Suitability of machine learning models for prediction of clinically defined Stage III/IV periodontitis from questionnaires and demographic data in Danish cohorts.

AIM: To evaluate if, and to what extent, machine learning models can capture clinically defined Stage III/IV periodontitis from self-report questionnaires and demographic data. MATERIALS AND METHODS: Self-reported measures of periodontitis, demographic data and clinically established Stage III/IV periodontitis status were extracted from two Danish population-based cohorts (The Copenhagen Aging and Midlife Biobank [CAMB] and The Danish Health Examination Survey [DANHES]) and used to develop cross-validated machine learning models for the prediction of clinically established Stage III/IV periodontitis. Models were trained using 10-fold cross-validations repeated three times on the CAMB dataset (n = 1476), and the resulting models were validated in the DANHES dataset (n = 3585). RESULTS: The prevalence of Stage III/IV periodontitis was 23.2% (n = 342) in the CAMB dataset and 9.3% (n = 335) in the DANHES dataset. For the prediction of clinically established Stage III/IV periodontitis in the CAMB cohort, models reached area under the receiver operating characteristics (AUROCs) of 0.67-0.69, sensitivities of 0.58-0.64 and specificities of 0.71-0.80. In the DANHES cohort, models derived from the CAMB cohort achieved AUROCs of 0.64-0.70, sensitivities of 0.44-0.63 and specificities of 0.75-0.84. CONCLUSIONS: Applying cross-validated machine learning algorithms to demographic data and self-reported measures of periodontitis resulted in models with modest capabilities for the prediction of Stage III/IV periodontitis in two Danish cohorts.

Lewis and AB0 blood group-phenotypes in periodontitis, cardiovascular disease, obesity and stroke.

The AB0 blood group has been linked to ischaemic heart disease, stroke, and periodontal disease, while the Lewis blood group has been linked to ischaemic heart disease and obesity, all of which have been associated with periodontitis. AB0 or Lewis blood group phenotype may therefore constitute common hereditary components predisposing to these disorders. In this study, we investigated if blood group phenotype associated with periodontitis in a subpopulation consisting of 702 participants from a Danish cross-sectional cohort and, secondarily, attempted to confirm their association with hypertension, ischaemic heart disease, stroke, and obesity. No significant association between blood group phenotype and periodontitis was detected, nor were previously reported associations between blood group phenotype and hypertension, ischaemic heart disease, stroke, and obesity confirmed. This may, at least partly, be attributed to differences in study type, outcome definitions, cohort sizes, and population attributable factors. However, our results suggested a strong association between self-reported stroke and the Lewis (a-b-) phenotype (P = 0.0002, OR: 22.28; CI 95: 4.72-131.63).

Peptidylarginine deiminase-4 gene polymorphisms are associated with systemic lupus erythematosus and lupus nephritis.

OBJECTIVE: Peptidylarginine deiminase-4 (PAD4) is highly expressed by neutrophils and essential for citrullination occurring during the formation of neutrophil extracellular traps, which have been implicated in the pathogenesis of systemic lupus erythematosus (SLE) and lupus nephritis (LN). Single-nucleotide polymorphisms (SNPs) in PADI4 influence PAD4 expression and functionality. Here, we investigate whether SNPs in PADI4 influence the risk of SLE or LN. METHOD: Altogether, 234 SLE patients and 484 controls were genotyped for nine PADI4 SNPs known to alter PAD4 functionality and/or expression, or to be associated with other autoimmune diseases, using an in-house multiplex Luminex assay. All analyses were adjusted for age and gender. RESULTS: Heterozygosity for rs1748033, and heterozygosity and homozygosity for rs1635564, were associated with increased occurrence of SLE [odds ratio (OR) 1.55, 95% confidence interval (CI) 1.08-2.23; OR 1.52, 95% CI 1.06-2.19; and OR 2.06, 95% CI 1.08-3.93, respectively]. Homozygosity for rs1635564 was also associated with increased occurrence of LN (OR 3.35, 95% CI 1.2-10.97). Notably, gene dose effects of the rs1635564 variant allele were observed for SLE (p = 0.005) and LN (p = 0.01). Carriage of minor alleles of five other SNPs (rs11203366, rs11203367, rs874881, rs2240340, and rs11203368) was associated with increased occurrence of LN and hypertension. CONCLUSION: The rs1635564 polymorphism of PADI4 is a candidate risk factor for SLE, particularly with renal involvement. Additional PADI4 polymorphisms also conferred increased risk of LN. Overall, these findings support the notion of PAD4 contributing to the pathogenesis of SLE and LN.

