CYP2E1 Gene Polymorphism. The results of CYP2E1 genotyping are presented in Table 2. Note that we identified only the common allele, CYP2E1*1C, carrying six repeats in the promoter region, and the insertion allele CYP2E1*1D carrying eight repeats, and did not identify the wild type allele, CYP2E1*1A, which contains only five repeats. The CYP2E1*1D allele was rare, especially in the control group, and there was no difference in the frequency of this allele between our controls (2.5%) and a Caucasian population (2%) [12]. However, a considerably increased frequency of the CYP2E1*1D was observed in MND patients (14% versus 2.5% in controls; p <0.0001). Analysis of genotype distribu­tions did not identify the CYP2E1*1D/CYP2E1*1D genotype in the control group. These results are in accordance with those for a Caucasian population [12]. Genotype distributions showed statistically significant differences between our patients and controls (p = 0.0018). There was observed a considerably increased frequency of CYP2E1*1D homozy­gotes among our MND patients (see Table 2). The genotype distribution was consistent with Hardy-Weinberg equilibrium. Because genotypes that contain the insertion are associated with higher levels of enzyme activity, we considered CYP2E1*1D heterozygotes (a) and CYP2E1*1D homozy­gotes (b) as one genotype. The combined CYP2E1 genotype (a + b) was also essentially predominant in our patients when compared with the control group [RR 2,500, 95% confidence interval (CI) 1,159-5,392].

      We estimated the possible influence of the CYP2E1*1D allele, with increased enzyme activity, on the clinical expres­sion of MND. To this end, we studied the correlation between the combined CYP2E1 genotype and patients with different clinical features, as listed in Table 1. We found only one cor- relation between clinical diagnosis and the combined geno- type (gamma correlation: Z = 2.838, p = 0.005; Spearmen correlation: R = 0.245, p = 0.033). Further analysis showed that the common allele genotype (CYP2E1*1C/ CYP2E1*1C) is associated with cervical onset ALS, whereas the CYP2E1*1D allele is significantly correlated with other and more malignant forms of MND, such as truncal onset and diffuse onset ALS and progressive bulbar palsy [χ ² = 3.85, p = 0.0492; see Fig. 1(a)].

      CYP2D6 Gene Polymorphism. The CYP2D6 allele and genotype frequencies for our patients and controls are given in Table 3. No significant difference in genotype frequency was observed between our controls and British Caucasian controls [33], the frequency of CYP2D6*4 homozygotes being 4% in both. Neither patient nor control genotype fre­quencies varied significantly from the Hardy-Weinberg equi­librium.

      Although a previous study of ALS reported a consider­ably increased frequency of the CYP2D6*4 allele [10], a direct ^{χ 2} comparison of the CYP2D6 allele and genotype frequencies in our MND patients and the controls revealed no significant difference (Table 3). The number of CYP2D6*4 homozygotes was elevated in the patient group.

      Because only CYP2D6*4 homozygosity results in the PM phenotype [10], we combined the non CYP2D6*4 homozy­gous (a) and CYP2D6*4 heterozygous (b) genotypes as one genotype. Comparative distribution of the combined CYP2D6 genotype (a+b) and CYP2D6*4 homozygotes revealed no statistically significant difference between our patients and controls (Table 3). CYP2D6*4 homozygotes were more fre­quent among our patients, but this difference was not statisti­cally significant (p = 0.0754).

      We also studied the distribution of the combined CYP2D6 genotype in patients with different clinical features, as CYP2E1 analysis. There was no significant correlation between different clinical features and any genotype of the CYP2D6*4 polymorphism (data not shown).

      GSTT1 and GSTM1 Gene Polymorphisms. The geno­type frequencies of the GSTT1 and GSTM1 genes observed in our patients and control groups are presented in Table 4. Note that we could conclusively identify only the null genotypes GSTT1(0/0) and GSTM1(0/0). The heterozygotes and homo­zygotes for the normal allele were combined together and are treated as the GSTT1(+) or GTM1(+) genotypes.

