The biochip addressing the biotransformation gene polymorphisms was used to examine the DNA of 403 children with acute leukemia. The genotype frequencies in children with primary acute leukemia or relapse are summarized in Tables 2 to 11.

Distribution of CYP1A1 Genotypes. The CYP1A1 genotype frequency was studied in patients with primary leukemia or relapse (Table 2). We observed both common (*1/*1, *1/*2A, *1/*2B and *1/*4) and rare allele combinations such as *2B/*4, *2B/*2B, and*2A/*2A). Genotype frequency of the CYP1A1 genotype *1/*2A was higher in patients with ALL or AML relapse than in those with primary acute leukemia (Table 3). The difference was significant for patients with ALL (OR = 2.29, 95% CI = 1.09-4.80, p = 0.0322). The group of relapsed patients (not separated according to leukemia variant) showed a significant increase in frequency of the CYP1A1 genotype *1/*2A when compared with primary leukemia patients (OR = 2.11, 95% CI = 1.11-4.02, p = 0.0291). Because the CYP1A1 *2A/*2A genotype is very rare (genotype frequency is 0.4 in patients with primary ALL and 0.0 in all other groups), it was not taken into account.

Distribution of GSTT1 Genotypes. An opposite tendency was observed for GSTT1: the null genotype was far less frequent in ALL or AML relapse than in primary leukemia. The difference was significant in the case of ALL (OR = 0.48, 95% CI = 0.26-0.90, p = 0.0227; Table 4). The null genotype frequency in patients with relapse (not separated according to leukemia variant) was 1.8-fold less than in primary leukemia patients (OR = 0.55, 95% CI = 0.33-0.92, p = 0.0265).

Combined Analysis of CYP1A1 Genotype *1/*2A and GSTs Genotypes.Patients with ALL and/or AML relapse had a higher frequency of CYP1A1 genotype *1/*2A in combination with the GSTT1 non null genotype, GSTM1 non null genotype, or the double GSTT1/GSTM1 non null genotype when compared with primary leukemia patients (Table 5). The patients with relapse also had a significantly higher frequency of CYP1A1 genotype *1/*2A and the double GSTT1/GSTM1 non null genotype when compared with those with primary leukemia (OR = 3.48, 95% CI = 1.23-9.84, p = 0.0293).

Distribution of NAT2 Genotypes. The NAT2 allele frequency distribution in patients with ALL and AML revealed a characteristic genotype for each type of leukemia, whose presence increases the risk of relapse. Patients with ALL relapse had a higher frequency of heterozygotes and homozygotes for the NAT2 mutations (heterozygous and homozygous variants) 341C/– and 481T/–, and for their combination, when compared with primary leukemia patients (data not shown). An opposite tendency was observed for polymorphic variants 590G>A and 857G>A: the heterozygous and homozygous variants 590A/– and 857A/–, and their combination, were less frequent in ALL relapse than in primary leukemia. In total, patients with ALL relapse showed a significantly higher frequency of genotype 341C/–, 481T/–, 590G/G, 857G/G when compared with primary leukemia patients (OR = 1.64, 95% CI = 0.97-2.78, p = 0.074).

The patients with AML relapse had a lower frequency of heterozygous and homozygous variants 341C/– and 481T/-, and their combination, when compared with primary leukemia patients (data not shown). They also had a higher frequency of the heterozygous and homozygous variant 590A/- when compared with primary leukemia patients. In total, patients with AML relapse had a significantly higher frequency of genotype 341T/T, 481C/C, 590A/– when compared with primary leukemia patients (OR = 2.37, 95% CI = 0.86-6.54, p = 0.125).

Combined Analysis of NAT2 and GSTs Genotypes. The NAT2 genotype 341C/–, 481T/–, 590G/G, 857G/G in combination with the GSTT1 and/or GSTM1 non null genotype occurred in patients with ALL relapse at a significantly higher frequency than in those with primary leukemia (Table 6). Patients with AML relapse also displayed a higher frequency of NAT2 genotype 341T/T, 481C/C, 590A/– in combination with the GSTT1 and/or GSTM1 non null genotype when compared with primary leukemia patients. The difference was significant for the combination of the NAT2 genotype with the GSTT1 non null genotype (OR = 3.38, 95% CI = 1.08-10.57, p = 0.048; Table 7).

