Increased levels of Porphyromonas gingivalis are associated with ischemic and hemorrhagic cerebrovascular disease in humans: an in vivo study

Objective: This study investigated the role of periodontal disease in the development of stroke or cerebral infarction in patients by evaluating the clinical periodontal conditions and the subgingival levels of periodontopathogens. Material and Methods: Twenty patients with ischemic (I-CVA) or hemorrhagic (H-CVA) cerebrovascular episodes (test group) and 60 systemically healthy patients (control group) were evaluated for: probing depth, clinical attachment level, bleeding on probing and plaque index. Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans were both identified and quantified in subgingival plaque samples by conventional and real-time PCR, respectively. Results: The test group showed a significant increase in each of the following parameters: pocket depth, clinical attachment loss, bleeding on probing, plaque index and number of missing teeth when compared to control values (p<0.05, unpaired t-test). Likewise, the test group had increased numbers of sites that were contaminated with P. gingivalis (60%x10%; p<0.001; chi-squared test) and displayed greater prevalence of periodontal disease, with an odds ratio of 48.06 (95% CI: 5.96-387.72; p<0.001). Notably, a positive correlation between probing depth and the levels of P. gingivalis in ischemic stroke was found (r=0.60; p=0.03; Spearman's rank correlation coefficient test). A. actinomycetemcomitans DNA was not detected in any of the groups by conventional or real-time PCR. Conclusions: Stroke patients had deeper pockets, more severe attachment loss, increased bleeding on probing, increased plaque indexes, and in their pockets harbored increased levels of P. gingivalis. These findings suggest that periodontal disease is a risk factor for the development of cerebral hemorrhage or infarction. Early treatment of periodontitis may counteract the development of cerebrovascular episodes.

Antimicrobial therapy using a local drug delivery system (Arestin) in the treatment of peri-implantitis. I: Microbiological outcomes

OBJECTIVES: To assess the microbiological outcome of local administration of minocycline hydrochloride microspheres 1 mg (Arestin) in cases with peri-implantitis and with a follow-up period of 12 months. MATERIAL AND METHODS: After debridement, and local administration of chlorhexidine gel, peri-implantitis cases were treated with local administration of minocycline microspheres (Arestin). The DNA-DNA checkerboard hybridization method was used to detect bacterial presence during the first 360 days of therapy. RESULTS: At Day 10, lower bacterial loads for 6/40 individual bacteria including Actinomyces gerensceriae (P<0.1), Actinomyces israelii (P<0.01), Actinomyces naeslundi type 1 (P<0.01) and type 2 (P<0.03), Actinomyces odontolyticus (P<0.01), Porphyromonas gingivalis (P<0.01) and Treponema socranskii (P<0.01) were found. At Day 360 only the levels of Actinobacillus actinomycetemcomitans were lower than at baseline (mean difference: 1x10(5); SE difference: 0.34x10(5), 95% CI: 0.2x10(5) to 1.2x10(5); P<0.03). Six implants were lost between Days 90 and 270. The microbiota was successfully controlled in 48%, and with definitive failures (implant loss and major increase in bacterial levels) in 32% of subjects. CONCLUSIONS: At study endpoint, the impact of Arestin on A. actinomycetemcomitans was greater than the impact on other pathogens. Up to Day 180 reductions in levels of Tannerella forsythia, P. gingivalis, and Treponema denticola were also found. Failures in treatment could not be associated with the presence of specific pathogens or by the total bacterial load at baseline. Statistical power analysis suggested that a case control study would require approximately 200 subjects.

Microbiological composition associated with interleukin-1 gene polymorphism in subjects undergoing supportive periodontal therapy

