Re-Evaluation of Chronic Graft-Versus-Host Disease Using National Institute of Health Consensus Criteria.

Background: The National Institutes of Health (NIH) proposed new consensus criteria for the diagnosis and scoring of chronic graft-versus-host disease (cGVHD). Using the new system, we re-evaluated the patients with cGVHD that was diagnosed by classic criteria.

Methods: Of 618 patients who underwent allogeneic hematopoietic cell transplantation (HCT) from December 1994 to April 2008 at the Asan Medical Center, Seoul, Korea, we retrieved 236 patients who had cGVHD by classic criteria from the AMC BMT Registry. Among 236 patients, 20 patients with liver-only involvement could not be diagnosed as cGVHD by the NIH criteria, thus we reclassified and graded 216 patients according to the NIH criteria. We also evaluated the ability of the NIH criteria to stratify and predict the risk of cGVHD patients, as assessed by GVHD-specific survival (GSS).

Results: Twenty patients (9.3%) were reclassified as acute GVHD by NIH criteria (‘classic acute’ in 7 and ‘persistent, recurrent, or late-onset acute’ in 13) and 196 patients (90.7%) remained as chronic GVHD (‘classic chronic’ [Cl-Ch] in 170 and ‘overlap syndrome’ [Ov-Sy] in 26). Median age of 196 patients with cGVHD by NIH criteria, 119 males and 77 females, was 35.5 years (range, 15 to 57). Acute GVHD preceded cGVHD in 70 patients (35.7%). The probability of GSS at 5 years was 86.2% with 22 cGVHD-related deaths. The GSS was significantly different between two subtypes of cGVHD by NIH criteria: 88.6% for Cl-Ch vs 70.2% for Ov-Sy (p=0.002). NIH global scoring system stratified risk of cGVHD patients better than stage by classic criteria at both onset and peak of cGVHD (Table 1). We evaluated 12 variables at onset of cGVHD to determine their prognostic significance for GSS. Multivariate analysis demonstrated that NIH global score at onset (mild vs moderate, HR 6.1, p=0.027; mild vs. severe, HR 7.0, p=0.015), preceding aGVHD (no vs. yes, HR 6.2, p=0.001), and number of HLA ABDR mismatch (0 vs. 1, HR 2.0, p=0.555; 0 vs 2, HR 200.4, p=0.009) were independent predictors for GSS.

Conclusion: Our results indicate that a new NIH system can provide a proper risk-stratification of patients with cGVHD and global scoring system at onset of cGVHD can predict the prognosis of patients.

Table 1. GVHD-specific survival according to NIH global scoring system or stage by classic criteria