Prevalence and Antimicrobial Susceptibility of Mycoplasma hominis and Ureaplasma Species in Nonpregnant Female Patients in South Korea Indicate an Increasing Trend of Pristinamycin-Resistant Isolates

Mycoplasma hominis and Ureaplasma species, commonly found in the lower urogenital tract, have been associated with various urogenital infections. This study aimed to estimate the prevalence and antimicrobial susceptibility trend of M. hominis and Ureaplasma sp. in female patients and to evaluate the risk factors for the acquisition of pristinamycin-resistant mycoplasma. Endocervical swab specimens obtained between March 2016 and December 2018 were analyzed using a Mycoplasma IST2 kit.

have reported that the incidence of tetracycline resistance associated with acquisition of the tet(M) determinant has been increasing (1,3,4).
The prevalence and antibiotic susceptibility profiles vary geographically and depend on the use of different antibiotics and the history of previous antibiotic exposure (5,6). It is important to identify the prevalence and antibiotic susceptibility profiles of genital mycoplasmas so that sufficient information is available when selecting appropriate empirical antibiotics and when performing antibiotic stewardship. This study aimed to investigate the prevalence and antibiotic resistance profiles of M. hominis and Ureaplasma species isolates from nonpregnant female patients in South Korea.

RESULTS
Detection of genital mycoplasmas. Out of the 4,035 samples, the prevalence of genital mycoplasma infection was 39.4% (1,589 of 4,035). Of the 1,589 samples with a positive culture, 49 (1.2%) had M. hominis only, 1,243 (30.8%) had Ureaplasma species only, and 297 (7.4%) had both M. hominis and Ureaplasma species. Thus, Ureaplasma species infection was significantly more prevalent than M. hominis infection. The results for the distribution of M. hominis and Ureaplasma species, according to age group, are presented in Table 1. Genital mycoplasma infections were more prevalent in younger patients (cutoff value, 56; P Ͻ 0.001).
Trends in the prevalence of genital mycoplasmas during the 3-year period between 2016 and 2018 are shown in Fig. 1. The prevalence of Ureaplasma species infection significantly decreased during the analyzed period (P ϭ 0.04). Although there was no significant difference, the prevalence of M. hominis infection increased during that period (P ϭ 0.08).
Antimicrobial susceptibility patterns. The analysis of the antimicrobial susceptibility patterns of M. hominis, Ureaplasma species, and both M. hominis and Ureaplasma  erythromycin. However, the susceptibility rates of both M. hominis and Ureaplasma species to pristinamycin decreased annually (100%, 97.1%, and 87.3%, respectively; P Ͻ 0.001). The susceptibility rates of M. hominis, Ureaplasma species, and both M. hominis and Ureaplasma species are described in Fig. 2.

Risk factor analysis for infection by pristinamycin-resistant genital mycoplasmas.
The baseline characteristics of patients who were infected with pristinamycinresistant genital mycoplasmas are presented in Table 3. No patient was infected by Trichomonas vaginalis in either group. In the univariate analysis using a logistic regression model, resistance to erythromycin, josamycin, and tetracycline, infection by Ureaplasma species, and coinfection with Candida species were risk factors for pristinamycin-resistant mycoplasma infection. In the multivariate analysis, josamycin resistance and coinfection with Candida species were considered independent risk factors (Table 4).

