Table 4.
Frequencies of HLA risk genotypes and haplotypes at diagnosis in familial groups and children with sporadic diabetes (n = 4993)
| Genetics, % (95% CI) | 1. Affected father, n=253 (5.1%) | 2. Affected mother, n=141 (2.8%) | 3. Affected sibling, n=95 (1.9%) | 4. ≥2 affected family members, n=30 (0.6%) | 5. Sporadic, n=4474 (89.6%) | p value | Adjusted p valuea |
|---|---|---|---|---|---|---|---|
| DR3-DQ2/DR4-DQ8 | 21.3 (16.3, 26.4) | 23.4 (16.4, 30.4) | 29.5 (20.3, 38.6) | 26.7 (10.8, 42.5) | 21.1 (19.9, 22.2) | 0.306 | 0.238 |
| DR3-DQ2/xb | 11.1 (7.2, 14.9) | 17.0 (10.8, 23.2) | 12.6 (6.0, 19.3) | 13.3 (1.2, 25.5) | 15.6 (14.5, 16.6) | 0.320 | 0.313 |
| DR4-DQ8/yc | 56.1 (50.0, 62.2) | 51.1 (42.8, 59.3) | 42.1 (32.2, 52.0) | 53.3 (35.5, 71.2) | 47.5 (46.0, 48.9) | 0.052 | 0.048 |
| 1 vs 3: <0.05 | |||||||
| 1 vs 5: <0.05 | |||||||
| xb/yc | 11.5 (7.5, 15.4) | 8.5 (3.9, 13.1) | 15.8 (8.5, 23.1) | 6.7 (0, 15.6) | 15.9 (14.8, 17.0) | 0.030 | 0.044 |
| 2 vs 5: <0.05 | |||||||
| DR3-DQ2 homozygosity | 3.2 (1.0, 5.3) | 2.8 (0.1, 5.6) | 3.2 (0, 6.7) | 3.3 (0, 9.8) | 3.0 (2.5, 3.5) | 0.971 | 1.000 |
| DR4-DQ8 homozygosity | 15.0 (10.6, 19.4) | 10.6 (5.5, 15.7) | 8.4 (2.8, 14.0) | 10.0 (0, 20.7) | 7.6 (6.9, 8.4) | 0.001 | 0.001 |
| 1 vs 5: 0.001 | |||||||
| DR3-DQ2 | 32.4 (26.6, 38.2) | 40.4 (32.3, 48.5) | 42.1 (32.2, 52.0) | 40.0 (22.5, 57.5) | 36.6 (35.2, 38.0) | 0.381 | 0.336 |
| DR4-DQ8 | 77.5 (72.3, 82.6) | 74.5 (67.3, 81.7) | 71.6 (62.5, 80.6) | 80.0 (65.7, 94.3) | 68.5 (67.1, 70.0) | 0.012 | 0.015 |
| 1 vs 5: <0.05 |
Variables are presented as % (95% CI)
In case of significant differences in the analyses between the four familial subgroups and those with sporadic disease, paired comparisons by groups were also performed. Only significant p values are presented from the paired analyses
For comparing frequencies in each study group, cross tabulation and χ2 statistics with continuity correction or Fisher’s exact test when appropriate were used. Differences in levels were determined using Kruskal–Wallis test or the Mann–Whitney U test
Adjustment for confounding factors was performed using logistic/ordinal/multinomial regression for dichotomous/ordinal/categorical variables
aAdjusted for sex and age at diagnosis
bx ≠ DR4-DQ8
cy ≠ DR3-DQ2