TABLE 2.
Risk HLA alleles for CBZ-, LTG-, and OXC-induced MPE.
| Allele | HLA genotype/total, n/N (%) | Cases vs. controls | Prevalence (%) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | NNT | ||
|---|---|---|---|---|---|---|---|---|---|---|
| MPE a | Controls a | p Value | OR (95%CI) | |||||||
| CBZ | 3.7 | |||||||||
| B*38:02 | 18/145 (12.41) | 10/179 (5.59) | 0.03 | 2.40 (1.07–5.37) | 12.41 | 94.41 | 7.86 | 96.56 | 218 | |
| A*24:02 | 44/140 (31.43) | 37/177(20.90) | 0.033 | 1.73 (1.04–2.88) | 31.43 | 79.10 | 5.46 | 96.78 | 86 | |
| /B*38:02 | ||||||||||
| OXC | 9.0 | |||||||||
| DRB1*04:06 | 8/51 (15.69) | 1/94 (1.06) | 0.002 b | 17.30 (2.10–142.72) | 15.69 | 98.94 | 59.41 | 92.23 | 71 | |
| Pooled | 2.8 | |||||||||
| A*24:02 | 56/253 (22.13) | 48/308 (15.58) | 0.047 | 1.54 (1.00–2.36) | 22.13 | 84.42 | 3.93 | 97.41 | 162 | |
| A*30:01 | 14/253 (5.53) | 7/308 (2.27) | 0.043 | 2.52 (1.00–6.34) | 5.53 | 97.73 | 6.56 | 97.29 | 646 | |
| B*35:01 | 11/261 (4.21) | 4/344 (1.16) | 0.02 | 3.74 (1.18–11.88) | 4.21 | 98.84 | 9.47 | 97.28 | 848 | |
| B*38:02 | 28/261 (10.73) | 20/344 (5.81) | 0.03 | 1.94 (1.07–3.54) | 10.73 | 94.19 | 5.05 | 97.34 | 333 | |
| DRB1*04:06 | 17/259 (6.56) | 9/343 (2.62) | 0.02 | 2.61 (1.14–5.95) | 6.56 | 97.38 | 6.73 | 97.31 | 544 | |
CBZ, carbamazepine; CI, confidence interval; LTG, lamotrigine; MPE, maculopapular exanthema; NPV, negative predictive value; NNT, number need to test; OR, odds ratio; OXC, oxcarbazepine; PPV, positive predictive value.
a
Several individuals were not subjected to HLA genotyping because of insufficient DNA. Due to 25 individuals were tolerant to CBZ and OXC, and 3 to LTG and OXC, we calculated these individuals by once when pooled analysis.
b
Significant after Bonferroni’s correction, p c = 0.038, n = 19 for HLA-DRB1*04:06 correction in OXC-induced MPE.