RESUMO
BACKGROUND: Until now, the influence of sex, age, smoking, drinking on serum pepsinogen levels has been assessed by single regression analysis. However, the influence of those factors on pepsinogen levels should be assessed exactly by multiple regression analyses. SUBJECTS AND METHODS: 891 subjects were collected from by questionnaire and serum tests. Analyses were done with serum pepsinogen I (PG I), pepsinogen II (PG II) and pepsinogen I/II ratio (PG I/II) as a criterion variable and as categorized explanatory variables, sex, age, current or past smoking habit, and current drinking habit. And analyses are done by Mann-Whitney U test, correlation coefficient, single regression method, multiple regression method. RESULTS: PG I level is significantly higher in men than in women by Mann-Whitney U test. But the effect of sex factor is not remarkable by multiple regression analyses. PG II level increased and PG I/II level decreased with progression of age by all analyses methods. Current or past smoking elevates PG I level by Mann-Whitney U test, but current smoking dose and past smoking amount showed no dose-dependent associations with PG I level. Current drinking elevates PG I level by Mann-Whitney U test, but current drinking dose showed no dose-dependent associations with PG I. However, the effects of current smoking and current drinking to serum PG levels is not so large by multiple regression analyses. CONCLUSION: The effects of sex, current smoking and current drinking to serum PG levels are not remarkable by multiple regression analyses. Significantly, PG II level increased and PG I/II level decreased with progression of age. Therefore it may not be necessary to consider the effects of sex, smoking habit and drinking habit when serum pepsinogen levels are used as markers for gastric cancer.
