An Acute Gvhd Classification Based on the Dynamic of Gvhd Skin Involvement from Its Appearance to Start of Systemic Treatment

The currently used classifications of acute GVHD are based on signs and symptoms in the skin, gut and liver, graded at the start of first-line therapy. However, they do not consider the rate at which these symptoms developed. Since skin aGVHD is the most frequent initial sign, we conducted a prospective observational study aimed at investigating the prognostic value of the dynamic of skin involvement such as the rate at which aGVHD progresses from cutaneous involvement to visceral involvement.

One hundred and eight consecutive patients who underwent allogeneic hematopoietic transplantation in our centre were registered for this observational study, their aGVHD episodes were classified according to both the modified Glucksberg classification and according to a “GHVD skin dynamic” proposed by us. Our classification “GVHD skin dynamic” classified acute GVHD based on the time interval between rash initial appearance to time when treatment using corticosteroids (CS) was started. In patients left untreated, we considered the change in skin rash extension shown during the first week without treatment. Median age of treated patients was 49 years, 69% were affected by Acute Leukemia, 64% had an alternative donor type (MUD, HAPLO). Primary aGVHD treatment was started in all patients in the case of progression to Glucksberg Overall Grade 2 or a persisting Overall Grade 1, no local corticosteroids were used on the skin. Eighty-two presented a skin rash as a first sign of a-GVHD (80.3%) and were classified according both to Glucksberg classification and our scheme; 18/82 started CS within 48 hours (Group 1); 13/82 started CS within days 3-7 (Group 2). Fifty-one patients (62.2%) were left untreated for 7 days (Group 3) and subclassified 3a and 3b. Group 3a (n 31) was defined when Skin GVHD Overall Grade 1, left untreated for 1 week, showed an absolute increase in involved body surface area <5 %. Group 3b (n 20), was defined when Skin GVHD Overall Grade 1, left untreated at 1 week, had an absolute increase in involved body surface area >5% (Fig 1). Demographic, disease-related and transplant features (Age, Donor type, Disease status, HSC Source, ATG use, Conditioning RIC, Diagnosis) in the four groups were not significantly different.

The four groups, identified by “GVHD skin dynamic”, had distinctive 2-y OS, (Fig 2). In Group 1 (patients who were started on CS within 48 h) the OS at 2y was 50.1% and TRM 41.6%. In Group 2 (patients who started on CS after 2 days and within 7 days) the OS at 2 y was 64.8% and TRM 16.8%. In Group 3a (patients who were started on CS after 7 days and had no increase of skin rash during days 1-7) the OS was 75.1% and TRM 10.4%. In Group 3b (patients who were started on CS after 7 days and had an increase of skin rash during days 1-7) the OS was 39.8% and TRM 39.2%.

Patients were then grouped based on “GVHD skin dynamic” into “Dynamic type poor risk” (Group 1 and Group 3b) and “Dynamic type good risk” (Group 2 and Group 3a). Such grouping was found statistically important using Cox proportional hazard univariate analysis for OS (p=0.04). Such grouping “Dynamic type poor risk” and “Dynamic type good risk” retained importance for OS when evaluated in Cox proportional hazard multivariate analysis (HR 2.222 ; 95% CL 1.017-4.875; p=0.04) together with age, status of the disease and conditioning.

Patients who were left untreated because of skin-only acute GVHD and who during a week had a worsening skin extension (Group 3b) were compared to those who when left untreated for one week showed a stable skin-only involvement (Group 3a). Overall Survival at 2 y in Group 3b was significantly inferior compared to Group 3a: 39.8% vs 75.1% (log-rank: p= 0.03). Group 3b had also increased Treatment-Related Mortality: 39.2% versus 10.4% vs (Gray test: p=0.05) and increased rate in need of secondary anti-GVHD treatment: 76.4% vs 46.6% (chi-test p=0.002).

The dynamic of skin aGVHD during the first week from appearance may be used to classify acute GVHD and this proposed acute GVHD classification has good prognostic value.

Further, patients who have a skin-only acute GVHD and who, when left untreated for one week show an increase in skin involvement, have inferior prognosis and should start systemic CS promptly.