To evaluate the cost and safety of oral linezolid versus intravenous (IV) vancomycin for empiric treatment of suspected methicillin-resistant Staphylococcus aureus (MRSA) pneumonia.
This retrospective, single-center analysis evaluated consecutive cases of patients admitted between Jan. 1, 2016, and Oct. 31, 2017, who received either oral linezolid or IV vancomycin for suspected MRSA pneumonia. Patients in the linezolid group could have received initial doses of vancomycin if the therapy was limited to less than 72 hours prior to switching to linezolid. The primary endpoint was the total cost of MRSA antibacterial therapy per patient. Secondary outcomes included length of therapy, length of stay, 30-day readmission, and adverse effects.
Of the 29 patients meeting inclusion criteria, 14 received oral linezolid and 15 received IV vancomycin. Ten patients in the linezolid group were initially given IV vancomycin and were switched to linezolid. Patients who received linezolid incurred a lower average total cost compared to those who received vancomycin ($53.31 [95%CI 36.4–70.2) ] vs. $359.45 [95%CI 274.2–444.7), p < 0.00001) . Mean length of therapy in the linezolid group was numerically shorter (4.8 [95%CI 3.6–5.9] vs. 6.6 [95%CI 4.9–8.3] days, p = 0.07). Length of stay (16.4 [95%CI 6.8–26.1] vs. 15 [95%CI 7.4–22.6] days, p = 0.8) and 30-day readmission (1 vs. 3, p = 0.59) were similar between groups. Differences in mean platelet count (262.3 [95%CI 203.5–321.1] vs. 272.3 [95%CI 175.4–369.2] k/mm3, p = 0.85) and mean creatinine clearance at end of therapy (69.1 [95%CI 36.4–101.9] vs. 98.2 [95%CI 67.3–129.1 ml/min, p = 0.18) were not statistically significant.
Patients who received oral linezolid incurred less total cost for their empiric MRSA pneumonia therapy with a similar safety profile compared to those who received IV vancomycin. Larger-scale prospective research is warranted to further analyze the risks and benefits of initiating oral linezolid for this indication.
Methicillin-resistant Staphylococcus aureus (MRSA) accounts for approximately 20% to 40% of hospital-acquired and ventilator-associated pneumonia cases.1 Mortality can be as high as 55.5% in these patients, making empiric coverage critical in high- risk patients.1 The Infectious Diseases Society of America (IDSA) and the American Thoracic Society recommend combination antibiotic therapy that includes intravenous (IV) linezolid or vancomycin as first-line therapy for empiric treatment of MRSA in hospital-acquired pneumonia (HAP).2
Between these two recommended therapies, there has been evidence that IV linezolid has comparable or better results compared to IV vancomycin for the treatment of MRSA pneumonia. Previous studies suggested no significant difference in clinical cure rates or adverse events and better survival with IV linezolid.3,4 However, the acquisition cost of linezolid has historically remained a large barrier to its use as demonstrated by a decision model analysis study conducted in Spain, that demonstrated higher cure and survival rates but at greater expense with IV linezolid compared to IV vancomycin for the treatment of ventilator-associated pneumonia (VAP).5
This issue of cost may not be an issue with the oral form of linezolid, which has become more affordable due to its recent generic availability; however, evidence for using oral linezolid over IV antimicrobials for suspected MRSA pneumonia is sparse despite its favorable pharmacokinetic profile.6–8 With the 2015 FDA approvals of generic oral linezolid from Teva and Mylan, average wholesale price of oral linezolid dropped from $320 to $184 per 600 mg tablet, and the actual wholesale cost of the drug has continued to decline.
