The role of bacterial vaginosis, aerobic vaginitis, abnormal vaginal flora and the risk of preterm birth

Sir, Donders et al.1 add some insight into the role of bacterial vaginosis (BV), aerobic vaginitis (AV), abnormal vaginal flora (AVF) and the risk of preterm birth (PTB). The earlier in pregnancy that PTB occurs, the greater this is likely to be as a result infection. The earlier in pregnancy AVF is diagnosed, the greater is the risk of an adverse outcome, even if this resolves, suggesting that whatever infection/inflammatory damage that occurs as a result of AVF, occurs early and persists. If antibiotics are to reduce PTB, they should be: active against organisms associated with PTB, used in women with demonstrable AVF, and used early in pregnancy before infection/inflammation can cause irreversible fetomaternal tissue damage. The study was undertaken 8–9 years ago, and women who were chlamydia-positive on the Chlamydiazyme test were excluded from the analysis. The test is only about 50% sensitive compared with current molecular techniques. Accordingly, some women regarded as chlamydia-negative, were positive and remained in the study. The authors state that in studies in which strict criteria had been applied for the diagnosis of BV, there was no correlation with PTB, yet we are aware of well-respected studies, for example Hay et al.,2 in which strict diagnostic criteria were used, and in which a five-fold increased risk of PTB was demonstrated when BV was detected early (i.e. with a maximal gestational age at screening of 20 weeks). The authors state that ‘intermediate flora’ (scores of 4–6 on the Nugent Gram-stain) does not correspond with ‘intermediate BV’. We take ‘intermediate BV’ to mean ‘partial BV’, described by the authors as ‘patchy streaks of BV flora or sporadic clue cells mixed with other flora’. There is a distinct order in which organisms of different species increase in number according to the grade of flora on Gram staining. Gardnerella vaginalis and Mycoplasma hominis, for example, have maximal numbers in the grade-III category (i.e. BV) compared with the intermediate grade II. The latter is a distinct entity, and its amalgamation with grade III diminishes the sensitivity and positive predictive value of Gram-stain scoring, and women with this diagnosis fail to respond to clindamycin, unlike women with grade III.3 Finally, the authors feel that the finding of an association between M. hominis and PTB was remarkable. They seem to regard the existence of M. hominis as a separate entity, whereas, in large numbers, it is a feature of BV. A large number of organisms is the key, yet Donders et al., despite referring to a quantitative method of detection, do not mention the numbers of organisms present. Some 22 years ago it was demonstrated that M. hominis organisms in large numbers [≥ 105 ccu (colour change unit – a unit to measure the quantity of M. hominis present)] were isolated from 18% of women admitted in spontaneous preterm labour between 26 and 34 completed weeks of gestation, compared with 0% from women not in labour, but delivered electively at the same gestational age for fetomaternal indications (P < 0.05).4 Our perception of the role of various microbes and how they influence the changes that occur in BV has been widened by the recent identification of hitherto unknown vaginal bacteria that are associated with BV.5 The relation between the ‘new’ bacteria and the criteria espoused by Donders et al., based on vaginal wet preparations, remains to be seen.