BACKGROUND: Interleukin-1 gene polymorphism (IL-1 gene) has been associated with periodontitis. The present study examined the subgingival microbiota by IL-1 gene status in subjects undergoing supportive periodontal therapy (SPT). METHODS: A total of 151 subjects with known IL-1 gene status (IL-1A +4845/IL-1B -3954) (IL-1 gene) were included in this study. Clinical data and subgingival plaque samples (40 taxa) were collected. These taxa were determined by the checkerboard DNA-DNA hybridization method. RESULTS: Gender, smoking habits (n-par tests), age, and clinical periodontal conditions did not differ by IL-1 gene status. IL-1 gene-negative subjects had a higher total bacterial load (mean difference, 480.4 x 10(5); 95% confidence interval [CI], 77 to 884 x 10(5); P <0.02). The levels of Actinobacillus actinomycetemcomitans (mean difference, 30.7 x 10(5); 95% CI, 2.2 to 59.5 x 10(5); P <0.05), Eubacterium nodatum (mean difference, 4.2 x 10(5); 95% CI, 0.6 to 7.8 x 10(5); P <0.02), Porphyromonas gingivalis (mean difference, 17.9 x 10(5); 95% CI, 1.2 to 34.5 x 10(5); P <0.05), and Streptococcus anginosus (mean difference, 4.0 x 10(5); 95% CI, 0.2 to 7.2 x 10(5); P <0.05) were higher in IL-1 gene-negative subjects, an observation specifically found at sites with probing depths <5.0 mm. CONCLUSIONS: Bleeding on probing did not differ by IL gene status, reflecting clinical SPT efficacy. IL-1 gene-negative subjects had higher levels of periodontal pathogens. This may suggest that among subjects undergoing SPT, a lower bacterial load is required in IL-1 gene-positive subjects to develop the same level of periodontitis as in IL-1 gene-negative subjects.

Influence of the menstrual cycle on the oral microbial flora in women: a case-control study including men as control subjects

Changes in the levels of female sex hormones during the menstrual cycle may cause cyclic differences in subgingival bacterial colonization patterns. The purpose of the present study was to test the hypothesis that hormonal changes in the menstrual cycle cause changes in the oral microbiota. METHODS: Bacterial plaque samples were collected in 20 systemically and periodontally healthy women using no hormonal contraceptives (test group) over a period of 6 weeks. Twenty age-matched systemically and periodontally healthy men were assigned to the control group. Samples were processed by checkerboard DNA-DNA hybridization assay, and 74 species were analyzed. RESULTS: No cyclic pattern of bacterial colonization was identified for any of the 74 species studied in women not using hormonal contraceptives. Aggregatibacter actinomycetemcomitans (previously Actinobacillus actinomycetemcomitans) (Y4) was common at the beginning of menstruation (mean: 32%) and increased during the following 2 weeks (36%) in women (P <0.05). No cyclic differences in bacterial presence were found among the men (P values varied between 0.14 and 0.98). Men presented with significantly higher bacterial counts for 40 of 74 species (P <0.001), including Staphylococcus aureus and Pseudomonas aeruginosa but not Porphyromonas gingivalis (P = 0.15) or Tannerella forsythia (previously T. forsythensis) (P = 0.42). CONCLUSIONS: During a menstruation period, cyclic variation in the subgingival microbiota of periodontally healthy women of child-bearing age who were not using oral hormonal contraceptives could not be confirmed. Male control subjects presented with higher levels of many species but also without a cyclic pattern.

Does pregnancy have an impact on the subgingival microbiota?

BACKGROUND: We investigated clinical and subgingival microbiologic changes during pregnancy in 20 consecutive pregnant women > or =18 years not receiving dental care. METHODS: Bacterial samples from weeks 12, 28, and 36 of pregnancy and at 4 to 6 weeks postpartum were processed for 37 species by checkerboard DNA-DNA hybridization. Clinical periodontal data were collected at week 12 and at 4 to 6 weeks postpartum, and bleeding on probing (BOP) was recorded at sites sampled at the four time points. RESULTS: The mean BOP at week 12 and postpartum was 40.1% +/- 18.2% and 27.4% +/- 12.5%, respectively. The corresponding mean BOP at microbiologic test sites was 15% (week 12) and 21% (postpartum; not statistically significant). Total bacterial counts decreased between week 12 and postpartum (P <0.01). Increased bacterial counts over time were found for Neisseria mucosa (P <0.001). Lower counts (P <0.001) were found for Capnocytophaga ochracea, Capnocytophaga sputigena, Eubacterium saburreum, Fusobacterium nucleatum naviforme, Fusobacterium nucleatum polymorphum, Leptotrichia buccalis, Parvimonas micra (previously Peptostreptococcus micros or Micromonas micros), Prevotella intermedia, Prevotella melaninogenica, Staphylococcus aureus, Streptococcus anginosus, Streptococcus intermedius, Streptococcus mutans, Streptococcus oralis, Streptococcus sanguinis, Selenomonas noxia, and Veillonella parvula. No changes occurred between weeks 12 and 28 of pregnancy. Counts of Aggregatibacter actinomycetemcomitans (previously Actinobacillus actinomycetemcomitans), Porphyromonas gingivalis, Tannerella forsythia (previously T. forsythensis), and Treponema denticola did not change. Counts of P. gingivalis and T. forsythia at week 12 were associated with gingivitis (P <0.001). CONCLUSIONS: Subgingival levels of bacteria associated with periodontitis did not change. P. gingivalis and T. forsythia counts were associated with BOP at week 12. A decrease was found in 17 of 37 species from week 12 to postpartum. Only counts of N. mucosa increased.

