Editor—The journal has previously published a letter entitled “Glycogen storage disease type II: frequency of three common mutant alleles and their associated clinical phenotypes studied in 121 patients”.1 The title of the letter has apparently led to the mistaken impression that all three of the mutations discussed (IVS1-13T→G, Δ525T, and Δexon 18) are common in the general patient population. The IVS1-13T→G mutation was found to be “common” in diverse white populations (Dutch and “non-Dutch”), confirming our previous report.2However, careful reading of the authors’ results and comments indicates that while the Δ525T and Δexon 18 mutations have a relatively high allele frequency in Dutch patients (>0.2), their frequency is much lower in patients from “non-Dutch” populations. Nonetheless, a subsequently published paper suggests that the exon 18 deletion should be preferentially screened for in non-Dutch patients,3 perpetuating the mistaken impression that Δexon 18 is very common in the general white population. We have been prompted to report our own unpublished data because of three manuscripts recently received by us for review, indicating the same misconceptions as to the frequency of the Δ525T and Δexon 18 mutations in diverse white populations.

We have screened white and non-white patients from the USA and Canada for the Δ525T and Δexon 18 mutations and would therefore like to confirm and emphasise that both mutations, although present in multiple, unrelated, apparently “non-Dutch” patients, are relatively rare compared to their frequency in “Dutch” patients (table 1). In addition, based on limited data, our results suggest that both mutations may have a higher allele frequency in French Canadians than in the general white population.

In order to compare data from patient populations with different proportions of adult and infantile onset patients, we have used “chromosomes at risk” rather than total number of chromosomes as the denominator for calculating mutant allele frequency. In infantile onset patients, both alleles are at risk for the Δ525T and Δexon 18 mutations since homozygosity for either of these mutations results in infantile onset disease. In adult/juvenile onset patients, only one allele is at risk for carrying these mutations.

For the Δ525T mutation, we studied 34 chromosomes at risk for the mutation in 25 white, non-Canadian patients (nine infantile onset and 16 adult/juvenile onset patients). The Δ525T mutation was found in three of the 34 chromosomes at risk (in heterozygosity in two infantile and one adult onset patient). Based on conversion to “chromosomes at risk”, it can be seen (table 1) that the allele frequency of the Δ525T in these patients was 0.09, much lower than the 0.32 frequency in Dutch populations. Additionally, the mutation was not found in 16 chromosomes at risk from non-white patients (primarily infantile onset Afro-Americans).

For the Δexon 18 mutation we studied 61 chromosomes at risk in 53 white, non-Canadian patients (eight infantile onset and 45 adult/juvenile onset patients). We found the Δexon 18 mutation in heterozygosity in three adult onset patients (table 1). The gene frequency for the Δexon 18 mutation in the population we studied was therefore 0.05, again much lower than the 0.23 allele frequency in the Dutch population.1

The allele frequencies for both mutations were somewhat higher than previously reported for “non-Dutch” populations (table 1). However, our analysis was restricted to white populations and inclusion of non-whites would lower the frequency. Thus, an additional 20 chromosomes at risk from 10 non-white infantile onset patients did not show any carrying the Δexon 18 or the Δ525T mutations.

The results were strikingly different in a small group of Canadian infantile onset patients studied. We examined DNA from seven patients, two patients referred directly and five cell lines obtained from the Canadian Mutant Repository. Three patients were eliminated from our data because of Dutch origin; two were represented by cell lines that were submitted from Holland (WG1099, heterozygous for Δexon 18 and WG1102), and the third was one of the referred patients, homozygous for Δexon 18, who was of Dutch descent. Of the three non-Dutch cell lines, two (WG285 and WG482), reportedly from French Canadians, were both heterozygous for Δexon 18.4 5 The remaining cell line (WG173) was of unknown origin and was homozygous for the Δ525T mutation. The other referred patient did not carry either of the mutations. Thus, in the non-Dutch Canadians, 4/8 chromosomes at risk carried either the Δexon 18 or Δ525T mutation. These results, albeit from a very small sample, suggest that both mutations may be more common in French Canadians as well as in subjects of known Dutch descent in Canada. Nonetheless, inclusion of these data in the overall calculations does not appreciably affect the allele frequencies for white populations.

It would appear likely that the relatively high frequency of these two mutations in the Dutch population is the result of founder effect. However, since both mutations occur in the context of sequence found at high frequency surrounding small and larger deletions,4 6 the occurrence of these mutations in other populations could either represent common descent or independent recurrence. In summary, although the bulk of the over 40 mutations identified in GSDII to date are each found in only a few patients, some mutations are common in defined ethnic groups. These include D645E in Taiwanese (0.80),7 R854Stop in Afro-Americans (0.5),8 and Δ525T and Δexon 18 in the Dutch (0.25-0.34)1 (and possibly in French Canadians). Only the IVS1-13T→G splice site mutation has a high frequency in adult/juvenile patients of diverse white populations (0.6 of chromosomes at risk),2 suggesting that it is an ancient mutation. It remains to be determined if the IVS1 mutation is common in Asian or African adult onset patients.