In addition to the acquisition cost savings, there may be further cost benefits to using oral linezolid because the total cost of using vancomycin includes several factors: drug levels, pharmacist monitoring time, potential for central line placement, or prolonged hospitalization to complete IV therapy. Studies suggest treatment with linezolid compared to vancomycin may be associated with shorter hospital length of stay and therefore lower hospital cost.6,9
To date, we know of no studies comparing the efficacy of oral linezolid and IV vancomycin for the treatment of MRSA pneumonia. Given the new economic feasibility of using oral linezolid and potential for significant cost savings with comparable clinical outcomes, we aimed to compare the cost and safety of using oral linezolid compared to IV vancomycin for the empiric treatment of MRSA pneumonia.
This retrospective, single- center observational study was conducted at a 643-bed academic medical center in Northern California. We evaluated cases of adult patients admitted between Jan. 1, 2016, and Oct. 31, 2017, and treated for suspected MRSA pneumonia with IV vancomycin or oral linezolid. The study was approved by the Institutional Review Board (IRB).
Pharmacists practicing on the general medicine service identified patients who were being treated for empiric MRSA pneumonia and recommended either initiating oral linezolid or switching therapy to oral linezolid. This intervention was implemented as a pilot in collaboration with the infectious diseases pharmacy team. Consecutive cases of patients receiving empiric MRSA pneumonia therapy with linezolid compared with patients receiving empiric vancomycin were included in the retrospective analysis. Inclusion criteria were as follows: age 18 years or older, suspected MRSA pneumonia based on clinical judgment of the treating physician, and receipt of oral linezolid or IV vancomycin. If patients had multiple courses of antibiotic treatment, only the first course was reported. Patients were excluded if they had complicated co-infections involving bacteremia or empyema, if they received 72 hours or more of IV vancomycin before initiation of oral linezolid, or if they had cystic fibrosis.
Data variables were collected retrospectively utilizing the institution’s electronic health record and include the following: antimicrobial received, length of stay, length of therapy, patient date of birth, number of vancomycin troughs taken, trough results, platelet count at beginning and end of therapy, changes in serum creatinine > 0.3 mg/dL and creatinine clearance at beginning and end of therapy. In our institution, the target serum trough concentration for vancomycin for MRSA pneumonia is typically 15–20 mcg/mL. Institutional policy requires that pharmacists monitor and adjust vancomycin to target concentrations.
The primary endpoint measure was the total cost in dollars of MRSA antibiotic therapy per patient. The following were included in the total cost calculation: the wholesale cost ($7.00/day for linezolid, $32.24/day for IV vancomycin (based on a dose of 1g q12h)) and vancomycin trough level laboratory cost ($110). If a patient received both vancomycin and linezolid, the cost of both were included in the total. Wholesale cost was used because this is consistent across wholesalers whereas Average wholesale price (AWP) overestimates the true cost of the drug for health care systems. Wholesale price more accurately represents the cost across different institutions.
Secondary endpoint measures included length of therapy, hospital length of stay, and readmission within 30 days. The following adverse effects were also included as a safety endpoint: renal dysfunction, thrombocytopenia, and signs and symptoms of serotonin syndrome. Renal dysfunction was defined as an increase in serum creatinine by more than 0.3 mg/dL at any time during the course of antibiotic therapy. Thrombocytopenia was defined as a decline of platelet count by 50% or greater and to a value of less than 150x109/L after initiation of antibiotic therapy.
Welch’s two sample t-test was used to describe baseline information of the two study groups as well as the primary and secondary endpoint metrics. Creatinine clearance was calculated using the Cockcroft-Gault equation and ideal body weight unless subjects were below their ideal body weight (actual body weight was used instead) or greater than 130% of their ideal body weight (adjusted body weight was used instead). Subjects on hemodialysis were assumed to have a creatinine clearance of 10 mL/min. For categorical variables, the chi-square test with Yates’ correction was performed. A p-value of less than 0.05 was considered significant. Microsoft Excel was used to generate graphs, and tables, and to calculate standard deviations for primary and secondary endpoint metrics. All other analyses were conducted using R software version 3.4.3 (R Foundation for Statistical Computing, Vienna).
