Infectious Disease Society of America (IDSA) guidelines recommend two first-line agents for treating methicillin-resistant Staphylococcus aureus (MRSA), vancomycin and linezolid.1 Looking at pharmacokinetic studies, optimal vancomycin dosing is important in order for the drug to reach adequate concentrations in the lungs; MRSA strains with higher minimum-inhibitory concentrations (MIC) (i.e. MIC >1) require higher vancomycin levels due to its decreased killing ability.2 In comparison, linezolid is suggested to easily reach efficacious concentrations in lung tissue than in plasma, making it theoretically a more suitable drug for treating pneumonia.3,4 However, clinical studies of MRSA pneumonia have shown no differences between the efficacy of linezolid vs. vancomycin in regards to clinical or microbiologic cure as well as mortality.5 The only evidence that demonstrates linezolid to be superior to vancomycin was in the recent randomized, controlled study performed by Wunderink et al.6
Wunderink’s and colleagues randomized 1225 hospitalized individuals with documented nosocomial or health care-associated pneumonia, radiologic evidence of pneumonia, along with other accompanying findings associated with pneumonia (e.g. worsening cough, ausculatory pneumonia findings, fever, white blood cell count >10,000/mm^3, …). A modified intention-to-treat population was created by excluding 736 patients without a positive MRSA culture specimen. Further exclusions created a per-protocol population (PP) of 172 in the linezolid arm and 176 in the vancomycin arm. The study aimed to assess “clinical success” at both end of study (EOS) and end of therapy (EOT) along with mortality and safety outcomes. The study reported linezolid’s superiority in the PP population at EOS, 57.6% vs. 46.6% (P= .042), a similar 60-day mortality, and less nephrotoxicity (8.4% vs. 18.2%). To effectively analyze these results understanding the population, comparator regimen, and outcome definitions is essential.
Exclusion criteria which may negatively impact the study’s external validity include: “no clinical response recorded at EOT”, “no clinical response recorded at EOS”, and “lack of dosing compliance”. More patients were excluded for these reasons in the linezolid arm, thus favoring its efficacy. Another imbalance within the study arms can be seen within the baseline characteristics. Specifically, vancomycin’s population had more preexisting kidney conditions (27.9% vs. 36.9%), which can explain why vancomycin was seen to be more nephrotoxic.
In regards to the linezolid’s comparator, vancomycin is inadequately dosed. The authors suggest that their protocol improved on previous studies’ regimens, but when looking at the vancomycin trough levels, 15 to 20 mcg/mL as per IDSA guidelines1, therapeutic dosing as not reached until day nine. On day three, in which there was adequate time for a trough based dosing adjustment, the median trough level was 12.3 mcg/mL. This is especially important when a portion of MRSA specimens have a vancomycin MIC >1 (~15% of patients, this statistic was not provided by investigators, but calculated from the subgroup analysis).
Finally, outcome definitions are comprised of weak components. “Clinical cure”, the major efficacy outcome, can be simply interpreted as resolved signs and symptoms from baseline, a chest x-ray no worse ...