This study was conducted to determine the relationship between lactobacilli and bacterial species associated with bacterial vaginosis in pregnancy and the prevalence of H2O2-producing and non-producing strains of lactobacilli in pregnant women whose vaginal flora had already been analysed. Information was available for 174 pregnant women whose vaginal flora had been evaluated previously by examining gram-stained vaginal smears: 50 had grade III flora (bacterial vaginosis). 50 grade II flora, 41 flora graded as abnormal which then reverted to grade I (revertants) and 33 normal flora (controls). Lactobacilli were isolated from 19 of 50 women whose vaginal flora was grossly abnormal culturally and categorised as grade III by Gram staining. In 6 of these 50 women lactobacilli were isolated in large numbers, i.e. 10(5)-10(6) cfu/ml. H2O2-producing strains of lactobacilli were isolated from 11 of 12 women with grade III flora who were randomly selected from this group. Thus, in those 11 women it appears that H2O2-producing lactobacilli had not protected them from developing bacterial vaginosis. Bacterial species associated with vaginosis were isolated in high numbers from a large proportion of women in the revertant and grade II groups in association with high counts of lactobacilli. Thus, in some women it is possible that a change to an abnormal flora could occur before the complete disappearance of lactobacilli. It is concluded that bacterial vaginosis may develop in some women despite the presence of H2O2-producing strains of lactobacilli and that other factors, as yet unidentified, might be conducive to the appearance of abnormal bacterial flora with progression to vaginosis.

A quantitative method was developed for the measurement of micromolar quantities of H2O2 produced in Rogosa broth and peptonized milk broth by vaginal strains of lactobacilli isolated from women. The production of substantial amounts reproducibly was dependent on the growth of the organisms in acid media (pH less than or equal to 6.0) under anaerobic or micro-aerophilic conditions with continuous agitation. The addition to the media of the enzyme inhibitor, 3-amino-1,2,4-triazole, with or without catalase sometimes induced the production of H2O2 especially in non-agitated cultures. However, other agents such as concanavalin and o-dianisidine had no enhancing effect, and catalase or peroxidase alone completely inhibited H2O2 production. The H2O2 produced in the acid media was stable for more than a month at 5 degrees C but not in media at pH greater than or equal to 7.0. Of five strains of lactobacilli tested by the quantitative method and by a chromogenic qualitative method (Rogosa-catalase or -per-oxidase agar), three consistently produced H2O2 measurable by the former method, but none did so after growth of the organisms on Rogosa-catalase/peroxidase agar which suggested that the qualitative method was unreliable. The fact that H2O2 was produced in substantial quantities by some strains and not at all by others enabled H2O2-producers and non-producers to be distinguished easily.

SUMMARY Although up to about halfthe cases ofacute non-gonococcal urethritis (NGU) are caused by Chlamydia trachomatis organisms (chlamydiae) and a smaller, ill-defined, proportion probably by Ureaplasma urealyticum organisms (ureaplasmas), the aetiology of all cases is not understood. Clarification of the role of the anaerobe, Bacteroides ureolyticus, was sought in the current study. Seventy five chlamydia negative patients with NGU were treated on a double blind placebo controlled basis with metronidazole. After seven days more of the 35 patients given this drug tended to improve clinically than the 40 given the placebo, but the difference was not significant. Of 23 chlamydia negative but anaerobe positive men, however, 78% (7/9) receiving metronidazole responded clinically, but only 7% (1/14) receiving placebo responded (p<0 001). Furthermore, whereas 78% of the anaerobe positive men given metronidazole recovered, only 23% (6/26) of the anaerobe negative men did so (p < 0.02). No further evidence for the role of ureaplasmas in the aetiology ofNGU was obtained, but the data suggest that B ureolyticus organisms, and perhaps other anaerobes, have an important role in a small proportion of cases and that the beneficial effects of metronidazole given on an empirical basis will be confined to anaerobe positive urethritis.Evidence that Chlamydia trachomatis organisms (chlamydiae) are a cause of non-gonococcal urethritis (NGU) is overwhelming.' Chlamydial infection, however, is probably responsible for no more than about half the cases. Ureaplasma urealyticum organisms (ureaplasmas) In a study of the anaerobic microbial flora of the urethra, we found a Gram negative anaerobic bacillus appreciably more often in patients with NGU than in those withoit, and termed it a "NGU-associated anaerobe".5 Subsequent work showed that this

SUMMARY Sixty-four men with non-gonococcal urethritis (NGU), seven with gonococcal urethritis (GU), and 30 who had no symptoms or signs of urethritis were studied. Chlarnydia trachomatis was isolated from urethral specimens taken from 22% of the men with NGU, and 1 8/o with GU, but not from those who did not have urethritis even though 20 (67%) of them had a history of NGU, GU, or both. The chlamydial isolation rate for men having NGU for the first time was 30%7o. Ureaplasma urealyticum was isolated from 42% of the men with NGU, 43% of men with GU, and 27% of those without urethritis.In addition to aerobes anaerobes were isolated frequently from men whether or not they had urethritis, the most common being anaerobic Corynebacteria, peptococci, and micro-organisms of the Bacteroides-Fusobacteria group. There was no appreciable difference in the overall isolation of anaerobes from men with NGU (89%) or from those without disease (8007o). The rate of isolation of a Gram-negative anaerobic bacillus from men with NGU (50%) was, however, strikingly different to that from men with GU (14%) or from those without disease (13%o). Furthermore, this bacillus was recovered from 28 (56Db) of 50 men with NGU who were considered chlamydia-negative and from 19 (61 %) of 31 men who were both chlamydia-negative and ureaplasma-negative. The effect of antibiotic treatment was evaluated in a few patients only, so that although those from whom the Gram-negative bacillus was eliminated by tetracycline recovered clinically there is a need for a prospective therapeutic trial to evaluate further the importance of the bacillus.

A numerical taxonomic study of 64 strains of anaerobic Gram-negative bacilli isolated from men with non-gonococcal urethritis, two unclassified laboratory strains of 'corroding bacilli', and 12 other strains of anaerobic Gram-negative bacilli, including nine received as anaerobic curved rods and three as 'Bacteroides corrodens' (B. ureolyticus), isolated from women with bacterial vaginosis, was undertaken. Seventeen reference anaerobic strains belonging to the genera Bacteroides, Fusobacterium, Mobiluncus, Mitsuokella and Wolinella were included. Morphological, biochemical and physiological characteristics were examined in 103 tests. The resemblance between the 95 strains was calculated using the SSM, SJ and DP coefficients for cluster analyses based on the UPGMA method. All three approaches gave similar groupings, and the estimated average probability of test error was 2.46%. The strains fell into 10 phenons. The unclassified strains from men and three from women with lower genital-tract infections, and the laboratory strains of 'corroding bacilli' clustered in one phenon with the reference strains of B. ureolyticus, indicating that they correspond to B. ureolyticus. The other unclassified strains of anaerobic curved rods clustered as a distinct phenon. They correspond to species of the newly described genus Mobiluncus. The taxonomic data and the compilation of diagnostic tables serve as a useful guide for the laboratory identification of clinical isolates regarded as B. ureolyticus.