The impact of the stone age diet on gingival conditions in the absence of oral hygiene

BACKGROUND: The objective of this study was to assess the oral microbiota and clinical data in subjects without access to traditional oral hygiene methods and who ate a diet available in the Stone Age. METHODS: Ten subjects living in an environment replicating the Stone Age for 4 weeks were enrolled in this study. Bleeding on probing (BOP), gingival and plaque indices, and probing depth (PD) were assessed at baseline and at 4 weeks. Microbiologic samples were collected at the mesio-buccal subgingival aspects of all teeth and from the dorsum of the tongue and were processed by checkerboard DNA-DNA hybridization methods. RESULTS: No subject had periodontitis. Mean BOP decreased from 34.8% to 12.6% (P <0.001). Mean gingival index scores changed from 0.38 to 0.43 (not statistically significant) and mean plaque scores increased from 0.68 to 1.47 (P <0.001). PD at sites of subgingival sampling decreased (mean difference: 0.2 mm; P <0.001). At week 4, the total bacterial count was higher (P <0.001) for 24 of 74 species, including Bacteroides ureolyticus, Eikenella corrodens, Lactobacillus acidophilus, Capnocytophaga ochracea, Escherichia coli, Fusobacterium nucleatum naviforme, Haemophilus influenzae, Helicobacter pylori, Porphyromonas endodontalis, Staphylococcus aureus (two strains), Streptococcus agalactiae, Streptococcus anginosis, and Streptococcus mitis. Bacterial counts from tongue samples were higher at baseline (P <0.001) for 20 species, including Tannerella forsythia (previously T. forsythensis), Aggregatibacter actinomycetemcomitans (previously Actinobacillus actinomycetemcomitans; serotype a), and Streptococcus spp. CONCLUSIONS: The experimental gingivitis protocol is not applicable if the diet (e.g., Stone Age) does not include refined sugars. Although plaque levels increased, BOP and PD decreased. Subgingival bacterial counts increased for several species not linked to periodontitis, whereas tongue bacterial samples decreased during the study period.

Detection of RTX toxin genes in gram-negative bacteria with a set of specific probes

The family of RTX (RTX representing repeats in the structural toxin) toxins is composed of several protein toxins with a characteristic nonapeptide glycine-rich repeat motif. Most of its members were shown to have cytolytic activity. By comparing the genetic relationships of the RTX toxin genes we established a set of 10 gene probes to be used for screening as-yet-unknown RTX toxin genes in bacterial species. The probes include parts of apxIA, apxIIA, and apxIIIA from Actinobacillus pleuropneumoniae, cyaA from Bordetella pertusis, frpA from Neisseria meningitidis, prtC from Erwinia chrysanthemi, hlyA and elyA from Escherichia coli, aaltA from Actinobacillus actinomycetemcomitans and lktA from Pasteurella haemolytica. A panel of pathogenic and nonpathogenic gram-negative bacteria were investigated for the presence of RTX toxin genes. The probes detected all known genes for RTX toxins. Moreover, we found potential RTX toxin genes in several pathogenic bacterial species for which no such toxins are known yet. This indicates that RTX or RTX-like toxins are widely distributed among pathogenic gram-negative bacteria. The probes generated by PCR and the hybridization method were optimized to allow broad-range screening for RTX toxin genes in one step. This included the binding of unlabelled probes to a nylon filter and subsequent hybridization of the filter with labelled genomic DNA of the strain to be tested. The method constitutes a powerful tool for the assessment of the potential pathogenicity of poorly characterized strains intended to be used in biotechnological applications. Moreover, it is useful for the detection of already-known or new RTX toxin genes in bacteria of medical importance.