Of the 29 cases that met inclusion criteria, 14 patients received oral linezolid and 15 received IV vancomycin. Nine patients (64.3%) in the linezolid group were initiated on vancomycin prior to linezolid therapy; six patients had been on vancomycin for one day, while three patients received two days of therapy. No statistically significant differences in baseline demographics were observed between the two groups (Table 1).
Two patients (14.2%) in the linezolid group had thrombocytopenia at baseline compared to five subjects (33%) in the vancomycin group (p = 0.39). Three patients (21.4%) in the linezolid group had end-stage renal disease on hemodialysis at baseline. Four patients (27%) in the linezolid group had a concurrent urinary tract infection (UTI) compared to three subjects (20%) in the vancomycin group (p = 0.67). UTI organisms included Escherichia coli (n = 1), Pseudomonas aeruginosa (n = 1), Citrobacter freundii (n = 1), Candida (n = 3) and vancomycin-resistant Enterococcus faecium (n = 1). Additionally, two patients (13%) had concurrent oral candidiasis in the linezolid group (p = 0.46, Table 1). Concurrent infection did not appear to affect duration of empiric MRSA coverage based on provider documentation.
Twenty-six patients (89.6%) received MRSA surveillance testing. Six patients (20%) had positive nasal swab cultures, one patient (3%) had a history of positive urine culture for MRSA, and one patient (3%) had a positive blood culture for MRSA.
Patients in the linezolid group incurred a lower average total cost based on wholesale cost than the vancomycin group $53.31 [95%CI 36.4–70.2) vs. $359.45 [95%CI 274.2–444.7), p < 0.00001) (Figure 2). A total of 20 trough levels were drawn for 15 patients in the vancomycin group, seven of which were supratherapeutic (Figure 3). An average of 1.3 trough levels were drawn for patients in the vancomycin group, equating to approximately $146.67 per patient for laboratory costs. None of the patients in the vancomycin group received a central line for outpatient therapy or had documented outpatient placement issues that prolonged their hospital stay.
Mean duration of therapy was numerically shorter (4.8 [95%CI 3.6–5.9] vs. 6.6 [95%CI 4.9–8.3] days, p = 0.07) in the linezolid group, but did not reach statistical significance (Table 2). One patient was discharged with a prescription for linezolid; we assumed this patient completed their outpatient therapy course. All readmissions were unrelated to infection except for one patient in the vancomycin group who was readmitted for MRSA pneumonia confirmed by sputum culture.
Mean platelet count (262.3 [95%CI 203.5–321.1] vs. 272.3 [95%CI 175.4–369.2] k/mm3, p = 0.85) and mean creatinine clearance (69.1 [95%CI 36.4–101.9] vs. 98.2 [95%CI 67.3–129.1 ml/min, p = 0.18) at the end of therapy were similar between the two groups. There were no incidents of renal dysfunction, and one patient in the linezolid group experienced thrombocytopenia after drug initiation. One patient in the vancomycin group developed a rash during therapy and was discontinued on antibiotics after five days of treatment. Serotonin syndrome was not reported for the two patients in the linezolid group with co-administration of serotonergic agents.
This is the first study to our knowledge to compare both the cost and clinical safety of oral linezolid versus IV vancomycin for the empiric treatment of suspected MRSA pneumonia and demonstrate significant cost savings with the use of linezolid. While the IDSA clinical practice guidelines for the treatment of MRSA infections support the use of oral linezolid for the management of pneumonia, many institutions have been reluctant to change their empiric guidelines away from vancomycin.10 These results may support the consideration of first-line use of oral linezolid for empiric treatment of suspected uncomplicated MRSA pneumonia in select patient populations.
