Background We compared 6586 UBMT (1993–2005) and 2713 UCBT (1997–2005) in Japan, where genetic background is more homogeneous and medical practices are more uniform than in USA or European countries. Patients and Methods Data on UBMT were collected by JMDP and UCBT by JCBBN. Patients were divided into children under 16 years, young adults (16–49 years) and elderly (50 years or older). Proportion of underlying diseases was similar in each group, except that transplants in severe aplastic anemia and chronic myelogenous leukemia were less in UCBT than in UBMT. The proportion of second or more transplants was higher in UCBT than in UBMT (22.9% vs 5.4%, p<0.001). Therefore, comparisons between UBMT and UCBT were made only in the first transplants. HLA compatibility was analyzed by allele typing.


  1. Engraftment

    1. Neutrophil(500/μl); median days in UBMT and in UCBT were 18 vs 23 (children), 17 vs 25 (young adults), and 17 vs 24 (elderly). Differences were statistically significant (p<0.001).

    2. Platelet (50K/μl):31 vs 59 (children), 30 vs 54 (young adults), and 32 vs 73 (elderly). (p<0.001)

  2. GVHD

    1. Acute GVHD ≥II; probability in UBMT and in UCBT were 50% vs 47% (children), 45% vs 44% (young adults), and 46% vs 44% (elderly). No significant differences.

    2. Chronic GVHD; 40% vs 20% (children), 50% vs 30% (young adults), and 47% vs 24% (elderly). (p<0.001)

  3. Survival in childhood non-malignant diseases

    1. Overall survival; 75%(N=56) vs 69% (N= 54) (immunedeficiency, n.s.), 81%(N=197) vs 51%(N=15) (aplastic anemia, P=0.001), and 75%(N=56) vs 69%(N=54) (inborn error of metabolism, n.s.).

    2. Event free survival (EFS); 64% vs 53% (ID, n.s.), 69% vs 32% (SAA, p<0.001), and 62% vs 19% (IEM, p=0.001).

  4. 5-year-EFS in malignant diseases

    1. childhood ALL; 1+2CR: 66% (HLA matched BM, N=179), 56% (mismatched BM, N=106), 63% (MA CB, N=19) and 45% (MM CB, N=127) (p<0.001). Non CR: 25%(MA BM, N=53), 18%(MM BM, N=45), 12% (MM CB, N=30). (n.s.)

    2. childhood AML; 1+2CR: 65% (MA BM, N=100), 58% (MM BM, N=47), and 53% (MM CB, N=50) (n.s.) Non CR: 16% (MA BM, N=30), 20% (MM BM, N=25) and 15% (MM CB, N=40).(n.s.)

    3. young adult ALL; 1+2CR: 52% (MA BM, N=374), 36% (MM BM, N=188), and 38% (MM CB, N=96).(p<0.001) Non CR: 15% (MA BM, N=135), 14% (MM BM, N=112), and 12% (MM CB, N=40). (n.s.)

    4. young adult AML; 1+2CR: 60% (MA BM, N=421), 53% (MM BM, N=171), and 55% (MM CB, N=106).(p<0.001) Non CR: 20% (MA BM, N=235), 21% (MM BM, N=161), and 13% (MM CB, N=112). (p=0.012.)

    5. elderly ALL; 1+2CR: 52% (MA BM, N=48), 48% (MM BM, N=16), and 33% (MM CB, N=39).(p<0.001) Non CR: 6% (MA BM, N=20), 20% (MM BM, N=10), and 12% (MM CB, N=17). (n.s.)

    6. elderly AML; 1+2CR: 37% (MA BM, N=68), 51% (MM BM, N=23), and 32% (MM CB, N=47).(p<0.001) Non CR: 7% (MA BM, N=64), 26% (MM BM, N=18), and 14% (MM CB, N=92). (n.s.)

Conclusion HLA allele matched UBMT gave better outcome than mismatched UBMT or matched/mismatched UCBT, but allele mismatched UBMT and UCBT were not different. UCBT should be chosen in early stages of the disease if UBMT from an HLA allele matched unrelated donor cannot be scheduled in the suitable period.

Author notes

Disclosure: No relevant conflicts of interest to declare.