Table 3.
Distribution of C4 polymorphisms in Graves' disease patients with or without myxedema
| Variations | Myxedema |
P value, individuala [OR (95%CI), individual]c |
P value b | OR (95%CI)d | |
|---|---|---|---|---|---|
| No, N (%) | Yes, N (%) | ||||
| C4 CNV | |||||
| 4 | 265 (50.5) | 48 (49.0) | 0.826 | (Reference) | |
| < 4 | 100 (19.0) | 34 (34.7) | 0.001 [1.884 (0.538-6.597)] | 4.900 × 10-4 | 1.714 (0.447-6.575) |
| > 4 | 160 (30.5) | 16 (16.3) | 0.005 [0.617 (0.166-2.289)] | 0.761 (0.186-3.122) | |
| C4A CNV | |||||
| 2 | 336 (64.0) | 58 (59.2) | 0.364 | (Reference) | |
| < 2 | 57 (10.9) | 22 (22.5) | 0.003 [0.627 (0.164-2.404)] | 0.008 | 0.511 (0.122-2.134) |
| > 2 | 132 (25.1) | 18 (18.4) | 0.159 | 0.496 (0.117-2.106) | |
| C4B CNV | |||||
| 2 | 317 (60.4) | 59 (60.2) | 1 | (Reference) | |
| < 2 | 115 (21.9) | 28 (28.6) | 0.152 | 0.168 | 1.163 (0.298-4.542) |
| > 2 | 93 (17.7) | 11 (11.2) | 0.072 | 0.552 (0.125-2.443) | |
| C4 polymorphisms | |||||
| A2B2 | 217 (41.3) | 36 (36.7) | 0.434 | 0.050 | (Reference) |
| A2B1 | 63 (12.0) | 15 (15.3) | 0.405 | 1.895 (0.307-11.710) | |
| A3B2 | 58 (11.0) | 6 (6.1) | 0.202 | 1.371 (0.163-11.522) | |
| A2B3 | 41 (7.8) | 3 (3.1) | 0.130 | 0.558 (0.029-10.789) | |
| A3B1 | 26 (5.0) | 6 (6.1) | 0.619 | 0.333 (0.032-3.499) | |
| A1B2 | 23 (4.4) | 11 (11.2) | 0.013 [1.094 (0.137-8.709)] | 1.009 (0.103-9.841) | |
| Other | 97 (18.5) | 21 (21.4) | 0.735 (0.163-3.310) | ||
Abbreviations: GD, Graves' disease; GO, Graves' ophthalmopathy; CNV, copy number variation; OR, odds ratio; CI, confidence interval; N, number.
aIndividual C4 CNVs and polymorphisms between GD patients with or without myxedema were evaluated by Fisher's exact test using 2 × 2 contingency tables.
b CNV of C4, C4A and C4B between GD patients with or without myxedema were evaluated by Fisher's exact test using 3 × 2 contingency tables.C4 polymorphisms between GD patients with or without myxedema were evaluated by Fisher's exact test using 7 × 2 contingency tables. The p value was estimated by 100,000 Monte Carlo simulations with 99% confidence intervals (CI).
cORs and 95% CIs were estimated from logistic regression models adjusting for age, nodular hyperplasia, GO and vitiligo.
dORs and 95% CIs were estimated from logistic regression models adjusting for age, nodular hyperplasia, GO and vitiligo.