Multivariable regression models for association between permissive and nonpermissive DPB1 mismatches and aGVHD II-IV and TRM using the classic and “core” vs “non-core” TCE models
| Endpoints . | TCE (classic) . | . | TCE3 core vs non-core . | . | ||
|---|---|---|---|---|---|---|
| DP matching . | HR (95% CI) . | (Overall) P . | DP matching . | HR (95% CI) . | (Overall) P . | |
| aGVHD II-IV | Match | 1.00 | (<.0001) | Match | 1.00 | (.0002) |
| Permissive | 1.18 (1.03-1.35) | .0197 | Core permissive | 1.12 (0.98-1.28) | .1012 | |
| Nonpermissive | 1.32 (1.16-1.50) | <.0001 | Non-core permissive | 1.24 (1.06-1.46) | .0082 | |
| Non-permissive | 1.32 (1.16-1.50) | <.0001 | ||||
| TRM | Match | 1.00 | (.0328) | Match | 1.00 | (.0042) |
| Permissive | 0.99 (0.84-1.15) | Core permissive | 0.87 (0.72-1.05) | .1439 | ||
| Nonpermissive | 1.12 (0.97-1.30) | Non-core permissive | 1.07 (0.90-1.27) | .4482 | ||
| Nonpermissive | 1.12 (0.97-1.30) | .1209 | ||||
| Endpoints . | TCE (classic) . | . | TCE3 core vs non-core . | . | ||
|---|---|---|---|---|---|---|
| DP matching . | HR (95% CI) . | (Overall) P . | DP matching . | HR (95% CI) . | (Overall) P . | |
| aGVHD II-IV | Match | 1.00 | (<.0001) | Match | 1.00 | (.0002) |
| Permissive | 1.18 (1.03-1.35) | .0197 | Core permissive | 1.12 (0.98-1.28) | .1012 | |
| Nonpermissive | 1.32 (1.16-1.50) | <.0001 | Non-core permissive | 1.24 (1.06-1.46) | .0082 | |
| Non-permissive | 1.32 (1.16-1.50) | <.0001 | ||||
| TRM | Match | 1.00 | (.0328) | Match | 1.00 | (.0042) |
| Permissive | 0.99 (0.84-1.15) | Core permissive | 0.87 (0.72-1.05) | .1439 | ||
| Nonpermissive | 1.12 (0.97-1.30) | Non-core permissive | 1.07 (0.90-1.27) | .4482 | ||
| Nonpermissive | 1.12 (0.97-1.30) | .1209 | ||||
Models were adjusted or stratified for disease type, donor age, donor/recipient CMV match, GVHD prophylaxis, HCT-CI score, conditioning regimen, year of transplant, Karnofsky score, time from diagnosis to transplant, graft type, TBI use, and disease status as required. P values for specific groups are shown only for those models in which the overall P value was significant.