Is It Time for the United States and Canada to Reconsider Macrolides as the First-line Empiric Treatment for Males With Symptomatic Urethritis?

(See the Major Article by Li et al on pages 805–10.)

Mycoplasma genitalium (MG), a fastidious, slow-growing bacterium, was first isolated as a human pathogen in males with nongonococcal urethritis (NGU) in 1980 [1]. A meta-analysis of 27 studies up to 2010 confirmed that MG is strongly associated with NGU and has been detected in up to 40% of men with persistent and recurrent urethritis [2, 3]. It has also been associated with female reproductive tract disease and sequelae [4]. In highly developed countries, the prevalence of MG in the general population (typically low-risk, asymptomatic individuals) is an estimated 1.3% (95% confidence interval 1.0–1.8%), with higher prevalences (3.2–3.7%) among men who have sex with men and in countries with lower development (3.9%) [5]. MG is now recognized as an important sexually transmitted infection, second in population-level prevalence only to Chlamydia trachomatis (CT) [5]. The disease symptoms are indistinguishable from CT [6], and access to diagnostic testing is limited in many settings. This often results in syndromic management for CT, with antibiotics which may not be optimal for MG infections. Unfortunately, MG has demonstrated a remarkable propensity to develop resistances to nearly all antimicrobials used to treat this infection. Due to the lack of a bacterial cell wall, few antimicrobial classes have activity against mycoplasmas, and include tetracyclines, macrolides, and fluoroquinolones (FQ) [7].

In this issue, Li et al [8] report findings from the largest study to date of the etiology of males with symptomatic urethritis. They determined the prevalence of MG and other sexually transmitted infections (STIs) in 1816 heterosexual men with symptomatic urethritis in Nanjing, China; estimated the prevalence of resistance to macrolides and FQs; and evaluated the impact of recent antimicrobial use on the distribution of MG in single or mixed infections. The overall MG prevalence was 19.7%, with 54% of infections occurring in the absence of other STI agents. This was lower than the prevalence of either Neisseria gonorrhoeae (NG; 46.4%) or CT (33%), although the high prevalence of NG may be related to the parent study’s goal to evaluate gonococcal transmission. In addition, males with single MG infections were more likely to have been treated with antibiotics that were either self-administered or prescribed in the 30 days prior to enrollment, potentially signaling treatment failure. Overall, 89% of MG-positive specimens possessed mutations in the 23S rRNA gene, and this was higher among males who had taken macrolides in the prior 30 days, compared to those who had not (96.7% vs 86.8%, respectively; P = .026). The single nucleotide polymorphism macrolide resistance–mediating mutations in region V of the 23S RNA gene were first reported in 2008 and are highly correlated with azithromycin treatment failure [9]. Both single-dose and extended regimens of azithromycin may select for macrolide resistance [10, 11], and this occurs approximately 10% of the time [12]. In contrast to other studies, which have reported rising resistance to macrolides in MG over time [13], this study found no increase in macrolide resistance over a 5-year period. However, the prevalence of the 23S mutations at the beginning of the 5-year observation period was high (88%), and there was little room for increase. Macrolide resistance rates vary significantly geographically, but ranged from 50–88% in 2016–2018 in studies from Australia, the United States, and Canada [14–19]. This is unacceptably high.

Adding to the concern over macrolide resistance is the fact that our ability to use moxifloxacin, recommended for the treatment of macrolide-resistant MG, is now being challenged by rising rates of resistance to FQ, particularly in Asian and Western Pacific regions [7]. Li and colleagues [8] reported an astonishingly high prevalence of mutations associated with FQ resistance (89% had parC mutations; 12.4% had gyrA mutations). Prior to this, the highest reported prevalences of mutations in parC were 37–40% in Japan and the United States [19–21]. The prevalences of FQ-associated mutations in the United Kingdom, Australia, and Canada have been lower, ranging from 5% to 15% [17, 18, 22], although this is still concerning. The higher prevalence of antibiotic resistance in Asian and Western Pacific regions has been attributed to multiple factors, including the misuse, overuse, inappropriate selection, unrestricted access, suboptimal quality, and dosing of antibiotics [7]. In this era of rapid global travel, spread from these areas is highly likely, again raising concern over the high prevalence of resistance markers in the Li et al study [8].