DISCUSSION
In this study, the prevalence of genital mycoplasma was 39.4% in symptomatic female patients. The prevalence of genital mycoplasmas did not show a significant decrease during the study period. Although direct comparison is difficult because there is no existing study assessing the trend of prevalence of genital mycoplasmas in symptomatic female patients in South Korea, Lee et al. reported that there was no significant difference in the prevalence of overall genital mycoplasmas between 2009 and 2013 every year in pregnant women (7). The prevalence of Ureaplasma species did not change significantly over the study period according to Lee et al.'s study, but it significantly decreased in our study. Interestingly, the prevalence of M. hominis and mixed infection increased significantly in both studies.
The prevalence of Ureaplasma species (30.8%) was higher than that of M. hominis (1.2%). According to several studies conducted in South Korea, the prevalence of Ureaplasma species in symptomatic patients was higher than that of M. hominis. The prevalence of Ureaplasma species and M. hominis was 21.3% and 2.9%, as reported by Moon et al. (8), 65.6% and 11.8% by Kweon et al. (9), and 48.8% and 25.3% by Jang et al. (10), respectively. Similar values were reported in Poland (11) and China (12).
Regarding antimicrobial susceptibility, the majority of genital mycoplasmas were most susceptible to pristinamycin, followed by josamycin, doxycycline, and tetracycline, but most of them were resistant to ciprofloxacin and ofloxacin. The majority of M. hominis isolates were resistant to erythromycin, clarithromycin, and azithromycin, whereas more than 85% of Ureaplasma species isolates were susceptible to these antibiotics. M. hominis is intrinsically resistant to C14-and C15-membered macrolides, for example, erythromycin, clarithromycin, and azithromycin, but susceptible to C16membered macrolides, for example, josamycin. Pereyre et al. reported that intrinsic erythromycin resistance of M. hominis was linked to the G2057A transition in domain V of the 23S rRNA sequence, and the high-level macrolide resistance might be associated with additional C2610U transition or the presence of a putative efflux mechanism (13). They also described that two josamycin-resistant isolates of M. hominis contained A2059G and C2611U mutations. Resistance to pristinamycin (these drugs are not available for clinical prescription in South Korea) was not frequently observed among both Ureaplasma species and M. hominis isolates but showed an upward trend during the study period. Resistance to doxycycline also showed an upward trend among Ureaplasma species isolates. Studies assessing the resistant trend for antibiotics in South Korea have not been conducted yet. Kasprzykowska et al. reported that there was no isolate with resistance for pristinamycin or josamycin among genital mycoplasmas between 2003 and 2015 (11). A study from Hungary reported that 10% of Ureaplasma species were resistant to josamycin from the specimen collected from men and women during 2 years from 2008 (14). However, they used another antibiotic susceptibility test kit. There are some in vitro studies investigating macrolide resistance in M. hominis (15) and Ureaplasma parvum (16) by the same team. They selected macrolide-resistant mutants of M. hominis and U. parvum by serial passages of M. hominis isolates in medium containing subinhibitory concentrations of macrolides, lincosamides, streptogramins, and ketolides. Selection of the pristinamycin-resistant M. hominis strains was performed with clindamycin, pristinamycin, quinopristin-dalfopristin, telithromycin, and josamycin. Selection of the pristinamycin-resistant U. parvum strains also was performed with erythromycin, josamycin, quinopristin/dalfopristin, and telithromycin.
In our age-and date-matched case-control study, resistance to josamycin and coinfection with Candida species were considered independent clinical risk factors for pristinamycin-resistant mycoplasma infection.
This study has some limitations. First, the Mycoplasma IST2 kit test was unable to differentiate U. urealyticum from U. parvum. U. parvum differs from U. urealyticum in its antimicrobial susceptibility; thus, inability to differentiate these species in a sample may lead to inappropriate reporting of antibiotic susceptibility (5). Second, antimicrobial susceptibility test using the Mycoplasma IST2 kit may not be compatible with the standardized guidelines, such as those of the Clinical and Laboratory Standards Institute or the European Committee on Antimicrobial Susceptibility Testing (5,11). Third, we could not conduct laboratory studies to find the mechanism of pristinamycin resistance of Ureaplasma species and M. hominis, because the specimens were discarded after they were tested with Mycoplasma IST2 kits. Fourth, there could be some selection bias in our case-control study.

Conclusions.
The results of this study provide important epidemiological data concerning the prevalence and antimicrobial susceptibility patterns of Ureaplasma species and M. hominis over the recent 3-year period. The analysis showed that (i) approximately 40% of symptomatic female patients may have genital mycoplasma infection, (ii) doxycycline and tetracycline are good treatment options, and (iii) antimicrobial susceptibility tests for patients showing genital candidiasis should be considered. More comprehensive and large-scale surveillance studies with standardized methodologies are required, especially when assessing the resistant trend of new macrolides.

