Probable Immortal Time Bias in Comparison of Daptomycin and Vancomycin for Methicillin-resistant <i>Staphylococcus Aureus</i> Bloodstream Infections

To the Editor—We read with interest the article by Schweitzer et al wherein [1] they conducted a large observational study and reported that in a multivariable analysis switching to daptomycin within the first 3 days of therapy was associated with reduced 30-day mortality (hazard ratio, 0.48; 95% confidence interval, 0.25–0.92). We believe the interpretation of that finding requires caution for 2 reasons. First, for patients with methicillin-resistant Staphylococcus aureus in the definitive daptomycin versus vancomycin randomized controlled trial, [2] the risk ratio for mortality with daptomycin was 1.43 (95% confidence interval, 0.65–3.2). The corresponding probability that daptomycin reduced mortality based on that trial data is 19% overall and only 7.5% for a risk ratio of 0.8 or lower. Although risk ratio and hazard ratio are not fully interchangeable, it is concerning that the results of this observational analysis are substantially different from the probability of that effect size based on the randomized trial. Second, there is an important risk of immortal time bias. although the authors are careful to exclude patients who die within the first 3 days, there remains asymmetry between the daptomycin and vancomycin groups in terms of time at risk. Specifically, the authors categorize patients as undergoing a daptomycin switch “if they received 3 or more doses of daptomycin treatment within 3–5 consecutive days.” Consequently, whereas vancomycin patients could die on any day after day 3, patients who were switched to daptomycin days 2 and 3 must live to days 5 and 6, respectively, or as long as up to day 8 if renally adjusted. It is unclear whether patients who received daptomycin before day 3 but died before their third dose were counted in the vancomycin group or excluded. We suggest this provides a varying degree of immortal time to the daptomycin group. Since many deaths due to Staphylococcus aureus bacteremia occur early, this seemingly small difference in immortal time could contribute to substantial differences in hazard ratio. In fact, the unadjusted χ 2 has a Fragility Index of only 3 [3]. Could the authors perform additional sensitivity analyses to control for this immortal time bias such as: (1) restricting the analysis to patients who have survived at least 6–7 days or (2) modelling daptomycin switch within the first 3 days as a time-varying covariate while removing the requirement for 3 or more days of use or (3) using a prevalent new-user cohort design [4]? Overall, daptomycin may have some advantages in terms of ease of dosing and nephrotoxicity. Some of the vancomycin nephrotoxicity will be mitigated with area under the curve–based dosing strategies [5]. With generic formulations available, the price difference may be less relevant than it was a decade ago; however, daptomycin stewardship might also help preserve susceptibility within the hospital biome. Overall, vancomycin remains the agent with the most experience, and an accurate real-world comparison of the 2, as envisioned by the authors, remains an important piece of work. Potential conflicts of interest. T. C. L. receives research salary support from the Fonds de Recherche du Québec—Santé. B. D. has nothing to declare. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.