Distribution of the CYP1A1 Genotype *1/*2A and GSTs Genotypes in Boys with Acute Leukemia. Combinations of rhe CYP1A1 Genotype *1/*2A and GSTs Genotypes. Analysis of the CYP1A1 genotype frequency showed that boys with ALL relapse had a significantly higher frequency of CYP1A1 genotype *1/*2A when compared with boys with primary leukemia (OR = 2.90, 95% CI = 1.16-7.22, p = 0.033; Table 8). Boys with ALL or AML relapse had a lower frequency of the GSTT1 null genotype when compared with boys with primary leukemia. The difference was significant only for ALL (OR = 0.35, 95% CI = 0.15-0.85, p = 0.0163). The frequency of the GSTT1 null genotype in boys with relapse (not separated according to leukemia variant) was 2.7-fold lower than in boys with primary leukemia (OR = 0.37, 95% CI = 0.18-0.79, p = 0.0085). There was a significant increase in frequency of CYP1A1 genotype *1/*2A in combination with GSTT1 non null genotype for boys with ALL relapse when compared with boys with primary ALL (OR = 3.09, 95% CI = 1.17-8.15, p = 0.0254).

Distribution of the CYP1A1 Genotype *1/*2A and GSTs Genotypes in Girls with Acute Leukemia. Combinations of the CYP1A1 Genotype *1/*2A and GSTs Genotypes. Girls with ALL or AML relapse had a higher frequency of CYP1A1 genotype *1/*2A when compared with girls with primary leukemia (Table 9). The patients with ALL or AML relapse had a lower frequency of the GSTM1 null genotype when compared with girls with primary leukemia. The difference was significant only for ALL (OR = 0.37, 95% CI = 0.15-0.90, p = 0.0303). Frequency of the GSTM1 null genotype in girls with acute leukemia relapse (not separated according to leukemia variant) was 2.4-fold lower than in those with primary leukemia (OR = 0.41, 95% CI = 0.19-0.85, p = 0.0175). Girls with leukemia relapse had a higher frequency of the CYP1A1 genotype *1/*2A in combination with the GSTM1 non null genotype when compared with girls with primary leukemia (Table 9).

The NAT2 Genotype Distribution in Girls with Acute Lymphoblastic Leukemia. Girls with ALL relapse had a higher frequency of heterozygous and homozygous variants 341C/– and 481T/– of NAT2 and a lower frequency of heterozygous and homozygous variants 590A/- and 857A/- when compared with these with primary ALL (data not shown). Girls with ALL relapse showed a higher frequency of genotype 341C/–, 481T/–, 590G/G, 857G/G when compared with those with primary leukemia (OR = 1.71, 95% CI = 0.73-4.04, p = 0.267).

The NAT2 Genotype Distribution in Boys with Acute Myoloblastic Leukemia. Boys with AML relapse had a lower frequency of heterozygous and homozygous variants 341C/– and 481T/– of NAT2 and their combinations when compared with those with primary AML (Table 10). An opposite tendency was observed for the NAT2 polymorphic variant 590G>A, the frequency of the heterozygous and homozygous variant 590A/- in AML relapse being higher than in primary leukemia. Boys with AML relapse had a higher frequency of genotype 341T/T, 481C/C, 590A/- when compared with those with primary AML (OR = 2.40, 95% CI = 0.55-10.39, p = 0.297).

Combined Analysis of NAT2 and GSTs Genotypes in Girls with Acute Lymphoblastic Leukemia. The NAT2 genotype 341C/–, 481T/–, 590G/G, 857G/G in combination with the GSTT1 and/or GSTM1 non null genotype occurred in girls with ALL relapse at a higher frequency than in those with primary ALL (Table 11), the difference being significant for combinations of the NAT2 genotype with the GSTM1 non null genotype or the double GSTT1/ GSTM1 non null genotype.

Combined Analysis of NAT2 and GSTs Genotypes in Boys with Acute Myoloblastic Leukemia. Boys with AML relapse had a higher frequency of NAT2 genotype 341T/T, 481C/C, 590A/– in combination with the GSTT1 non null genotype when compared with those with primary leukemia (OR = 5.60, 95% CI = 0.97-32.21, p = 0.063). For CYP1A1 (4887Ñ>À and 4889A>G), CYP2D6 (1934 G>A and DelA2637), MTHFR (677C>T and 1298A>C), MTRR (66A>G), NQO1 (609C>T), CYP2C9 (430C>T and 1075A>C), and CYP2C19 (681G>A) no significant difference in genotype frequencies was observed between the patients with relapse and those with primary leukemia.