Table 2. Frequency of enterovirus RNA in faecal samples prior to islet autoimmunity and matched controls and influence of IFIH1 polymorphisms.
| Cases* | Controls* | OR (95% CI)† | OR (95% CI)† | |
| (n = 25 subjects) | (n = 50 subjects) | Unadjusted | Adjusted | |
| EV infection and later development of islet autoimmunity: | ||||
| EV− | 282 | 555 | 1.00 (reference) | 1.00 (reference) |
| EV+ | 42 (13.0%) | 93 (14.4%) | 0.95 (0.58–1.55), P = 0.83 | 0.99 (0.62–1.60)‡ |
| The association of IFIH1 rs1990760 polymorphisms with islet autoimmunity: | ||||
| rs1990760 Ala− | 9 (36.0%) | 20 (40.0%) | 1.00 (reference) | 1.00 (reference) |
| rs1990760 Ala+ | 16 (64.0%) | 30 (60.0%) | 1.18 (0.44–3.14); P = 0.74 | OR = 1.2 (0.44–3.26)§ |
| EV infection and later development of islet autoimmunity, stratified by IFIH1 rs1990760: | ||||
| EV−, rs199076, Ala− | 181 (86.2%) | 300 (86.0%) | 1.00 (reference) | |
| EV+, rs199076, Ala− | 29 (13.8%) | 49 (14.0%) | 0.77 (0.34–1.77) | |
| EV−, rs199076, Ala+ | 101 (88.6%) | 255 (85.3%) | 1.00 (reference) | |
| EV+, rs199076, Ala+ | 13 (11.4%) | 44 (14.7%) | 1.04 (0.50–2.15) | P(interaction) = 0.74∥ |
EV, enterovirus; OR, odds ratio; CI, confidence interval.
*Cases defined as repeated positivity in consecutive blood samples for two or three islet autoantibodies (anti-insulin, anti-GAD, anti-IA2). Controls matched by high risk HLA genotype, date of birth, time of follow-up and county of residence. See reference [13] for more details.
Estimated from a three level random intercept logistic regression model with enterovirus positivity as dependent variable and case/control status and IFIH1 SNPs as independent variables. Nested random effects were specified for individuals (samples within individuals) and for matched set (individuals within matched sets of a case and 1–2 controls).
Enterovirus adjusted for IFIH1 common variant rs1990760 and vice versa. Rare variants could not be adjusted for because of zero observations in one of the comparing groups. Instead subjects/samples with rare variants of these SNPs were excluded from the analysis (1 case, 6 controls; in total 184 samples).
Adjusted for the cumulative number of enterovirus infections within a child until persistent autoimmunity develops.
Testing whether the model stratified for Thr/Thr is significantly different from the model stratified for carriers of at least one Ala allele.