Analysis of Predictive Power of Prednisone Dose at 3 Months of Treatment on CGVHD Prognosis.

Current CGVHD prognostic and staging systems are still undergoing development and have identified plt count; CGVHD types progressive(P), quiescent(Q), de novo(DN); KPS, and GI involvement as significant risk factors affecting outcome. A simple reproducible staging system such as used for AGVHD to apply in clinical trials is still lacking. We evaluated whether the PSE dose required to control CGVHD at 3 months from diagnosis would have a prognostic effect on survival and in and of itself serve as a criteria for secondary intervention or investigational therapy. We hypothesized that by 3 months from start of treatment patients would be on the lowest dose dictated by medical necessity rather than by physician driven dose preference. A retrospective analysis of charts from 109 patients diagnosed with CGVHD between 6/2000 and 6/2003 was done. Data collected included age, donor type(mud/sib), plt count, CGVHD type(P/Q/DN), KPS, GI involvement. Outcome analysis included survival and cause of death. PSE dose was calculated in mg/kg at 3 months from first diagnosis of CGVHD. With a median follow up of 47.2 months(range 6.6–67.2) relapse censored survival of patients on a 3 month PSE dose of more than .3 mg/kg was 53% compared to all lower doses 86%(P.03). In a univariate analysis only PSE dose (P .05) and KPS (P <.01) were significant with plt count approaching significance (P .11). In a multivariate analysis again only PSE dose and KPS were significant (Table 1).

Table

.	.	.	.	Univariate .	.	Multivariate .	.
Variable .	Value .	N .	# of deaths .	Hazard Rate Ratio (95% CI) .	p value .	Hazard Rate Ratio (95% CI) .	p value .
Prednisone at 3 month	0	20	1	Baseline	0.05	Baseline
	0.1–0.14	18	2	3.0 (0.3–32.7)		3.7 (0.3–43.4)	0.33
	0.15–0.29	22	3	2.9 (0.3–28.2)		3.4 (0.3–34.5)	0.32
	0.3–0.59	16	7	9.8 (1.2–79.4)		9.3 (1.1–76.0)	0.04
	0.6–0.99	0	0	n/a		n/a	n/a
	≥1.0	0	0	n/a		n/a	n/a
cGVHD class	Denovo	29	5	Baseline	0.54
	Progressive	23	7	1.9 (0.6–6.0)
	Quiescent	57	13	1.3 (0.5–3.8)
Donor type	Sibling	48	9	Baseline	0.40
	Unrelated	61	16	1.4 (0.6–3.2)
Platelet count at 3 month	<100	28	9	Baseline	0.11	Baseline
	≥100	81	16	0.5 (0.2–1.1)		0.7 (0.2–2.3)	0.59
KPS at 3 month	<80	14	7	Baseline	<0.01	Baseline
	≥80	95	18	0.2 (0.1–0.5)		0.3 (0.0–1.3)	0.03
Lower GI at 3 month	No	98	22	Baseline	0.78
	Yes	9	2	1.2 (0.3–5.2)

.	.	.	.	Univariate .	.	Multivariate .	.
Variable .	Value .	N .	# of deaths .	Hazard Rate Ratio (95% CI) .	p value .	Hazard Rate Ratio (95% CI) .	p value .
Prednisone at 3 month	0	20	1	Baseline	0.05	Baseline
	0.1–0.14	18	2	3.0 (0.3–32.7)		3.7 (0.3–43.4)	0.33
	0.15–0.29	22	3	2.9 (0.3–28.2)		3.4 (0.3–34.5)	0.32
	0.3–0.59	16	7	9.8 (1.2–79.4)		9.3 (1.1–76.0)	0.04
	0.6–0.99	0	0	n/a		n/a	n/a
	≥1.0	0	0	n/a		n/a	n/a
cGVHD class	Denovo	29	5	Baseline	0.54
	Progressive	23	7	1.9 (0.6–6.0)
	Quiescent	57	13	1.3 (0.5–3.8)
Donor type	Sibling	48	9	Baseline	0.40
	Unrelated	61	16	1.4 (0.6–3.2)
Platelet count at 3 month	<100	28	9	Baseline	0.11	Baseline
	≥100	81	16	0.5 (0.2–1.1)		0.7 (0.2–2.3)	0.59
KPS at 3 month	<80	14	7	Baseline	<0.01	Baseline
	≥80	95	18	0.2 (0.1–0.5)		0.3 (0.0–1.3)	0.03
Lower GI at 3 month	No	98	22	Baseline	0.78
	Yes	9	2	1.2 (0.3–5.2)

View Large

View largeDownload slide

Overall survival by Prednisone dose

View largeDownload slide

Overall survival by Prednisone dose

Close modal

The analysis of this data to date suggests the PSE dose at 3 months may be an important independent clinical marker for subsequent CGVHD prognosis which could be used to decide which patients to include in 2^nd line and experimental therapy trials.