Adjunctive linezolid versus clindamycin for toxin inhibition in β-lactam-treated patients with invasive group A streptococcal infections in 195 US hospitals from 2016 to 2021: a retrospective cohort study with target trial emulation
Ahmed Babiker, Sarah Warner, Xiaobai Li, et al.
Lancet Infect Dis 2024 Available online 10 October 2024
https://doi.org/10.1016/S1473-3099(24)00507-3
Summary
Background
Adjunctive clindamycin use is associated with survival in invasive group A streptococcus (GAS) infections but increasing clindamycin resistance in GAS has called into question its durability for this indication. Linezolid also inhibits GAS toxin and virulence factor production, but clinical efficacy data remain sparse.
Methods
We retrospectively emulated a target multicentre, non-blinded, non-inferiority trial to assess the efficacy of adjunctive linezolid compared with clindamycin in adult inpatients with invasive GAS infection treated with a β-lactam using the PINC AI database between 2016 and 2021. Patients were eligible if they had a monomicrobial GAS culture and received adjunctive therapy within 3 days of culture either concurrently or after β-lactam initiation and completed at least 3 days of β-lactam therapy. The primary outcome was adjusted risk ratio (aRR) of in-hospital mortality assessed by overlap-weighting using propensity scores. Secondary outcomes were length of stay among survivors and Clostridium difficile infection.
Findings
Of 1095 β-lactam-treated patients with GAS, 829 (76%) received clindamycin and 266 (24%) received linezolid. In the overlap weighted cohort, the receipt of linezolid was not associated with a statistically significant different aRR of in-hospital mortality compared with clindamycin (linezolid: 9·8% [26/266] vsclindamycin: 7·0% [58/829]; aRR: 0·92 [95% CI 0·42 to 1·43]; p=0·76). The risk difference was –0·005 (95% CI –0·05 to 0·04; p=0·81) and fell within the non-inferiority margin of 0·05. The primary analysis results were consistent across important subgroups and sensitivity analyses. Among survivors, median length of stay (adjusted ratio 0·96 [95% CI 0·16 to 0·08]; p=0·47) and C difficileinfection risk (aRR 1·76 [95% CI 0·37 to 1·75]; p=0·29) were not statistically significantly different between the two groups.
Interpretation
In this emulated trial of adult patients with invasive GAS infections treated with β-lactam, linezolid appeared non-inferior to clindamycin suggesting linezolid as an alternative for adjunctive antitoxin therapy.
Funding
The Intramural Research Program of the US National Institutes of Health Clinical Center and the National Institute of Allergy and Infectious Disease.