Non-HLA gene polymorphisms in juvenile idiopathic arthritis: associations with disease outcome.

OBJECTIVE: To test the hypothesis that non-HLA single-nucleotide polymorphisms (SNPs) associated with the risk of juvenile idiopathic arthritis (JIA) are risk factors for an unfavourable disease outcome at long-term follow-up. METHODS: The Nordic JIA cohort is a prospective multicentre study cohort of patients from the Nordic countries. In all, 193 patients met the inclusion criteria of having an 8 year follow-up assessment and available DNA sample. Seventeen SNPs met the inclusion criteria of having significant associations with JIA in at least two previous independent study cohorts. Clinical endpoints were disease remission, actively inflamed joints and joints with limitation of motion (LOM), articular or extra-articular damage, and history of uveitis. RESULTS: Evidence of associations between genotypes and endpoints were found for STAT4, ADAD1-IL2-IL21, PTPN2, and VTCN1 (p = 0.003-0.05). STAT4_rs7574865 TT was associated with the presence of actively inflamed joints [odds ratio (OR) 20.6, 95% confidence interval (CI) 2.2-> 100, p = 0.003] and extra-articular damage (OR 7.9, 95% CI 1-56.6, p = 0.057). ADAD1_rs17388568 AA was associated with a lower risk of having joints with LOM (OR 0.1, 95% CI 0-0.55, p = 0.016). PTPN2_rs1893217 CC was associated with a lower risk of having joints with LOM (OR 0.2, 95% CI 0-0.99, p = 0.026), while VTCN1_rs2358820 GA was associated with uveitis (OR 3.5, 95% CI 1-12.1, p = 0.029). CONCLUSION: This exploratory study, using a prospectively followed JIA cohort, found significant associations between long-term outcome and SNPs, all previously associated with development of JIA and involved in immune regulation and signal transduction in immune cells.

In vitro complement activation, adherence to red blood cells and induction of mononuclear cell cytokine production by four strains of Aggregatibacter actinomycetemcomitans with different fimbriation and expression of leukotoxin.

BACKGROUND AND OBJECTIVE: The periodontal pathogen Aggregatibacter actinomycetemcomitans has been proposed as pro-atherogenic, and complement-mediated adherence to red blood cells (RBCs) may facilitate its systemic spread. We investigated the ability of four strains of A. actinomycetemcomitans with differential expression of leukotoxin A (LtxA) and fimbriae to activate complement, adhere to RBCs and elicit cytokine responses by mononuclear cells (MNCs). MATERIAL AND METHODS: Aggregatibacter actinomycetemcomitans serotype b strains HK 921, HK 1651, HK 2092 and HK 2108 were fluorescence-labeled, incubated with human whole blood cells in the presence of autologous serum, and assessed for RBC adherence by flow cytometry and for capacity to induce cytokine production by cytometric bead array analysis. The levels of IgG to A. actinomycetemcomitans serotype b were quantified by ELISA, as was consumption of complement. RESULTS: The JP2 clone variants HK 1651 and, to a lesser extent, HK 2092, consumed complement efficiently, while HK 2108 (= strain Y4) consumed complement poorly. Nonetheless, the four tested strains adhered equally well to RBCs in the presence of autologous serum, without causing RBC lysis. The JP2 clone variant HK 2092, selectively lacking LtxA production, induced higher production of tumor necrosis factor (TNF)-α, interleukin (IL)-1ß, IL-6 and IL-10 by MNCs than did the other three strains, while the four strains induced similar production of IL-12p70. RBCs facilitated the HK 2092-induced production of TNF-α and IL-1ß, and IL-6 was enhanced by RBCs, and this facilitation could be counteracted by blockade of complement receptor 3 (CD11b/CD18). CONCLUSION: Our data suggest that the JP2 clone of A. actinomycetemcomitans, most closely resembled by the variant HK 1651, activates complement well, while strain Y4, represented by HK 2108, activates complement poorly. However, all strains of A. actinomycetemcomitans adhere to RBCs and, when capable of producing LtxA, prevent production of inflammatory cytokines by MNCs. This "immunologically silent" immune adherence may facilitate systemic spread and atherogenesis.

