Approximately 50% of individuals in most human populations completely lack activity of the detoxifying enzyme glutathione S-transferase M1. The medical consequences of this deficiency have been extensively investigated in molecular epidemiological studies, but possible differences of the highly active homozygous genotype versus the moderately active heterozygous genotype could not be considered because many currently used polymerase chain reaction (PCR) assays cannot distinguish the homozygous genotypes GSTM1 *A/A and GSTM1*B/B from the heterozygous genotypes GSTM1*A/0 and GSTM1*B/0. Here, we describe that long PCR is suitable for this purpose by specifically producing a signal for the GSTM1*0 allele. Based on the published cluster of GSTM genes (GSTM1 to -M5), a 13-kb segment spanning the site of the GSTM1 deletion was amplified using a GSTM2-specific forward primer (5'-CATCGCTTATGATGTCCTTGAGAGAAACCAAG-3') and a reverse primer, which is specific for the upstream region of GSTMS (5'-GCGTTTCTGAGGACTGGACTGATGATCG-3'). Any failure in the amplification was controlled by co-amplification of a 15-kb human tissue plasminogen activator gene fragment. The GSTM1*0-specific 13-kb amplicon appears in all carriers of one or two null alleles, but in none of the 14 tested GST1*A/B samples which served as controls since individuals with this genotype are known to lack any GSTM1*0 allele. While conventional PCR assays for detection of the GSTM1 deletion differentiated homozygously deficient from hetero- or homozygously active individuals, this long PCR assay differentiates homozygously active from hetero- or homozygously deficient individuals. Using both assays, an unambiguous differentiation into carriers of zero, one or two active alleles of GSTM1 is possible.