In the same work,
nonfermented soyfoods were thought to lower the risk to 0.72, based on the analysis
of the data of ten pooled studies. The authors, however, warned of possible confounders
such as salt and fruits-vegetable intake.66 Responding to the Wu et al.66 study, Ji
and colleagues reanalyzed 1,124 original data from the 1998 case-control study,22
which had showed a decreased cancer risk in men with increased soy intake, and a
small insignificant increased risk in women. The reanalysis showed that adjustment
for salt intake and salt preference lowered the odd ratio slightly, while nonfermented
soyfoods lowered the cancer risk in men, but not in women. It was also shown that
fermented bean curd was responsible for the increased risk among women.67 A prospective
study involving a Japanese immigrant population in Hawaii reported that
miso (a Japanese fermented soybean product) was mildly associated with gastric
cancer, while salt consumption showed a stronger association.