Single nucleotide polymorphisms in genes encoding toll-like receptors 7, 8 and 9 in Danish patients with systemic lupus erythematosus.

Several studies indicate a role for toll-like receptors (TLRs) in the pathogenesis of systemic lupus erythematosus (SLE). We aimed to investigate the risk of SLE and typical clinical and serological manifestations of SLE potentially conferred by selected single nucleotide polymorphisms (SNPs) of genes encoding TLR7, TLR8, and TLR9. Using a multiplexed bead-based assay, we analyzed eight SNPs in a cohort of 142 Danish SLE patients and a gender-matched control cohort comprising 443 individuals. Our results showed an association between the rs3853839 polymorphism of TLR7 and SLE (G vs. C, P = 0.008, OR 1.60, 95 % CI 1.12-2.27 in females; P = 0.02, OR 4.50, 95 % CI 1.18-16.7 in males) confirming recent findings in other populations. Additionally, an association between the rs3764879 polymorphism of TLR8 and SLE (G vs. C, P < 0.05, OR 1.36, 95 % CI 0.99-1.86 in females; P = 0.06, OR 4.00, 95 % CI 0.90-17.3 in males) was found. None of the other investigated SNPs were associated with SLE but several SNPs were associated with clinical and serological manifestations. In summary, a previously shown association between the rs3853839 SNP of TLR7 and SLE in Asian patients was also found in Danish patients. Together with the association of several other SNPs of TLR8 and TLR9 with various clinical and serological manifestations of SLE these findings corroborate the pathogenic significance of TLRs in SLE.

Genetic polymorphisms of dsRNA ligating pattern recognition receptors TLR3, MDA5, and RIG-I. Association with systemic lupus erythematosus and clinical phenotypes.

This study aimed to demonstrate possible associations between genetic polymorphisms in Toll-like receptor 3, interferon induced with helicase C domain 1 (IFIH1) and DEAD (Asp-Glu-Ala-Asp) box polypeptide 58 and systemic lupus erythematosus (SLE), including the phenotypes lupus nephritis and malar rash, as well as the presence of autoantibodies against nucleic acid-containing complexes. Genotyping was carried out in two Danish cohorts [Copenhagen (CPH) and Odense (ODE)] totaling 344 patients and was compared with 641 previously genotyped healthy controls. In the ODE cohort, the patients were only genotyped for the rs1990760 polymorphism of IFIH1. Single nucleotide polymorphisms (SNPs) were determined by a multiplex bead-based assay (CPH cohort) or real-time PCR (ODE cohort). Associations were investigated using the Cochran-Armitage trend test. The odds ratio (OR) for minor allele homozygotes versus major allele homozygotes suggested a protective effect of the IFIH1 rs1990760 SNP for SLE in the ODE cohort [OR 0.52, 95 % confidence intervals (95 % CI) 0.31-0.88, Pcorr. = 0.05] but not in the CPH cohort, although the OR suggested a trend in the same direction, and when combining the two patient cohorts, ORs were 0.57, 95 % CI 0.37-0.88. None of the other investigated polymorphisms showed any association with SLE. Regarding phenotypes, we found a statistically significant association between rs1990760 and malar rash in the CPH cohort, with ORs suggesting a protective effect (OR 0.28, 95 % CI 0.13-0.62 for heterozygotes and OR 0.11, 95 % CI 0.03-0.41 for homozygotes, Pcorr. = 0.0001). There were no significant associations between rs1990760 and presence of anti-dsDNA, anti-U1RNP, or anti-Smith antibodies. Our study supports previous findings of an association between the rs1990760 polymorphism of IFIH1 and SLE and indicates that this SNP may also be associated with malar rash in SLE patients although this finding needs confirmation.