Some of the concern over the high prevalence of FQ resistance markers has been mitigated by observations that not all mutations in the parC and gyrA genes have been strongly associated with resistance and treatment failure [21]. Unfortunately, in the Chinese men with urethritis studied by Li et al [8], the S83I mutation, which is most strongly associated with resistance and treatment failure, was present in 84% of participants. Among Australian patients with the S83I mutation, treatment failure after moxifloxacin occurred in 4/7 (57.1%) participants [23]. From an optimistic perspective, 43% of MG infections with this mutation may still be effectively treated with moxifloxacin. However, from a pessimistic perspective, an estimated 165 moxifloxacin treatment failures would have occurred among the Chinese men in the Li study [8]. Globally, our arsenal of alternatives is limited. Pristinamycin was initially effective for moxifloxacin treatment failures, but it is not available in many countries and resistance developed rapidly once it was implemented as a third-line therapy in Melbourne [24]. Some older drugs, such as minocycline and spectinomycin, have been effective in limited cases, but also are not uniformly available [20, 25]. Without the creative use of existing antimicrobials and/or the development of new antibiotic compounds, the number of untreatable MG infections is likely to far outweigh the number of untreatable gonorrhea infections.

The high prevalence of MG infections and macrolide resistance in men with symptomatic urethritis, along with evidence that doxycycline is likely more effective than azithromycin for the treatment of rectal chlamydia infections (and possibly for urogenital CT infections) [26, 27], has called into question the routine use of macrolides (single-dose azithromycin) for the empiric treatment of NGU. Recent British, European, and Australian guidelines for the management of NGU now recommend against the routine use of azithromycin as the first-line treatment for NGU, and instead recommend doxycycline at 100 mg orally twice daily for 7 days [28–30]. Current Centers for Disease Control and Prevention and Canadian guidelines, however, continue to recommend either azithromycin or doxycycline regimens for the empiric treatment of NGU [31, 32]. Despite the higher costs of azithromycin, compared to doxycycline, the cited rationale for using azithromycin as the preferred agent includes the higher rates of compliance with directly observed, single-dose therapy that can be dispensed in a clinic setting [31, 32]. Although some public health sexually transmitted disease clinics in the United States now recommend doxycycline as a first-line therapy, this is unusual and azithromycin is likely to continue to be used by many until official guidelines change. With this continued use will certainly come the further expansion of macrolide resistance.

Transitioning to doxycycline as a first-line therapy would affect MG in a nonintuitive way. Despite in vitro data suggesting that most MG strains are susceptible to doxycycline, treatment with doxycycline results in only 22–45% microbiological cure rates [7]. The reasons for this are unclear, but may be related to inappropriate break points for susceptibility to doxycycline in MG [7]. However, pretreating patients with doxycycline may be beneficial, as highlighted by Read and colleagues [30], who reported the sequential treatment of MG, using doxycycline as the first-line agent for the treatment of NGU and then providing subsequent treatment based on the presence or absence of macrolide resistance. They reported curing >92% of infections and the infrequent selection of macrolide resistance. This resistance-guided therapy approach not only reduced the overall use of azithromycin, but also reduced the MG bacterial load, which is hypothesized to contribute to the high proportion cured and the low rate of selected macrolide resistance [30]. This initial success has sparked considerable enthusiasm for extending the approach to FQ resistance testing. In order for this to be successful, however, further work is needed to determine which mutations are critical for FQ resistance. This will require careful, longitudinal studies that measure a wide spectrum of mutations and treatment outcomes and yield clinical isolates on which minimum inhibitory concentration can be performed. In the meantime, it will be critical to preserve our current antimicrobials. Reducing the use of azithromycin is key to achieving this.

Finally, while this study adds to the body of literature supporting the removal of macrolides as empiric, first-line treatments for symptomatic NGU, additional questions remain regarding diagnostic testing for MG. Until recently, access to testing for MG has been extremely limited in the United States and Canada. The first Food and Drug Administration–approved assay is now on the market in the United States, and access to testing is gradually expanding. However, even where testing is available, the results of testing for macrolide and FQ resistance may not be available or may only be available with long turnaround times. In an ideal world, initial testing for urethritis will consist of tests for NG, CT, and MG, with resistance testing available at the point of care. Point-of-care assays for MG that incorporate resistance testing are already in development [33] and additional studies on sequential treatment for MG are underway [30]. This will be essential to reduce antimicrobial resistance and optimize global treatment guidelines, including in the United States and Canada, for this important clinical syndrome.

Note

Potential conflicts of interest. L. M. has received donations of test kits, reagents, and equipment for related research studies and has received consulting fees from Hologic, Inc. 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.