Phylogeny of genes for secretion NTPases: Identification of the widespread tadA subfamily and development of a diagnostic key for gene classification

Macromolecular transport systems in bacteria currently are classified by function and sequence comparisons into five basic types. In this classification system, type II and type IV secretion systems both possess members of a superfamily of genes for putative NTP hydrolase (NTPase) proteins that are strikingly similar in structure, function, and sequence. These include VirB11, TrbB, TraG, GspE, PilB, PilT, and ComG1. The predicted protein product of tadA, a recently discovered gene required for tenacious adherence of Actinobacillus actinomycetemcomitans, also has significant sequence similarity to members of this superfamily and to several unclassified and uncharacterized gene products of both Archaea and Bacteria. To understand the relationship of tadA and tadA-like genes to those encoding the putative NTPases of type II/IV secretion, we used a phylogenetic approach to obtain a genealogy of 148 NTPase genes and reconstruct a scenario of gene superfamily evolution. In this phylogeny, clear distinctions can be made between type II and type IV families and their constituent subfamilies. In addition, the subgroup containing tadA constitutes a novel and extremely widespread subfamily of the family encompassing all putative NTPases of type IV secretion systems. We report diagnostic amino acid residue positions for each major monophyletic family and subfamily in the phylogenetic tree, and we propose an easy method for precisely classifying and naming putative NTPase genes based on phylogeny. This molecular key-based method can be applied to other gene superfamilies and represents a valuable tool for genome analysis.

Genomic DNA fingerprinting by restriction fragment end labeling.

A typing method for bacteria was developed and applied to several species, including Escherichia coli and Actinobacillus actinomycetemcomitans. Total genomic DNA was digested with a restriction endonuclease, and fragments were enabled with [alpha-32P]dATP by using the Klenow fragment of DNA polymerase and separated by electrophoresis in 6% polyacrylamide/8 M urea (sequencing gel). Depending on the restriction endonuclease and the bacterium, the method produced approximately 30-50 well-separated fragments in the size range of 100-400 nucleotides. For A. actinomycetemcomitans, all strains had bands in common. Nevertheless, many polymorphisms could be observed, and the 31 strains tested could be classified into 29 distinct types. Furthermore, serotype-specific fragments could be assigned for the three serotypes investigated. The method described is very sensitive, allowing more distinct types to be distinguished than other commonly used typing methods. When the method was applied to 10 other clinically relevant bacterial species, both species-specific bands and strain-specific bands were found. Isolates from different locations of one patient showed indistinguishable patterns. Computer-assisted analysis of the DNA fingerprints allowed the determination of similarity coefficients. It is concluded that genomic fingerprinting by restriction fragment end labeling (RFEL) is a powerful and generally applicable technique to type bacterial species.

Modulation of the antibody response by Porphyromonas gingivalis and Fusobacterium nucleatum in a mouse model

Successive immunization of mice with Fusobacterium nucleatum and Porphyromonas gingivalis has been shown to modulate the specific serum IgG responses to these organisms. The aim of this study was to investigate these antibody responses further by examining the IgG subclasses induced as well as the opsonizing properties of the specific antibodies. Serum samples from BALB/c mice immunized with F. nucleatum (gp1-F), P. gingivalis (gp2-P), P. gingivalis followed by F. nucleatum (gp3-PF) F. nucleatum followed by P. gingivalis (gp4-FP) or saline alone (gp5-S) were examined for specific IgG1 (Th2) and IgG2a (Th1) antibody levels using an ELISA and the opsonizing properties measured using a neutrophil chemiluminescence assay. While IgG1 and IgG2a subclasses were induced in all immunized groups, there was a tendency towards an IgG1 response in mice immunized with P. gingivalis alone, while immunization with F. nucleatum followed by P. gingivalis induced significantly higher anti-P. gingivalis IgG2a levels than IgG1. The maximum light output due to neutrophil phagocytosis of P. gingivalis occurred at 10 min using nonopsonized bacteria. Chemiluminescence was reduced using serum-opsonized P. gingivalis and, in particular, sera from P. gingivalis-immunized mice (gp2-P), with maximum responses occurring at 40 min. In contrast, phagocytosis of immune serum-opsonized F. nucleatum demonstrated peak light output at 10 min, while that of F. nucleatum opsonized with sera from saline injected mice (gp5-S) and control nonopsonized bacteria showed peak responses at 40 min. The lowest phagocytic response occurred using gp4-FP serum-opsonized F. nucleatum. In conclusion, the results of the present study have demonstrated a systemic Th1/Th2 response in mice immunized with P. gingivalis and/or F. nucleatum with a trend towards a Th2 response in P. gingivalis-immunized mice and a significantly increased anti-P. gingivalis IgG2a (Th1) response in mice immunized with F. nucleatum prior to P. gingivalis. Further, the inhibition of neutrophil phagocytosis of immune serum-opsonized P. gingivalis was modulated by the presence of anti-F. nucleatum antibodies, while anti-P. gingivalis antibodies induced an inhibitory effect on the phagocytic response to F. nucleatum.

