Table 3.
Univariate and multivariate analysis of risk factors for progression of prior IA after transplantation
|
Progression < d 30
|
Progression after d 30
|
Overall follow-up
|
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
|
Multivariate
|
Multivariate
|
Multivariate
|
|||||||||
| No. patients | 2-y incidence of IA progression (95% CI) | Univariate P | P | HR (95% CI) | Univariate P | P | HR (95% CI) | Univariate P | P | HR (95% CI) | |
| Impact during the entire posttransplantaton period | |||||||||||
| Duration of neutropenia*† | — | < .001 | < .01 | 10 (4-67)‡ | .01 | .01 | 10 (2.6-40) | < .001 | < .001 | 10.6 (5.4-37) | |
| Status of the underlying disease | .07 | .3 | .01 | .15 | < .01 | .01 | 7 (1.6-30) | ||||
| Not early | 85 | 30 (16-43) | |||||||||
| Early | 44 | 5 (-1-11) | |||||||||
| Wk between start TxIA-Allo-HSCT†§ | |||||||||||
| Less than 6 | 18 | 34 (9-58) | .04 | .02 | 4.6 (1.3-16) | .06 | .11 | < .01 | .01 | 3.6 (1.4-9.4) | |
| At least 6 | 111 | 16 (7-25) | |||||||||
| Response status of the IA at HSCT§ | |||||||||||
| SD or progression | 21 | 32 (10-54) | .1 | NI§ | NI§ | .04 | NI§ | NI§ | .01§ | — | —§ |
| CR or PR | 108 | 17 (8-26) | |||||||||
| Pharmacologic prophylaxis for GVHD | .03 | .11 | — | .04 | .2 | .03 | |||||
| CsA or Tacro + MTX | 76 | 16 (9-30) | .08 | — | |||||||
| Other non-MTX | 53 | 26 (16-36) | |||||||||
| Impact during the early posttransplantation period (< d 30)∥ | |||||||||||
| Type of conditioning | .02 | .054 | 3.4 (0.98-12) | .5 | — | — | .4∥ | — | — | ||
| CONV | 72 | 21 (9-33) | |||||||||
| RIC | 57 | 22 (8-32) | |||||||||
| Impact during the late posttransplantation period (> d 30)∥ | |||||||||||
| CMV disease* | NA | NA | NA | < .001 | |||||||
| Yes | 12 | 40 (10-70) | .02 | 4.2 (1.4-17) | .02∥ | .04 | 3.7 (1.3-11) | ||||
| No | 117 | 15 (7-23) | |||||||||
| Stem cell source | 21 (8-34) | .4 | — | — | < .001 | ||||||
| BMT or CBT | 30 | 15 (3-27) | .01 | 9.8 (9-99) | .4∥ | — | — | ||||
| PBSC | 99 | ||||||||||
| Acute GVHD ≥ grade II that required HD steroids (> 1 wk) and/or ATG* | NA | NA | NA | .01 | .04 | 10 (1.7-29) | .04 | .3 | — | ||
| Yes | 53 | 31 (15-47) | |||||||||
| No | 76 | 16 (6-26) | |||||||||
| Variables with no impact | |||||||||||
| Patient age† | — | .96 | — | — | .93 | — | — | .96 | — | — | |
| Donor type | .9 | — | — | .93 | — | — | .9 | — | — | ||
| Alternative donor | 58 | 22 (8-36) | |||||||||
| HLA-identical sibling | 71 | 18 (7-29) | |||||||||
| Type of IA | .9 | — | — | .9 | — | — | .9 | — | — | ||
| Proven | 49 | 22 (9-35) | |||||||||
| Probable | 80 | 18 (8-28) | |||||||||
| Surgical resection before allo-HSCT | .4 | — | — | .3 | — | — | .35 | — | — | ||
| Yes | 28 | 18 (0-36) | |||||||||
| No | 101 | 23 (13-333) | |||||||||
| Chronic GVHD* | NA | NA | NA | .8 | — | — | .8 | — | — | ||
| Yes | 45 | 24 (9-39) | |||||||||
| No | 84 | 20 (9-31) | |||||||||
| ATG or alemtuzumab in conditioning | .4 | — | — | .5 | — | — | .6 | — | — | ||
| Yes | 60 | 23 (5-40) | |||||||||
| No | 69 | 17 (6-28) | |||||||||
| Ex vivo T-cell depletion | .8 | — | — | .7 | — | — | .8 | — | — | ||
| Yes | 22 | 22 (6-38) | |||||||||
| No | 107 | 26 (9-27) | |||||||||
| 2 or more post-HSCT serum samples with positive GM, index ≥ 0.8 (%)* | .03 | NT§ | NT§ | .2 | — | — | .002† | NT | NT¶ | ||
| No | 43 | 2/43 (4.7) | |||||||||
| Yes | 24 | 11/24 (46) | |||||||||
| Not done¶ | 62 | 14/62 (22.6) | |||||||||
Data for progression before day 30 reflect 13 patients out of 129 evaluable patients; for progression after day 30, data reflect 14 patients out of 107 evaluable. Overall follow-up data are for 27 patients.
NA indicates not applicable, since no patient developed this outcome variable before day 30 after transplantation; HD (high-dose) steroids, prednisone ≥ 2 mg/kg; TxIA-Allo-HSCT, interval from start of AFT for IA and allo-HSCT; NT, not tested (see “Statistical analysis” and “Results” for details); NI, not included; —, not applicable (since P > .01).
Time-dependent variables.
Quantitative variables that were found to have an impact on any outcome were reanalyzed as categorical variables.
The HR shown refers to risk increase with every 5-day delay in the time for recovery of the absolute neutrophil count.
These 2 variables show collinearity, and for multivariate analysis the variable used was < 6 weeks from start of treatment of IA and the Allo-HSCT, because it showed a higher hazard ratio in univariate Cox regression and it is a more objectively measurable variable than the response of the IA at the time of transplantation.
These variables had a nonproportional hazard ratio over time of their impact on the risk of progression of the IA, which differed before and after day + 30 post-HSCT (see “Statistical analysis”).
Since only 67 patients (52%) were screened in the immediate posttransplantation period with serum GM performed at least twice weekly, this variable was not included in the multivariate analysis.