METHODS

From June 1997 to April 1998, we conducted an unmatched case-control study to assess the impact of syphilis and HIV-2 on pregnancy outcomes in Bissau, Guinea-Bissau. The inclusion and exclusion criteria, the demographic information pertaining to the study participants, and the laboratory methods have been previously described.⁹ Briefly, women living in Bissau who gave birth or aborted at the Simao Mendes Hospital obstetrical ward were invited to participate in the study within 24 hours of delivery or abortion. Controls were women who had delivered a term neonate with a birth weight over 2500 g. Cases were classified into four groups of mothers according to the outcome of pregnancy: stillbirths, spontaneous abortions, premature deliveries, and small for gestational age (SGA) babies. Informed consent was obtained, and women were interviewed by midwives for demographic information and sexual, medical, and obstetric histories. They were examined by a physician, blood was collected for syphilis and HIV serology, and a cervical swab was obtained. Seven days later, when the mothers came back to obtain the results of the syphilis serology (and treatment when required), a vaginal swab to identify Trichomonas vaginalis (wet preparation) and a second cervical swab were obtained. The cervical swabs were stored at 4°C in Amplicor transport medium (Roche Diagnostic Systems) until transportation to the University of Sherbrooke, Canada, where the first and second swabs were pooled together and submitted for PCR detection of N gonorrhoeae.

The same pools of first and second swabs were later (after 24–28 months of storage) used for the detection of M genitalium, using a seminested PCR procedure adapted from Jensen et al.⁵ The details of the protocol used for amplification and the primers’ sequences have been described elsewhere.^2,3 Of the 1341 women who participated in the original study, cervical samples for PCR detection of M genitalium were available for the first 1014 women enrolled (June-December 1997), as the last specimens obtained were destroyed during the civil war that broke out in Bissau in June 1998. Data were entered on the epi-info 6.0 package and analysed with epi-info and stata 5.0. Proportions were compared with the χ² test or with Fisher’s tests if numbers were small. The sample size achieved was sufficient to detect a 2.3-fold increased risk of prematurity among women infected with M genitalium (β=0.2; α=0.05).

RESULTS

Among the 1014 women included in this study, 6.2% were infected with M genitalium. The prevalence of M genitalium according to demographic, behavioural, clinical or laboratory characteristics is presented in table 1 for the entire group of women for whom cervical samples were obtained (controls and all categories of cases). Among women in the control group (n=600), 6.0% were infected with M genitalium, and M genitalium infection tended to be more common in the presence of maternal HIV-1 infection (either single or dual): 20.0% (3/15) of HIV-1 infected women were infected with M genitalium as opposed to 5.9% (33/555) of HIV negative women (odds ratio 3.95; 95% confidence interval 0.68–15.61; p=0.06). The prevalence of M genitalium infection among controls and each category of cases is shown in table 2. M genitalium infection was not significantly associated with any of the adverse outcomes of pregnancy studied.

DISCUSSION

We identified M genitalium infection in 6.2% of women who had delivered or aborted at the national reference hospital of Guinea-Bissau. To our knowledge, this study is the first to report on the prevalence of M genitalium among pregnant women, using cervical samples. Through our case-control design, we were able to assess the impact of such infections on various outcomes of pregnancy. We did not find a statistically significant association between M genitalium infection and stillbirth, abortion, prematurity, and delivery of a SGA baby. Our findings extend on those of Lu et al who did not find an increased risk of preterm delivery in women with vaginal M genitalium infection at mid-trimester.⁸ Given our sample size, we can not rule out M genitalium having a weak effect on the outcome of pregnancy, as the upper limits of the confidence intervals for the odds ratios that we measured were between 2.5 and 3.0, but this would have little public health importance since this pathogen is found in only 6.2% of pregnant women in Bissau.

The role of other mycoplasmas (M hominis and Ureaplasma urealyticum) on the outcome of pregnancy has been studied by others,¹⁰ with inconsistent results owing to differences in study design and failure to identify and adjust for potential confounding factors. Furthermore, the ubiquity of these organisms (M hominis and U urealyticum colonisation is found in 5–49% and 43–81% of pregnant women, respectively)¹¹ and the high frequency of coinfection with other STD agents make it challenging to assess their own contribution to such disorders. Overall, there is little evidence implicating M hominis as a cause of adverse outcomes of pregnancy¹ but the potential role of U urealyticum in adverse pregnancy outcomes continues to be debated.^1,10,12 As expected for most genital pathogens, an animal model of U urealyticum infection suggested that if this organism was involved in adverse outcomes of pregnancy, it would probably do so by invasion of the upper reproductive tract of only a subpopulation of those colonised in the lower tract.¹⁰ This phenomenon might occur with M genitalium as well and could explain the absence of correlation with adverse outcomes of pregnancy in our study population since chorioamnion infection was not assessed.

We observed a trend for M genitalium being more frequent in HIV-1 infected mothers than among their seronegative counterparts in the control group. This association has also been observed in a smaller case-control study using PCR detection of M genitalium in urethral specimens of asymptomatic male patients.¹³ It deserves additional studies especially since M genitalium has been identified recently as a potential cofactor of transmission of HIV among American discordant couples.^4,14

In conclusion, although M genitalium is now considered an aetiological agent of male urethritis rather than an innocent bystander co-transmitted with a true pathogen,^1,2,4 M genitalium appears not to have a deleterious impact on the outcome of pregnancy. However, its role in cervicitis and PID, as well as its association with HIV, need further investigations.