L-arginine-dependent nitric oxide production of a murine macrophage-like RAW 264.7 cell line stimulated with Porphyromonas gingivalis lipopolysaccharide

The aim of this study was to determine nitric oxide (NO) production of a murine macrophage cell line (RAW 264.7 cells) when stimulated with Porphyromonas gingivalis lipopolysaccharides (Pg-LPS). RAW264.7 cells were incubated with i) various concentrations of Pg-LPS or Salmonella typhosa LPS (St-LPS), ii) Pg-LPS with or without L-arginine and/or N-G-monomethyl-L-arginine (NMMA), an arginine analog or iii) Pg-LPS and interferon-gamma (IFN-gamma) with or without anti-IFN-gamma antibodies or interleukin-10 (IL-10). Tissue culture supernatants were assayed for NO levels after 24 h in culture. NO was not observed in tissue culture supernatants of RAW 264.7 cells following stimulation with Pg-LPS, but was observed after stimulation with St-LPS. Exogenous L-arginine restored the ability of Pg-LPS to induce NO production; however, the increase in NO levels of cells stimulated with Pg-LPS with exogenous L-arginine was abolished by NMMA. IFN-gamma induced independent NO production by Pg-LPS-stimulated macrophages and this stimulatory effect of IFN-gamma could be completely suppressed by anti-IFN-gamma antibodies and IL-10. These results suggest that Pg-LPS is able to stimulate NO production in the RAW264.7 macrophage cell model in an L-arginine-dependent mechanism which is itself independent of the action of IFN-gamma.

Effect of periodontal treatment on the serum antibody levels to heat shock proteins

We have shown previously that both humoral and cellular immune responses to heat shock protein 60 (HSP60) are elevated in chronic periodontitis patients compared with non-diseased subjects. The aim of the present study was to determine whether periodontal treatment could influence the level of serum antibodies to human HSP60 and Porphyromonas gingivalis GroEL, a bacterial homologue of human HSP60. Sera were obtained from 21 patients with moderate to advanced chronic periodontitis at the baseline examination and again after completion of treatment. Antibody levels were determined using an enzyme-linked immunosorbent assay. The mean anti-P. gingivalis GroEL antibody levels were down-regulated significantly by periodontal treatment when recombinant P. gingivalis GroEL was used as an antigen, whereas antibody levels to P. gingivalis GroEL-specific peptide were significantly elevated following successful periodontal therapy. The mean level of anti-human HSP60 antibody remained unchanged although individual levels of antibody either increased or decreased after periodontal treatment, suggesting that synthesis of these antibodies might be regulated independently during the course of periodontal infection. Although their regulatory mechanisms in chronic infection are not understood, further study would provide insight not only into the role of these antibodies in the pathogenesis of periodontitis but also into the possible link between periodontitis and systemic diseases such as coronary heart disease.

Cross-reactivity of GroEL antibodies with human heat shock protein 60 and quantification of pathogens in atherosclerosis

Background/aims: Chronic infections such as those caused by Chlamydia pneumoniae and periodontopathic bacteria such as Porphyromonas gingivalis have been associated with atherosclerosis, possibly due to cross-reactivity of the immune response to bacterial GroEL with human heat shock protein (hHSP) 60. Methods: We examined the cross-reactivity of anti-GroEL and anti-P. gingivalis antibodies with hHSP60 in atherosclerosis patients and quantified a panel of six pathogens in atheromas. Results: After absorption of plasma samples with hHSP60, there were variable reductions in the levels of anti-GroEL and anti-P. gingivalis antibodies, suggesting that these antibodies cross-reacted with hHSP60. All of the artery specimens were positive for P. gingivalis. Fusobacterium nucleatum, Tannerella forsythia, C. pneumoniae, Helicobacter pylori, and Haemophilus influenzae were found in 84%, 48%, 28%, 4%, and 4% of arteries, respectively. The prevalence of the three periodontopathic microorganisms, P. gingivalis, F. nucleatum and T. forsythia, was significantly higher than that of the remaining three microorganisms. Conclusions: These results support the hypothesis that in some patients, cross-reactivity of the immune response to bacterial HSPs including those of periodontal pathogens, with arterial endothelial cells expressing hHSP60 may be a possible mechanism for the association between atherosclerosis and periodontal infection.

