Discussion
In our cohort of 607 pregnant individuals who were at high risk for
preterm birth, we found that those who had a positive Ureaplasma culture
but did not receive treatment had comparable gestational age at delivery
compared to those who have negative Ureaplasma culture as well as those
who had a positive Ureaplasma culture and received treatment. Further,
the rate of treatment failure after the initial treatment was very high
(78.6%). We also found that cerclage was associated with lower odds of
recurrence compared to a history of preterm birth.
We found that treatment failure of Ureaplasma was very high even after
treatment of pregnant individuals and their sexual partners. This
finding was consistent with a previous study that showed that treatment
of pregnant individuals with preterm labor or PPROM with Azithromycin
resulted in a 93.3% treatment failure rate. Interestingly, studies of
intra-amniotic Ureaplasma infection showed that treatment of infection
resulted in a high rate of clearance of infection (79%). Given that
Ureaplasma species are normal genital flora, eradication of Ureaplasma
might be challenging. We found that pregnant individuals who had a
cervical culture due to cerclage had lower odds of recurrence compared
to those who had a cervical culture due to a history of preterm births.
We speculate that these lower odds could be due to abstinence from
intercourse after cerclage placement. In non-pregnant individuals,
sexual intercourse permits the spread of the various Ureaplasma species
with their different resistance profiles. In pregnant individuals, those
who had sexual intercourse 3 times or greater per week had a Ureaplasma
positive rate of 37%, which was much higher than 18% for those who had
intercourse less than 3 times per week (P=0.02). It is noteworthy that
even though our institution treated all sexual partners with
Doxycycline, the treatment failure rate remained high.
We found that giving appropriate antibiotics was not associated with an
improvement in gestational age at delivery. Despite the use of different
antibiotics, our finding was consistent with previous studies that
showed treatment of Ureaplasma was not associated with improvement of
adverse pregnancy outcomes compared to no treatment. For example, a
double-blind randomized control trial that randomized pregnant
individuals who tested positive for Ureaplasma Urealyticum to either
erythromycin treatment or placebo did not show significant differences
in gestational age at delivery, PPROM, or neonatal outcomes. Our study
is unique, however, since our protocol adopts an Azithromycin regimen
and treats their sexual partners.
It is important to note that giving antibiotics to pregnant individuals
could change bacterial flora and could have unknown consequences. Even
though Azithromycin is not associated with short-term neonatal adverse
outcomes, little is known regarding the long-term outcomes. A previous
study has indicated that antibiotics may be associated with an infant
gut microbiota imbalance during the first year of life. Future studies
that examine long-term neonatal and infants’ outcomes associated with
prenatal Azithromycin exposure would be useful.
Our study has many strengths. In 2014, our institution started a quality
improvement project that tests and treats pregnant individuals with a
positive Ureaplasma cervical culture as well as their sexual partners.
Our protocol was to repeat Ureaplasma cervical culture after the initial
treatment and treat again if treatment failure is noted. Though previous
studies treated pregnant individuals only once, we thought repeating
Ureaplasma cervical culture was important due to the high treatment
failure rate. We believe the quality of our data is high: data were
obtained from a chart review of outpatient, inpatient, and anesthesia
records. We think that this study sets a good example of a quality
improvement project; given the high rate of treatment failure despite
treatment of pregnant individuals and sexual partners, our institution
decided to stop routine Ureaplasma cervical culture for pregnant
individuals who were at high risk for preterm birth. Our study
underscores the importance of assessing the effects of a quality
improvement project and changing our practice based on study findings.
Our study is not without limitations. Given the retrospective nature of
the study, we were not able to assess the compliance rates of treatment
for pregnant individuals and their sexual partners. We tried to adjust
for covariates using cox or multivariable logistic regression models.
However, there could be residual confounding. Since this was a
retrospective cohort study, we were only able to assess association but
not causation. Our study was from a single academic center. Therefore,
our findings may not be generalizable in other clinical settings.
Unfortunately, due to our protocol of treating all pregnant individuals
who had positive Ureaplasma culture, the sample size of pregnant
individuals who had a positive Ureaplasma culture and did not receive
treatment was low. Since our study was powered to examine gestational
age at delivery, other analyses were exploratory. Lastly, we did not
differentiate the Ureaplasma species. It is possible that some species
of Ureaplasma such as Ureaplasma parvum but not Ureaplasma Urealyticum
are associated with spontaneous preterm birth.