Impact of HLA-B alleles on transplantation-related outcomes based on linkage to high- and low-risk HLA alleles
| Clinical endpoint . | Odds ratio (95% CI) . | P . |
|---|---|---|
| GI aGVHD (grade II-IV)* | .059 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.04 (0.75-1.46) | |
| ≥ 1 high-risk allele | 0.63 (0.38-1.04) | |
| GI aGVHD (grade III-IV)* | .88 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.90 (0.58-1.40) | |
| ≥ 1 high-risk allele | 0.80 (0.44-1.44) | |
| Overall aGVHD (grade II-IV)* | .41 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.99 (0.79-1.24) | |
| ≥ 1 high-risk allele | 1.18 (0.89-1.57) | |
| Overall aGVHD (grade III-IV)* | .60 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.90 (0.68-1.19) | |
| ≥ 1 high-risk allele | 1.05 (0.73-1.51) | |
| Chronic aGVHD* | .15 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.15 (0.89-1.48) | |
| ≥ 1 high-risk allele | 1.38 (1.0-1.90) | |
| Relapse* | .17 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.35 (0.94-1.93) | |
| ≥ 1 high-risk allele | 1.01 (0.63-1.60 | |
| Treatment-related mortality* | .11 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.76 (0.59-0.98) | |
| ≥ 1 high-risk allele | 0.80 (0.58-1.12) | |
| Disease-free survival† | .64 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.96 (0.84-1.11) | |
| ≥ 1 high-risk allele | 0.91 (0.76-1.10) | |
| Overall survival† | .45 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.95 (0.82-1.10) | |
| ≥ 1 high-risk allele | 0.89 (0.73-1.07) |
| Clinical endpoint . | Odds ratio (95% CI) . | P . |
|---|---|---|
| GI aGVHD (grade II-IV)* | .059 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.04 (0.75-1.46) | |
| ≥ 1 high-risk allele | 0.63 (0.38-1.04) | |
| GI aGVHD (grade III-IV)* | .88 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.90 (0.58-1.40) | |
| ≥ 1 high-risk allele | 0.80 (0.44-1.44) | |
| Overall aGVHD (grade II-IV)* | .41 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.99 (0.79-1.24) | |
| ≥ 1 high-risk allele | 1.18 (0.89-1.57) | |
| Overall aGVHD (grade III-IV)* | .60 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.90 (0.68-1.19) | |
| ≥ 1 high-risk allele | 1.05 (0.73-1.51) | |
| Chronic aGVHD* | .15 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.15 (0.89-1.48) | |
| ≥ 1 high-risk allele | 1.38 (1.0-1.90) | |
| Relapse* | .17 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 1.35 (0.94-1.93) | |
| ≥ 1 high-risk allele | 1.01 (0.63-1.60 | |
| Treatment-related mortality* | .11 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.76 (0.59-0.98) | |
| ≥ 1 high-risk allele | 0.80 (0.58-1.12) | |
| Disease-free survival† | .64 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.96 (0.84-1.11) | |
| ≥ 1 high-risk allele | 0.91 (0.76-1.10) | |
| Overall survival† | .45 | |
| 2 low-risk alleles | 1 | |
| Intermediate | 0.95 (0.82-1.10) | |
| ≥ 1 high-risk allele | 0.89 (0.73-1.07) |
Allogeneic transplant recipients (n = 1676) were assigned to high-, intermediate-, or low-risk groups based on known linkage between HLA-B and MICA.5-8 MICA alleles (and the associated HLA-B alleles) that trended toward GI aGVHD protection (*008) were considered low-risk. MICA alleles associated with increased GI aGVHD risk (*004, *006, *009, *044, and *049) were considered high risk. All others were coded as intermediate risk. To investigate risk factors for MICA-associated HLA-B alleles and aGVHD in the cohort analysis, cumulative incidence rates of aGVHD (the chance a patient will have experienced aGVHD event before time t, and where death without the aGVHD event is the competing risk) were calculated using methods previously described.9 Multivariate analyses were applied to adjust for the effects of other covariates on the cumulative incidence of aGVHD using the pseudovalue approach of Klein10,11 with a logistic link function. A forward stepwise regression model using a generalized linear model for the pseudovalues was used. Shown are the P values comparing the clinical endpoint for recipients with 2 low-risk alleles to either intermediate or ≥ 1 high-risk allele.
CI indicates confidence interval.
Pseudovalue technique.
Cox regression.