MATERIALS AND METHODS
This study was conducted at H Plus Yangji Hospital, a 350-bed general hospital in Seoul, South Korea. This hospital performs genital mycoplasma culture, including antimicrobial susceptibility tests, with an endocervical or vaginal swab for patients who have specific symptoms or signs of genital tract infection or have abnormal Pap smear results during health screening. Clinical and microbiological data of female patients were collected from the database between 1 May 2016 and 31 December 2018. Patients older than 18 years were included in the study. Considering the large number of cases (more than 3,000 tests were performed annually) in this study, we selected the data of patients who were tested from May to July and November to December for each year. The following data were collected: age, results of genital mycoplasma tests, including the presence of M. hominis or Ureaplasma species, and results of antimicrobial susceptibility tests.
We also conducted a 1:2 age-and date-matched case-control study to identify the clinical risk factors for the acquisition of pristinamycin-resistant M. hominis and Ureaplasma species. Medical records were reviewed for the case and control groups. The following data were collected: age, sex, presence of specific symptoms, history of previous antibiotic administration within 3 months before examination, history of infection by M. hominis and Ureaplasma species within 3 months before examination, and presence of coinfection with Candida species or Trichomonas vaginalis. The specific symptoms were considered present when the patient described having abnormal discharge, dysuria, or other voiding symptoms, dyspareunia, bleeding between menstrual periods or after sexual intercourse, fever, and low abdominal pain (11). This study was approved by the Institutional Review Board of the H Plus Yangji Hospital.
A Mycoplasma IST2 kit (bioMérieux, Marcy-l'Etoile, France) was used for the detection, enumeration, identification, and antibiotic susceptibility testing for M. hominis and Ureaplasma species. Clinical specimens were inoculated in liquid transport medium R1 containing selective agents to inhibit the growth of contaminating flora in the sample. The samples in the R1 transport medium were centrifuged for 10 s and used to rehydrate the lyophilized selective growth medium R2. This medium was subsequently dispensed into 22 test wells, each well with a depth of 55 l, and two drops of mineral oil were overlaid on each compartment to prevent desiccation. The strips were incubated at 37°C for 48 h and observed for color changes. Positive results were observed when the color of the culture medium changed from yellow to red due to alkalization and when the estimated density of each organism was Ն10 4 CFU.
There were two concentration assay wells for each of the nine antibiotics (doxycycline, josamycin, ofloxacin, erythromycin, tetracycline, ciprofloxacin, azithromycin, clarithromycin, and pristinamycin). The development or absence of red color on the strip provided an index of the resistance or susceptibility, respectively, to each antimicrobial agent. The absence of red discoloration in either of the wells implied Mycoplasma sensitivity, whereas the presence of red discoloration in both wells signified Mycoplasma resistance. Mycoplasma was considered moderately susceptible to the antibiotic tested if the lowconcentration assay wells turned red. The breakpoints for the antimicrobials tested were the following: tetracycline susceptible (S), Յ4; resistant (R), Ն8; doxycycline S, Յ4; R, Ն8; azithromycin S, Յ0.12; R, Ն4; clarithromycin S, Յ1; R, Ն4; erythromycin S, Յ1; R Ն4; josamycin S, Յ2; R, Ն8; ciprofloxacin S, Յ1; R, Ն2; ofloxacin S, Յ1; R, Ն4; and pristinamycin S, Յ1; R, Ն2 (17).
Statistical analysis. Student's t test and the Mann-Whitney U test were used to compare continuous variables, and the chi-square test and Fisher's exact test were used for comparison of categorical variables. P values of Ͻ0.05 were considered statistically significant. To identify risk factors for the acquisition of pristinamycin-resistant M. hominis and Ureaplasma species, a logistic regression model was used to control for confounding variables. All P values were two-tailed. Variables that were statistically significant (P Ͻ 0.2) in the univariate analyses were included as candidates for multivariate analysis, in addition to the main variable of interest. The final logistic regression model was selected by stepwise backward elimination. Statistical analyses were performed using R, version 3.4.4.