The efficacy of intravenous linezolid was directly compared to vancomycin in a study that demonstrated linezolid was associated with improved 30- day mortality (0% vs. 41%, P = .02); however, patients in the vancomycin cohort had a higher severity of illness as measured by sequential organ failure assessment (SOFA) score, mechanical ventilation, and bacteremia.11 One meta-analysis found superiority of linezolid to vancomycin in terms of resolution of pneumonia (RR 1.18, 95% CI (1.10- - 1.27), p < 0.00001) and microbiological clearance (RR 1.39 (1.25- - 1.55), p < 0.00001) by the end of therapy.12
Our study focused mainly on a medical/surgical population, which likely had less severe or less complicated infections. One could therefore anticipate response to an oral agent would be similar or better, but when studied in critically ill patients with MRSA VAP, linezolid still appeared to be more effective.13
These results may be attributed to the limitations and complexity of dosing vancomycin. Despite the low cost of the drug itself, the overall cost of administration includes those associated with obtaining and monitoring drug serum concentrations as well as those associated with potential treatment of adverse effects, namely nephrotoxicity.14 In our study, we were only able to account for the direct costs of vancomycin ($32.24 per day) and the laboratory cost of obtaining drug concentrations ($110 per lab draw). Pharmacists in our institution provide all therapeutic drug monitoring for vancomycin, but we were unable to accurately quantify the costs of pharmacist monitoring time and nursing time to administer IV medications and draw labs. Thus, the cost of vancomycin reported in our study is underestimated. Furthermore, while our antimicrobial stewardship team also reviews orders for linezolid, the time spent reviewing patients for drug-drug interactions and adverse events is minimal compared to the time spent on patients receiving vancomycin who require lab ordering and chart documentation. With these considerations in mind, the true difference in cost between oral linezolid and IV vancomycin is likely to be even greater than what we identified.
Another limitation of vancomycin is the risk of nephrotoxicity, which is one of the most serious of its adverse effects and has been associated with prolonged hospitalization and increased mortality.15 The diagnosis of vancomycin-induced nephrotoxicity is also complex because there is no standardized definition, and patients can often sustain renal injury before the association is made.15, 16 We did not identify any incidents of vancomycin-induced nephrotoxicity; however, our sample size was small, the duration of exposure was relatively short (roughly six days), and we only evaluated serum creatinine at the conclusion of therapy. Of the patients who received vancomycin, five had a vancomycin trough greater than 20 mcg/mL, which put these patients at higher risk for developing this complication.17 Less than half (six patients) had trough values in the 10–20 mcg/mL range. More patients in the linezolid group had a lower average creatinine clearance than the vancomycin group, but this was not statistically significant. This effect may have been seen because patients with renal impairment were not ideal candidates for vancomycin therapy.
Considering the higher cost and limitations of administration in addition to adverse effects associated with vancomycin, oral linezolid remains an appealing treatment option for patients with suspected MRSA pneumonia. Estimated bioavailability per the medication’s package insert is 100%; although data for the use of oral linezolid for pneumonia is limited, some studies suggest a favorable pharmacokinetic profile even in patients with altered pharmacokinetics/pharmacodynamics. In one study that compared the pharmacokinetics of IV and oral linezolid in a small adult population of patients with cystic fibrosis, which has a known association with malabsorption, the average bioavailability of linezolid was still 85%, and standard doses of both formulations appeared to be sufficient against most MRSA isolates.18 One study evaluating lung penetration of linezolid demonstrated epithelial fluid lining levels of two to four times higher than the minimum inhibitory concentrations (MIC) established for Gram-positive organisms.19
With respect to clinical outcomes, one study evaluated switching from intravenous linezolid therapy to oral linezolid, and investigators found no difference in duration of therapy, reduction in C-reactive protein, reduction in platelet count, or readministration of anti-MRSA agent within 90 days; however, they showed a significant decrease in cost ($83/day vs. $305/day), which is comparable to what we demonstrated in our study. In this study, however, only 20 patients were treated for lower respiratory infection, with seven patients receiving the oral formulation. Another study that assessed oral and IV formulations of linezolid compared to IV vancomycin for the treatment of skin and soft tissue infections demonstrated a significant decrease in length of hospital stay and IV antibiotic use as well as higher microbiologic cure rates in the linezolid group.21 This effect was also demonstrated in the subgroup analysis of comparing only patients receiving oral linezolid to IV vancomycin.22