Genetic polymorphisms of tumour necrosis factor receptor superfamily 1b and fas ligand are associated with clinical efficacy and/or acute severe infusion reactions to infliximab in Crohn's disease.

BACKGROUND: Single nucleotide polymorphisms (SNPs) in TNF receptor superfamily (TNFRSF) 1A and 1B, and Fas ligand (FASLG) genes, have been associated with responsiveness to infliximab (IFX) in Crohn's disease. AIM: To investigate if SNPs in TNFRSF1A and 1B, and FAS (TNFRSF6) and FASLG (TNFSF6), associated with short- or long-term clinical and biological efficacy and with acute severe infusion reactions. METHODS: Observational, retrospective and explorative cohort study of IFX-treated Caucasian patients with Crohn's disease classified as primary nonresponders (n = 21), response failures on maintenance therapy (n = 37), maintained remission (n = 47) and occurrence of acute severe infusion reactions (n = 20). RESULTS: During IFX maintenance therapy, minor allele carriage of TNFRSF1B, rs976881 is associated with loss of response [OR 3.3 (1.2-9.1), P = 0.014]. Minor allele homozygosity increased the risk substantially (OR estimated 19, P = 0.006), and furthermore associated with a mean CRP increase of 17 mg/L as compared to a mean decrease of 17 mg/L in all others (P = 0.036). In contrast, minor allele carriage of TNFRSF1B, rs1061622 is associated with beneficial response to IFX induction [OR 4.2 (1.2-18.2), P = 0.014], and with persistence of remission during maintenance therapy [OR 5.5 (1.5-25.5), P = 0.007]. Carriage of the minor allele of FASLG, rs76110 increased risk of severe infusion reactions [OR 4.0 (1.1-22.4), P = 0.041]; minor allele carriage of TNFRSF1B, rs652625 decreased the risk (OR estimated 0.2, P = 0.043 ). CONCLUSIONS: The TNFRSF1B polymorphisms may contribute to predict efficacy of infliximab. Moreover, FASLG and TNFRSF1B polymorphisms may confer genetic susceptibility to severe infusion reactions. These findings could potentially aid clinical decisions if confirmed in larger studies.

Formation of antibodies against infliximab and adalimumab strongly correlates with functional drug levels and clinical responses in rheumatoid arthritis.

BACKGROUND: Tumour necrosis factor alpha (TNFalpha) neutralising antibody constructs are increasingly being used to treat rheumatoid arthritis (RA). OBJECTIVE: To determine potential differences in clinical responses, soluble drug levels and antibody formation between patients with RA receiving infliximab and adalimumab. METHODS: 69 patients with RA fulfilling the 1987 American College of Rheumatology criteria and about to start treatment with infliximab or adalimumab, were enrolled consecutively. All patients had active disease (28-joint count Disease Activity Score >3.2). Infliximab was given intravenously at 3 mg/kg at baseline and after 2, 6 and 14 weeks. Adalimumab was administered as 40 mg biweekly subcutaneously. Concomitant drug treatment was monitored and continued at constant dosage during the study. All serum samples were tested for infliximab/adalimumab levels and anti-infliximab/anti-adalimumab antibodies. RESULTS: 35 patients received infliximab, 34 received adalimumab. At 6 months, 15 (43%), 6 (17%) and 14 (40%) of the infliximab-treated patients fulfilled the EULAR criteria for good, moderate and non-responders, respectively, whereas the corresponding figures for adalimumab-treated patients were 16 (47%), 8 (24%) and 10 (29%). Clinical responses correlated with the levels of S-infliximab/adalimumab and the formation of anti-infliximab/anti-adalimumab antibodies. CONCLUSION: The clinical response to two anti-TNFalpha biological agents closely follows the trough drug levels and the presence of antibodies directed against the drugs. Further studies that focus on the underlying pathways leading to antibody formation are warranted to predict immunogenicity of these expensive biological agents and treatment outcomes.