The effects of intra-amniotic injection of periodontopathic lipopolysaccharides in sheep

Objective: Periodontal disease may cause several complications of pregnancy, including fetal death. The purpose of this study was to investigate in sheep the effects of the intra-amniotic injection of lipopolysaccharide from 3 periodontopathic organisms and to compare these effects with those resulting from similar injection of Escherichia coli lipopolysaccharide. The outcomes that were studied included the rates of fetal death and the features of inflammation and lung maturation in survivors. Study design: At 118 days of pregnancy, ewes that were bearing single fetuses were allocated at random to receive intra-amniotic injections of saline solution (n = 13 fetuses), or lipopolysaccharide from Porphyromonas gingivalis (in doses from 0.1 to 10 tug [n = 22 fetuses]), Actinobacillus actinomycetemcomitans (10 mg [n = 6 fetuses]; 1 mg [n = 6 fetuses]), Fusobacterium nucleation (10 mg [n = 6 fetuses]) or Escherichia coli (10 mg [n = 14 fetuses]; 1 mg [n = 7 fetuses]). Surviving fetuses were delivered abdominally at 125 days of gestation (term, 150 days). Results: When compared with Escherichia coli lipopolysaccharide at similar dosages, periodontopathic lipopolysaccharides had high rates of fetal lethality. Only 6 of 22 fetuses that were exposed to intra-amniotic Porphyromonas gingivalis lipopolysaccharide survived doses of 0.1 to 10 mg, and only 3 of 6 fetuses survived 10-mg Actinobacillus actinomycetemcomitans lipopolysaccharide. Escherichia coli lipopolysaccharide did not cause fetal loss when given at doses of 10 mg (n = 14 fetuses) or l mg (n = 7 fetuses). Fetuses that survived exposure to these lipopolysaccharides showed features of inflammation in amniotic fluid and cord blood at birth and enhanced lung maturation. Conclusion: Lipopolysaccharides from these 3 periodontopathic organisms have much higher rates of fetal lethality than Escherichia coli lipopolysaccharide but can cause similar intrauterine inflammatory responses and improvements in lung volumes in survivors. Sources of inflammation that are distant from the uterus may underlie a proportion of unexplained stillbirth and other complications of pregnancy. (c) 2005 Mosby, Inc. All rights reserved.

The distribution of Tannerella forsythia in an adolescent and adult population

Background: The fact that Tannerella forsythia, an important periopathogen, is difficult to cultivate from mixed infections has impeded precise estimates of its distribution within a given population. In order to discern T. forsythia alone from the mixed infection of plaque, the use of sensitive 16S ribosomal RNA based polymerase chain reaction (PCR) detection is necessary. Objectives: The aim of the present study was to determine the distribution of T. forsythia in an adult and in an adolescent population. Materials and methods: Subgingival plaque samples were obtained from 498 Australian adults and from 228 adolescent subjects from Manchester, UK. Tannerella forsythia was detected using PCR and confirmed by restriction analysis. Semi-quantitation of the organisms was carried out using two specific primers of differing sensitivities. Results: In the adolescent population, 25% were found to carry T. forsythia, albeit in relatively low numbers. In the adult population, a total of 37.8% and 11% were found to carry the organism with primer 2 and primer 1, respectively, suggesting that around 27% had between 10(3) and 10(7) organisms. Although there was an apparent increased proportion of T. forsythia positive subjects in those aged >= 50 years, this was not statistical significant. However, T. forsythia positive male smokers showed increased disease severity compared with T. forsythia negative subjects. Conclusion: This study has shown that at least 25% of the adolescent population carry low numbers of T. forsythia, whereas at least 37% of adults carry the organism, with some 11% having relatively high numbers. The relationship between T. forsythia and disease progression in these populations, however, remains to be determined.