      The frequencies of GSTT1 and GSTM1 homozygous deletions in our control group were similar to those reported for Caucasians [20,23,36]. The frequency of GSTT1-null subjects was much lower than that of GSTM1(0/0) genotype and was very similar in both groups (see Table 4). Our results for the GSTT1(0/0) genotype differ from those of Stroom­bergen et al. [23]. A preponderance of GSTT1-null subjects was observed in the patient group, whereas no association between GSTM1 homozygous deletions and MND were found in the British population [23]. In contrast, we found a significant difference in the GSTM1(0/0) genotype distribu­tion between our patients and controls. We also found signifi­cantly reduced GSTM1(0/0) frequency and elevated GSTM1(+) frequency in our patients group (Table 4). We found no significant correlation between the GSTT1-null and GSTM1-null genotypes and different clinical features.

      GSTP1 Gene Polymorphism. The results of GSTP1 genotyping are presented in Table 5. The analysis of 13 addi­tional patients did not affect our previously reported results [38]. The wild type allele (GSTP1*A) was the most common allele in the control and patient groups. The frequency of GSTP1*B homozygotes in the control population was similar to that reported for Caucasians [39,40]. The distributions of the GSTP1 genotypes and those predicted by allele frequen­cies showed that the control and patient groups were in Hardy-Weinberg equilibrium. A slight preponderance of GSTP*B homozygotes was observed among our patients but did not appear to be significant (see Table 5). Because previ­ous studies have demonstrated that the 105Val variant has altered heat stability and catalytic efficiency for various xeno­biotics [21], we combined heterozygotes (a) and homozygotes (b) for the GSTP1*B allele into one category. As shown in Table 5, the distributions of the GSTP1 genotype (a+b) and GSTP1*A homozygotes were very similar. We found no significant association between the combined GSTP1 geno­types and the different clinical features listed in Table 1 (data not shown).

      Only one correlation between variations in motor neuron involvement and the combined GSTP1 genotypes was found (gamma correlation: Z = 2.909, p  = 0.003; Spearmen correla­tion: R = 0.242, p  = 0.036). Further analysis showed that the wild type allele genotype (GSTP1*A/GSTP1*A) was associ­ated with classical upper and lower MN involvement, where­as the presence of the GSTP1*B allele was significantly cor­related with predominant lower MN involvement and pre­dominant upper MN involvement [^{χ 2} = 6.60; p = 0.0102, see Fig. 1(b)].

      NAT2 Gene Polymorphism. The three most common slow alleles, S1, S2, and S3 (NAT2*5, NAT2*6, and NAT2*7, respectively) and one wild type fast allele, F1 (NAT2*4), were identified. The number of S1, S2, and S3 alleles and the fre­quencies of various genotypes assessed separately were rather low (data not shown), so we combined all the alleles into slow (S) and fast (F1), and all the genotypes into rapid acetyl­ators and slow and intermediate acetylators. The frequency of slow and intermediate acetylators was slightly elevated in the MND patient group (57.3%) versus control group (50.5%), were not significant (^{χ 2} = 0.83, p = 0.36). No significant cor­relation was observed between acetylator types and clinical features.

Allele frequency	MND patients N (%)	Control N (%)	2	P	RR (0,95 %CI)
CYP2E1*1C (Ins96(-)) CYP2E1*1D (Ins96(+))	129 (86) 21 (14)	205 (97,5) 5 (2,5)	15,938a	<0,0001a	3,192 (1,445-7,050)
Genotype frequency
CYP2E1*1C/ CYP2E1*1C CYP2E1*1C/ CYP2E1*1D (a) CYP2E1*1D/ CYP2E1*1D (b)	60 (80,0) 11 (14,7) 4 (5,3)	100 (95) 5 (5) 0	11,571b	0,0031b
CYP2E1*1C/ CYP2E1*1C a+b	60 (80,0) 15 (20,0)	100 (95) 5 (5)	8,801a	0,003a	2,500 (1,159-5,392)
Total number	75	105
MND - M???r neuron disease, a -  2 is has been calculated with use  2x2 contingency table,  GraphPadInStat program,  marked correlation are significant at p<0,05 b - 2 is has been calculated with use of 2x3 contingency table,  GraphPadInStat program, marked correlation are significant at p<0,025 according Bonferroni’s correction

 a) 

b)