It is intuitive that length of stay would be shorter for patients who can be discharged on oral therapy, but this has also been demonstrated with IV linezolid in other studies. Tong and colleagues were able to demonstrate a significantly reduced hospital length of stay when comparing linezolid to vancomycin for MRSA pneumonia.23 A multinational, randomized trial by Li et al. demonstrated that the availability of oral linezolid shortened the length of hospital stay for those with evaluable MRSA infections, with a median length of stay two days shorter compared to the vancomycin group (14 vs. 16 days; p = 0.08).24 Length of stay in our study was comparable; however, length of therapy in was shorter in the linezolid group (4.8 [95%CI 3.6–5.9] vs. 6.6 [4.9–8.3] days; p = 0.07). Given our small sample size, it was likely underpowered to show statistical significance but may still have clinical significance. Additionally, in no instances did we recommend to extend vancomycin duration of therapy for MRSA pneumonia due to trough levels that below target range. As most patients resulted only one trough level, doses were adjusted accordingly and most patients were treated based on clinical response and the provider’s judgment.
Our study is not without its limitations. The single-center study design with small sample sizes limits the generalizability of this study. A larger cohort would be needed to achieve adequate power to show outcomes in hospital length of stay and duration of antibiotic therapy. Additionally, all patients in this study were treated empirically, which may make results less applicable to patients with culture-confirmed diagnoses. The practice of using MRSA nares swab as a screening tool to de-escalate antibiotic therapy was also not commonly done at the time of this study. So, the results of this study should be applied to the initial empiric treatment of patients with risk factors for MRSA pneumonia. Lastly, nine patients in the linezolid group were initiated on vancomycin before being switched to linezolid. Although the duration of vancomycin was less than 72 hours for all of these patients, it is difficult to discern what impact this had clinically.
Because of typical antimicrobial stewardship restrictions on linezolid, the applicability of our results may be limited. Linezolid, both oral and IV formulations, remains a restricted medication on our institution’s formulary. One reason for this restriction is to avoid the development of resistance via mutations in the 23S ribosomal subunit, or through acquisition of the chloramphenicol-florfenicol resistance (cfr) gene, both of which decrease the efficacy of linezolid and led to a linezolid-resistant MRSA outbreak in a Spanish hospital.25 Linezolid also has a broader spectrum of coverage when compared to vancomycin, further contributing to collateral damage and resistant organisms. Our institution’s antibiogram shows 100% susceptibility of MRSA to vancomycin and 99% susceptibility to linezolid (IV or PO) in non-ICU inpatients, which may support continued use of vancomycin in terms of efficacy. In areas or institutions in which rates of vancomycin-intermediate Staphylococcus aureus (VISA) or vancomycin-resistant Staphylococcus aureus (VRSA) are higher, linezolid may be a preferred option.
With respect to safety, patients in our study received a relatively short course of linezolid therapy because they had uncomplicated lower respiratory infections and no bacteremia. It is likely if patients were to have a more complicated diagnosis and require a longer duration of therapy, they would also be a higher risk of developing adverse events such as thrombocytopenia. Additionally, when selecting patients in which to utilize this approach, it is critical to evaluate for potential serotonergic drug interactions.
Linezolid appears to be a more cost-effective alternative to IV vancomycin therapy for the empiric treatment of suspected uncomplicated MRSA pneumonia with similar safety outcomes for this select patient population. Furthermore, linezolid may be associated with a shorter duration of therapy; however, this must be further investigated in larger studies. This pilot study may guide future research that includes more robust clinical outcome comparisons and cost evaluation. Future studies may investigate if these results can be replicated in definitive treatments of MRSA pneumonia.