RESUMO
OBJECTIVES: To investigate the relationship between MIC and clinical outcome in a randomized controlled trial that compared gentamicin 240 mg plus azithromycin 1 g with ceftriaxone 500 mg plus azithromycin 1 g. MIC analysis was performed on Neisseria gonorrhoeae isolates from all participants who were culture positive before they received treatment. METHODS: Viable gonococcal cultures were available from 279 participants, of whom 145 received ceftriaxone/azithromycin and 134 received gentamicin/azithromycin. Four participants (6 isolates) and 14 participants (17 isolates) did not clear infection in the ceftriaxone/azithromycin and gentamicin/azithromycin arms, respectively. MICs were determined by Etest on GC agar base with 1% Vitox. The geometric mean MICs of azithromycin, ceftriaxone and gentamicin were compared using logistic and linear regression according to treatment received and N. gonorrhoeae clearance. RESULTS: As the azithromycin MIC increased, gentamicin/azithromycin treatment was less effective than ceftriaxone/azithromycin at clearing N. gonorrhoeae. There was a higher geometric mean MIC of azithromycin for isolates from participants who had received gentamicin/azithromycin and did not clear infection compared with those who did clear infection [ratio 1.95 (95% CI 1.28-2.97)], but the use of categorical MIC breakpoints did not accurately predict the treatment response. The geometric mean MIC of azithromycin was higher in isolates from the pharynx compared with genital isolates. CONCLUSIONS: We found that categorical resistance to azithromycin or ceftriaxone in vitro, and higher gentamicin MICs in the absence of breakpoints, were poorly predictive of treatment failure.