episodes of bleeding due to thrombocytopenia or any complications of anemia were observed, and no patient received a blood transfusion for linezolid-induced anemia. Platelet counts and hemoglobin concentration returned to normal levels in all patients within 3 weeks after discontinuation of linezolid treatment.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1

Trends in platelet counts (A) and hematocrit (B) after institution of linezolid treatment. The horizontal dotted lines indicate the cutoff levels for thrombocytopenia and anemia. Red lines correspond to patients whose treatment was complicated by thrombocytopenia (A) or anemia (B), and blue lines indicate patients without respective toxicity.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Characteristics of patients at baseline.

Discussion. In this retrospective study, supplementation of treatment with pyridoxine did not prevent a high incidence of thrombocytopenia or anemia associated with linezolid treatment for a duration of >2 weeks. The reported incidence of thrombocytopenia following linezolid treatment has varied from 2.4% in phase III studies [8] to 32% in postmarketing studies [3]. In certain patient populations, such as patients with renal disease [9], a higher frequency of thrombocytopenia—up to 78.6%—has been reported. These results might be explained by the severity and type of underlying diseases, concomitant medication contributing to thrombocytopenia, and the varying duration of treatment. Other factors, including pharmacokinetic variability in the populations studied, cannot be excluded. In our patient population, these same factors, in addition to our conservative definition of thrombocytopenia as a platelet count <140 × 10⁹ platelets/L, might account for the high incidence of thrombocytopenia. Most studies define thrombocytopenia as a decrease in the platelet count to <75% of the lower limit of normal (which corresponds to 105 × 10⁹ platelets/L). However, the trend of a steadily decreasing platelet count without any signs of stabilization during linezolid treatment suggests that no major difference in the incidence of thrombocytopenia would ensue, even if a lower cutoff value was used. With all these factors taken into account, our results are in stark contrast with those of 2 studies involving patients treated with linezolid for bone infection—one of 45 patients with osteomyelitis [2] and another of 20 patients with prosthetic joint infection [10]—without any cases of thrombocytopenia, despite a mean duration of treatment of 15.9 weeks in the first study and a median duration of 7.2 weeks in the second study. In addition, the administration of linezolid to 10 patients with multidrug-resistant tuberculosis for 4–28 months resulted in only 1 case of thrombocytopenia and 5 cases of anemia [11]. This patient population, however, was characterized by younger age and fewer underlying conditions—factors that may account for the significantly different incidence and timing of myelosuppression.

The incidence of anemia was 25% among our patient population, in concordance with previously reported results [2]. There was a universal trend of decreasing hematocrit, and patients with low pretreatment hematocrit had a significantly increased risk of discontinuation of treatment because of anemia.

There was a significant increase in the rate of discontinuation among patients with renal failure, which corroborates previous findings [9]. Other underlying diseases may have influenced the incidence of myelosuppression, but the number of patients included in our study was not adequate to draw significant correlations. Concomitant myelosuppressive medication was not administered to our patients. Vancomycin and teicoplanin have been implicated as causes of thrombocytopenia [12]. Previous administration of these antibiotics did not correlate with discontinuation of linezolid because of either anemia or thrombocytopenia in our patients.

The limitations of this study are the retrospective and uncontrolled design and the small number of patients. However, we believe that the patient population here was similar to a typical population with bone and joint infections seen in clinical practice. Although an effect of pyridoxine in our patients cannot be excluded, a significant beneficial effect of pyridoxine supplementation on the incidence of myelosuppression appears to be rather improbable. A delay in the development of myelosuppression cannot be excluded, either. Pretreatment with pyridoxine before institution of linezolid treatment was not studied, but such a strategy is not easily applied in clinical practice.

Our results contrast with the favorable outcome of the 2 patients whose anemia and thrombocytopenia were reversed after pyridoxine administration [7]. However, apart from the limited number of patients, other factors might account for the observations of this report. Anemia might have been caused by chronic disease, and response in at least 1 patient could have been partly attributed to recombinant erythropoietin administration.

In conclusion, these results do not support the coadministration of pyridoxine for patients receiving linezolid for >2 weeks to prevent linezolid-induced anemia and thrombocytopenia. Strict weekly monitoring of blood cell counts and prompt discontinuation of treatment if anemia or thrombocytopenia occur are appropriate measures for this patient population.

• Received April 6, 2006.
• Accepted June 30, 2006.

• © 2006 Infectious Diseases Society of America