Validation of the NIH Criteria of the Severity Grading of Chronic Gvhd – Results of a Prospective German Multicenter Validation Trial.

Abstract 204

The NIH staging and response criteria offer for the first time the chance for uniform documentation of chronic graft-versus-host disease (cGVHD). Here we present the validation of the cGVHD staging criteria as part of an interim analysis of a prospective German/Austrian multicenter study on the NIH staging criteria in cGVHD. One hundred-seventeen patients (median age 44 years, range 21–72) after allogeneic hematopoietic stem cell transplantation (HSCT) were evaluated according to the NIH criteria based on cGVHD activity assessment including the NIH-cGVHD grading form, the physicians cGVHD activity assessment form, the NIH-cGVHD symptom self assessment form, and the Lee cGVHD symptom-scale (L-cGVHD-SC). Enrolment occurred a median of 443 (range, 100–4003) days after HSCT. The validation of the NIH staging criteria was performed using the first 3 surveys each one month apart. For the purpose of validity a second evaluation was performed including only patients with stable cGVHD (n=64) as judged by the physician (results in parentheses). The statistical significance p was <0.01 for all reported results. Severity of cGVHD was mild in 33 patients, moderate in 50, and severe in 34, respectively. For the global NIH cGVHD grading (mild, moderate, severe) the test-retest stability between the three evaluations showed a high correlation ranging from 0.67–0.75 (0.7–0.78). Interestingly, the physician′s severity grading using the 10 point scale demonstrated a higher correlation ranging from 0.82–0.87 (0.88–0.92). The test-retest stability was highest for skin: 0.82–0.85 (0.91–0.93), mouth: 0.7–0.75 (0.76–0.8) and joints: 0.75–0.79 (0.68–0.86) followed by liver: 0.58–0.76 (0.63), eye: 0.64–0.68 (0.6–0.62) and genital: 0.65–0.87 (0.83–0.95), while it was relatively low for gastrointestinal (GI): 0.43–0.44 (0.38) and lung manifestations: 0.46–0.49 (0.6) respectively. With regard to the patients' self assessment the test-retest stability ranged for the 10 point scale between 0.76–0.79 (0.8), and for the mild/moderate/severe grading from 0.68–0.75 (0.83–0.85). Mouth dryness was reported with a correlation of 0.74–0.79 (0.83–0.84), mouth pain with 0.68–0.79 (0.74–0.84) and mouth sensitivity with 0.77–0.80 (0.81–0.88). Eye complaints were reported with high stability: 0.82 (0.81–0.86). Comparison of physicians' organ grading with patients' organ specific self assessment of the NIH self assessment form and the L-cGVHD-SC organ scores revealed a moderate correlation of the physicians' skin grading with the skin L-cGVHD-SC subscale (r=0.55), the mouth grading with the mouth L-cGVHD-SC subscale (r=0.32), mouth dryness (r=0.42), mouth sensitivity (r=0.44), and mouth pain (r=0.5) of the NIH self assessment form. The physicians' eye grading correlated with the eye and mouth L-cGVHD-SC subscale (r=0.62) and eye complaints in the patients' NIH self assessment form (r=0.65). The lung grading correlated moderately with the breath subscale of the L-cGVHD-SC (r=0.5). While the joint grading correlated moderately only with the muscle and joint subscale of the L-cGVHD-SC, the GI grading did not correlate with the eating and digestion part of the L-cGVHD-SC. The results demonstrate a high test-retest stability for most of the organs and the global grading indicating its robustness in clinical use. The low test-retest correlation in the lung and GI grading including failure to correlate with the eating and digestion subscale of the L-cGVHD-SC may relate to the solely functional staging and the multifactorial pathogenesis of symptoms. This is supported by the correlation of lung symptoms in the L-cGVHD-SC breath subscale with lung grading.