Determinants of Platelet Counts in Patients with Chronic Graft Versus Host Disease.

Although a low platelet count in chronic graft-versus-host disease (cGVHD) patients is a consistent negative survival predictor across many cGVHD studies, this association is poorly understood. Additionally, a subgroup of cGVHD patients has elevated platelets, which is also unexplained. We hypothesized that such platelet abnormalities may reflect the pathophysiology of cGVHD and therefore be associated with specific clinical manifestations. This exploratory study was conducted to investigate determinants and clinical significance of platelet counts in a large cohort of cGVHD patients.

cGVHD patients were enrolled in the ongoing cross sectional National Cancer Institute protocol (clinicaltrials.gov#NCT00331968). A detailed history, physical and extensive laboratory evaluations were obtained using standardized methods. National Institutes of Health (NIH) severity scores were based on the extent of cGVHD manifestations, degree of functional impairment and the number of affected organs. Wilcoxon rank sum test, Kruskal-Wallis test, and Jonckheere -Terpstra trend test were used in statistical analyses. Results with p-values<0.01 were considered as statistically significant, and those with Spearman rank correlation coefficient 0.3< r <0.5 as moderate to weakly correlated. Multiple regression analysis was performed with potentially predictive variables.

The cohort included 120 adult cGVHD patients (median age 48 years, 53% male (M), 47% female (F)). Median time from transplantation to study enrollment was 2.7 years. Median time of cGVHD diagnosis was 0.5 year after transplantation, and median time from cGVHD diagnosis to enrollment was 1.6 years. 75% of subjects received related donor transplant; 84% received peripheral blood stem cells and 16% bone marrow; 63% had a history of acute GVHD (skin 88%, gastrointestinal 45%, liver 22%). Median total NIH organ severity score was 6 (range 2-14), median average total NIH organ severity score was for F 0.750 (range 0.250-1.625), and for M 0.857 (range 0.286-2.000), and median number of organs involved was 4 (range 1-8). Median platelet count was 239,000/μL (range 33,000/μL-648,000/μL). F had significantly higher platelet counts than M (median 275,500 /μL vs. 221,500/μL, p=0.004). Time from transplantation (r=0.35) and time from cGVHD diagnosis (r=0.32) correlated positively with platelet counts (p=0.001). Patients who received TBI based conditioning tended to have higher platelet counts (median 279,000/μL vs. 211,000/μL, p=0.021). Platelet counts varied among the four recipient-donor gender pairings (M recipient/M donor had median platelets 199,500/μL, M recipient/F donor had median platelets 239,500/μL, F recipient/F donor had median platelets 257,000/μL, F recipient/M donor had median platelets 301,000/μL, p=0.010). Among patients with history of acute GVHD, those with liver acute GVHD tended to have higher platelet counts (292,000/μL vs. 226,500 /μL, p=0.027). Absolute lymphocyte count (r=0.35), absolute CD3s (r=0.31), absolute CD4s (r=0.31), absolute NK cells count (r=0.32), and calcium levels (r=0.33) correlated positively with platelet counts (p<0.01). NIH scores for involvement of joints/fascia (p=0.021) and lungs (p=0.039), maximal NIH individual (0-3) organ score (p=0.013), number of involved organs with cGVHD (p=0.0007), total NIH score (p=0.0002), and average total NIH organ severity score (r=0.28; p=0.002) were positively associated with platelet counts. None among the panel of laboratory parameters of acute inflammation correlated with platelet counts. In the linear multiple regression model, female patients (p=0.001), higher absolute lymphocytes count (p<0.0001) and higher average total NIH score (p<0.0001) remained the strongest simultaneous predictors of higher platelet counts.

This exploratory analysis demonstrates strong positive association of platelet counts with the lymphocyte counts and with the average total NIH cGVHD organ severity scores. These data implicate abnormalities of platelet counts in cGVHD pathophysiology and warrant further research.

No relevant conflicts of interest to declare.