The positive results for IST3 correlated well with those for IST2. The IST2-positive samples were also positive with IST3 in 93.8% of samples. The Up-positive samples (10.6%) in which Up failed to grow again, all contained initial concentrations of Up ≥ 10⁴ CCU as enumerated by IST2. Possible causes for this failure in growth may be the loss of viability during storage or the one-tenth dilution of the specimens. Considering that the manufacturer recommends refrigerated storage of samples suspended in R1 media for up to 48 h for IST2 and up to 72 h for IST3 [9,10], the 24-h storage is unlikely to have resulted in deterioration in the viability of Up or Mh. Therefore, sample dilution is more likely to be the cause of the observed failure in growth. When tested with IST2, Mh tended to be identified as mixed with Up, and less than half of these samples were cultured successfully by IST3. A previous study evaluating IST3 showed high sensitivity and specificity for both Up and Mh [12]. In that study, the sensitivity and specificity for Mh was 95.7% and 100%, respectively, compared to that of the standard culture using A7 agar. The results of the present study raise the possibility of false positives for Mh in IST2, because it is difficult to explain the high rate of growth failure with IST3. After introducing IST3 into the laboratory, the overall culture positivity rate remained at 24%–25% over the previous 3 months without significant changes; however, mixed infections positive for both Up and Mh among all positive samples decreased from 14.4% to 1.9%, and the proportion of Mh-positive samples also decreased from 15.5% to 1.9% (unpublished data). Polymerase chain reaction (PCR)-based kits detect genital mycoplasmas more than twice as frequently as with IST [13,14]. Up accounted for 30%, whereas Mh and Mycoplasma genitalium accounted for 3.1% of all PCR-positive sexually transmitted pathogens in Korea [13]. Considering the ratio of Up and Mh by PCR, a significant proportion of mixed infections of Mh and Up may be false positives for Mh owing to the high load of Up, which may overcome the inhibitory effects in Mh-enumerating cupules in IST2.

The enumeration of Up with IST2 was reproducible with IST3 as 82.2% of Up samples with a load ≥ 10⁴ CCU were also identified as having a load ≥ 10⁴ CCU in IST3. This result was comparable to that of a previous study that compared IST3 to standard agar culture and reported an agreement of 83.7% for M. hominis and 86.3% for Up [12]. The change in bacterial load in three Up-only samples from ≥ 10⁴ CCU to < 10³ CCU and growth failure in four Up samples can be explained by the sample dilution effect. To date, no studies have compared IST2 and IST3 in parallel, as most specimens are swabs transported in R1 medium, leaving no equivalent specimens for repeat testing. In the present study, the R1 broth refrigerated after inoculation was successfully used for parallel testing to compare the performance of the old and new versions of the IST assay kit. It is necessary to perform parallel tests using patient samples when a new culture system is introduced to assess the performance specifications of the new system in a clinical laboratory. The residual R1 transport broth is a reasonable alternative to the swab itself.

When the performance of IST3 for ASTs in comparison to that of IST2 was evaluated for tetracycline and macrolide susceptibility of Up-only samples, two mismatches were observed: one ME and one VME in tetracycline. The occurrence of one VME was defined by the categorical interpretation results of each of the two systems; however, the MIC results were not discordant because the tetracycline concentrations tested were different. The result of one ME was from a sample that was diluted from ≥ 10⁴ to ≥ 10³ CCU, and the tetracycline MIC changed from > 8 μg/mL to 2 μg/mL. This ME was probably caused by a decrease in the bacterial load, which was inadequate to detect tetracycline resistance. Therefore, parallel testing using the residual R1 medium carries the risk of VME owing to sample dilution when the bacterial load in the original sample is close to 10⁴ CCU. All five Mh samples tested were susceptible to tetracycline, which is consistent with the results of IST2. Tetracycline and doxycycline susceptibilities vary geographically [7,8], and in Korea, tetracycline susceptibility is reported to be 71.4%–100% and 81.0%–96.2% for Mh and Up, respectively [3,15-17]. The tetracycline susceptibilities in this study were consistent with those reported in previous studies. However, two erythromycin-resistant Up samples were found in two mixed infection samples with IST3. Macrolide resistance is often considered as a contributor to Mh in mixed cultures. Furthermore, one of the erythromycin-resistant Up samples was also resistant to telithromycin, which suggests acquired resistance to macrolide-lincosamide-streptogramin-ketolide class drugs, because telithromycin has not been used in Korea. Erythromycin is the primary drug for AST of Up, but macrolide resistance due to mutations in 23 rRNA and L22 ribosomal proteins is not rare in Korea [15]. The erythromycin AST is essential, especially in neonatal infections, as shown by a fatal case of erythromycin-resistant Ureaplasma meningitis in a premature infant treated with erythromycin in Korea [18]. The IST2 kit overestimates resistance in a mixed cultures containing Mh, which is inherently resistant to macrolides, and Up, which is inherently resistant to lincosamides [12]. Therefore, IST3 has a great advantage in that it can be used to independently test the AST of Mh and Up, even in mixed infections. IST2 also lacks the ability to determine whether the bacterial load, internationally set between 10⁴ and 10⁵ CCU, is compliant with the resistance threshold concentrations validated by the CLSI [11]. Therefore, nine VME and one ME of fluoroquinolones appeared to occur because IST2 used drugs that are less active than the standardized fluoroquinolones to test for levofloxacin and moxifloxacin. The susceptibility rates for ofloxacin and ciprofloxacin (17.1% and 2.8%, respectively) were similar to those of a previous study that collected the largest-scale data using IST2 over 3 years from 2016 to 2018 in Korea, showing susceptibilities of 18.7% and 6.0% for ofloxacin and ciprofloxacin, respectively [17]. Sequence analysis of gyrA, gyrB, parC, and parE detected amino acid mutations in only half of the ciprofloxacin-resistant Ureaplasma parvum isolates tested using IST2 [15]. Therefore, susceptibility of Up to fluoroquinolone seems to be biased toward greater resistance with IST2 [8,17]. Unlike conventional AST, which uses cultured bacteria, IST uses direct samples. Standardized AST for genital mycoplasmas is needed not only for patient care but also for monitoring the regional epidemiology of acquired antimicrobial resistance [17].