Evaluation of NIH consensus criteria for classification of late acute and chronic GVHD

Abstract

Historically, graft-versus-host disease (GVHD) beyond 100 days after hematopoietic cell transplantation (HCT) was called chronic GVHD, even if the clinical manifestations were indistinguishable from acute GVHD. In 2005, the National Institutes of Health (NIH) sponsored a consensus conference that proposed new criteria for diagnosis and classification of chronic GVHD for clinical trials. According to the consensus criteria, clinical manifestations rather than time after transplantation should be used in clinical trials to distinguish chronic GVHD from late acute GVHD, which includes persistent, recurrent, or late-onset acute GVHD. We evaluated major outcomes according to the presence or absence of NIH criteria for chronic GVHD in a retrospective study of 740 patients diagnosed with historically defined chronic GVHD after allogeneic HCT between 1994 and 2000. The presence or absence of NIH criteria for chronic GVHD showed no statistically significant association with survival, risks of nonrelapse mortality or recurrent malignancy, or duration of systemic treatment. Antecedent late acute GVHD was associated with an increased risk of nonrelapse mortality and prolonged treatment among patients with NIH chronic GVHD. Our results support the consensus recommendation that, with appropriate stratification, clinical trials can include patients with late acute GVHD as well as those with NIH chronic GVHD.

Figure 1

Outcomes according to acute or chronic GVHD at initial presentation. Cumulative incidence of (A) nonrelapse mortality and (B) recurrent malignancy and (C) overall survival are shown for patients who presented initially with chronic GVHD according to NIH criteria (n = 388; —) compared with those who presented initially with late acute GVHD (n = 352; ). In all 3 panels, progression to NIH chronic GVHD was treated as a competing risk for patients who presented initially with late acute GVHD.

Figure 2

Cumulative incidence of discontinued systemic treatment. (A) The cumulative incidence of discontinued systemic treatment among patients who presented NIH chronic GVHD (—) and the competing risks of death or recurrent malignancy during systemic treatment (). (B) The cumulative incidence of discontinued systemic treatment among patients who presented initially with late acute GVHD (—), together with the competing risks of death or recurrent malignancy during systemic treatment with () or without (- - -) the additional competing risk of transition to NIH chronic GVHD. Areas in each panel are labeled according to the initial event after starting treatment. The “Therapy continuing” areas include only patients without a competing event.

Figure 3

Risk factor comparisons. (A) Nonrelapse mortality. (B) Discontinued systemic treatment. Hazard ratio estimates and 95% confidence limits for each risk factor are shown for patients with chronic GVHD according to NIH criteria (n = 488; solid lines, filled symbols) and for patients who presented initially with late acute GVHD (n = 352; dashed lines, open symbols). Progression to NIH chronic GVHD was treated as a competing risk in the analysis of late acute GVHD, and the analysis of NIH chronic GVHD began when manifestations of GVHD first fulfilled NIH consensus criteria for the diagnosis.