Background Treatment failure due the development of macrolide resistance seen at Melbourne Sexual Health Centre (MSHC) has risen from 15 to 46% over the last six years (p < 0.01). Macrolide resistance is conferred through point mutations in the M. genitalium 23S rRNA gene. Treatment failure with 1g azithromycin is either due to an infection with a pre-existing resistant strain or development of resistance during treatment, but the mechanism is not yet fully understood.

Methods A subset of M. genitalium positive cases seen at MSHC between 2007–9 (n = 67) and 2012 (n = 70) underwent detection of resistance mutations via high resolution melt analysis, with real-time PCR quantifying M. genitalium infectious load.

Result Of those M. genitalium cases that successfully responded to 1g azithromycin, the pre-treatment median loads for 2007–9 (n = 40) and 2012 (n = 48) were 8.5 x 102 and 1.7 x 103 copies per reaction respectively. For M. genitalium strains that possessed resistance in the pre-treatment sample, the loads were remarkably similar to those successfully treated, with 2.2 x 103 copies per reaction detected for 2007–9 (n = 9) and 5.7 x 103 copies for 2012 (n = 22). However, for M. genitalium cases that appeared to develop resistance following treatment, (ie. mutations only detected in follow-up test of cure sample), there was a significantly higher load detected with 3.1 x 104 copies per reaction for 2007–9 (n = 8) and 1.8 x 104 copies for 2012 (n = 8), when compared to either treatment success cases or those with baseline resistance (one sided p < 0.01).

Conclusions The higher infectious load in pre-treatment M. genitalium cases that developed detectable resistance after 1g of azithromycin compared to those with baseline resistance and those cured raises the possibility that heterotypic resistance and/or induced resistance may be contributing to macrolide failure in M. genitalium. These findings have implications for current recommended treatment for M. genitalium.