To the Editor—We thank Dr. Kenyon for his letter [1] in response to our recent study modeling the impact of human immunodeficiency virus (HIV) preexposure prophylaxis (PrEP) on the incidence of Neisseria gonorrhoeae(NG) and Chlamydia trachomatis among men who have sex with men (MSM) [2]. His letter questioned our model’s lack of inclusion of antimicrobial resistance (AMR) for NG, which may have led to overestimating the benefits of PrEP-specific biannual NG screening and treatment.

Resistant NG is a critical epidemiological issue. Reduced susceptibility to macrolides and cephalosporins has been rising, although it is still relatively rare in the United States: 2.5% for azithromycin, 0.8% cefixime, and 0.1% for ceftriaxone in the US general population, and approximately twice those levels for MSM [3]. On this basis, the Centers for Disease Control and Prevention (CDC) currently continues to recommend dual therapy with azithromycin and ceftriaxone [4].

Our model did not explicitly represent NG treatment regimens but assumed this standard dual therapy. We provided a sensitivity analysis (Table 2) varying the proportion of PrEP users who were successfully treated. Incomplete treatment there implied both clinical practice deficiencies and biological treatment failure for reasons such as AMR, although presumed NG treatment failures in clinical practice usually reflects partner reinfection [5]. We estimated that incomplete treatment would have to exceed 50% to reverse the prevention benefits of PrEP-related STI treatment, much higher than the levels of resistance currently observed.

Although not addressed to our study per se, Dr. Kenyon also argued that increased exposure to antimicrobials could induce higher levels of resistant NG. Although the selection pressures against many antibiotics increase with their use, there is limited empirical evidence of this occurring for cephalosporin-based treatment of NG. Several mathematical modeling studies have explored these complex issues: from biological and behavioral drivers of resistance [6], to mechanisms of fitness costs related to reduced duration of asymptomatic infection [7], to prevention strategies with point-of-care susceptibility testing [8]. But ongoing surveillance is needed to detect whether these predictions transpire.

The long-term goal of CDC’s NG treatment recommendations, for both PrEP users [9] and other MSM [4], is lower exposure to antimicrobials via disease eradication. Interventions that succeed in screening more MSM may temporarily result in higher drug use, but exposure should decline with incidence. NG epidemiology must also be understood in the context of HIV, which our model represented. NG increases the risk of HIV acquisition, particularly for asymptomatic rectal infections that often remain untreated [10]. If cephalosporin-based AMR rates increase, future models should evaluate the tradeoff between increased NG resistance with greater screening and increased HIV incidence with less screening.

We strongly agree with Dr. Kenyon’s call for more research to investigate AMR for NG in the context of PrEP interventions, but we would hypothesize that the lack of reduction in sexually transmitted infection (STI) incidence in open-label PrEP studies he pointed out is less likely due to AMR than reinfection from non-PrEP-using partners and overall increases in STI incidence over the past decade. Until further evidence emerges, there is insufficient evidence to warrant changes in the recommended NG screening and treatment of PrEP users or other MSM.