Significance analysis and relative hazards (RHs) for factors affecting the risks of non-SCC death and development of SCC in SLH
Factors* . | Non-SCC Deaths (48 events) . | . | SCC (11 events) . | . | ||
|---|---|---|---|---|---|---|
| Treatment factors . | RH (95% CI)† . | Raw P value‡(adjusted P value)§ . | RH (95% CI)† . | Raw P value (adjusted P value) . | ||
| Antithymocyte globulin | 4.1 (2.2-7.3) | 8.3 × 10-6 (6.7 × 10-5) | .0¶ | .44 | ||
| Other GVHD prophylaxis | 2.9 (1.6-5.1) | 2.6 × 10-4 (.001) | 1.8 | .51 | ||
| Total lymphoid irradiation at 5Gy | 0.4 (0.2-0.8) | .01 (.03) | 1.9 | .90 | ||
| CR with busulphan | 5.0 (1.3-13.6) | .02 (.04) | 0 | .99 | ||
| Total body irradiation at 5Gy | 2.0 | .11 | 0 | .69 | ||
| CD34+ cell selection | 1.9 | .18 | 0 | .87 | ||
| CR with fludarabine | 1.6 | .71 | 0 | .99 | ||
| CR with cyclophosphamide | ND∥ | 1.00 | ND∥ | .99 | ||
| Host factors | ||||||
| Acute GVHD: grade III + IV versus 0-II | 13.8 (7.8-22.2) | 4.0 × 10-17 (1.6 × 10-16) | 32.8 (2.7-392) | .006 (.026) | ||
| Age at transplantation: at least 10.8 y vs younger than 10.8 y | 2.7 (1.5-4.9) | 5.2 × 10-4 (.001) | 2.2 | .22 | ||
| Chronic GVHD: extensive vs none + limited; f/u beyond 6 mo | 2.3 | .29 | ∞¶ | .02 (.03) | ||
| Sex of patient: male vs female | 1.0 | .99 | 2.9 | .14 | ||
| Donor factors | ||||||
| HLA: alternative donors vs matched siblings | 5.4 (3.0-9.7) | 3.3 × 10-9 (1.0 × 10-8) | 0.8 | .99 | ||
| Sex of donor: male vs female | 1.8 (1.0-3.3) | .052 (.08) | 1.5 | .70 | ||
| Year of transplantation: 1993 or later vs before 1993 | 1.7 | .09 | 0 | .11 | ||
Factors* . | Non-SCC Deaths (48 events) . | . | SCC (11 events) . | . | ||
|---|---|---|---|---|---|---|
| Treatment factors . | RH (95% CI)† . | Raw P value‡(adjusted P value)§ . | RH (95% CI)† . | Raw P value (adjusted P value) . | ||
| Antithymocyte globulin | 4.1 (2.2-7.3) | 8.3 × 10-6 (6.7 × 10-5) | .0¶ | .44 | ||
| Other GVHD prophylaxis | 2.9 (1.6-5.1) | 2.6 × 10-4 (.001) | 1.8 | .51 | ||
| Total lymphoid irradiation at 5Gy | 0.4 (0.2-0.8) | .01 (.03) | 1.9 | .90 | ||
| CR with busulphan | 5.0 (1.3-13.6) | .02 (.04) | 0 | .99 | ||
| Total body irradiation at 5Gy | 2.0 | .11 | 0 | .69 | ||
| CD34+ cell selection | 1.9 | .18 | 0 | .87 | ||
| CR with fludarabine | 1.6 | .71 | 0 | .99 | ||
| CR with cyclophosphamide | ND∥ | 1.00 | ND∥ | .99 | ||
| Host factors | ||||||
| Acute GVHD: grade III + IV versus 0-II | 13.8 (7.8-22.2) | 4.0 × 10-17 (1.6 × 10-16) | 32.8 (2.7-392) | .006 (.026) | ||
| Age at transplantation: at least 10.8 y vs younger than 10.8 y | 2.7 (1.5-4.9) | 5.2 × 10-4 (.001) | 2.2 | .22 | ||
| Chronic GVHD: extensive vs none + limited; f/u beyond 6 mo | 2.3 | .29 | ∞¶ | .02 (.03) | ||
| Sex of patient: male vs female | 1.0 | .99 | 2.9 | .14 | ||
| Donor factors | ||||||
| HLA: alternative donors vs matched siblings | 5.4 (3.0-9.7) | 3.3 × 10-9 (1.0 × 10-8) | 0.8 | .99 | ||
| Sex of donor: male vs female | 1.8 (1.0-3.3) | .052 (.08) | 1.5 | .70 | ||
| Year of transplantation: 1993 or later vs before 1993 | 1.7 | .09 | 0 | .11 | ||
F/u indicates follow-up.
Treatment, host, and donor factors are analyzed as separate families of hypotheses, as described in “Patients, materials, and methods.” CR is conditioning regimen; GVHD is graft-versus-host disease. The median age at transplantation was 10.8 years, and the median year of transplantation was 1993. Other GVHD prophylaxis contrasts cyclosporin A (CSA) alone versus any of the following “other” regimens: CSA and monoclonal antibody, steroids, CSA and steroids, methotrexate, CSA and methotrexate, CSA and antithymocyte globulin and steroids. Total lymphoid irradiation (TLI) and total body irradiation (TBI) contrast TLI and TBI versus no TLI and no TBI, respectively.
Relative hazard (RH) and exact 95% confidence interval (CI) and P value estimated by rate-ratio analysis, except for analyses of host factors and SCC. In those analyses, RHs were obtained using the proportional hazards models, because time was associated with the risk factors acute and chronic GVHD and age at transplantation, and with the outcome SCC.
P values unadjusted for multiple comparisons.
P values adjusted for multiple comparisons using the Benjamini-Hochberg FDR procedure are shown whenever the raw P values are less than or nearly equal to .05.
ND indicates not determined. In 117 patients, 113 received cyclophosphamide and 4 did not. Rate-ratio analysis was not informative.
An RH of infinity is obtained if all 11 SCC events occur in a subgroup of interest. An RH of 0 is obtained if none of the 11 SCC events occur in a subgroup.