BACKGROUND The association between elevated γ-glutamyltransferase (GGT) at a certain point and incident cancer has been suggested; however, no study has evaluated the association between repeatedly elevated GGT and cancer incidence.

AIM To investigate the effects of repeatedly elevated GGT on the incidence of digestive cancers.

METHODS Participants who had undergone health screening from 2009 to 2012 and 4 consecutive previous examinations were enrolled. GGT points were calculated as the number of times participants met the criteria of quartile 4 of GGT in four serial measurements (0-4 points). Multivariable Cox proportional hazard regression models were applied.

RESULTS In total, 3559109 participants were included; among them, 43574 digestive cancers developed during a median of 6.8 years of follow-up. The incidence of total digestive cancers increased in a dose-response manner in men [adjusted hazard ratio (aHR) compared with those with 0 GGT points = 1.28 and 95% confidence interval (CI) = 1.24-1.33 in those with 1 point; aHR = 1.40 and 95%CI = 1.35-1.46 in those with 2 points; aHR = 1.52 and 95%CI = 1.46-1.58 in those with 3 points; aHR = 1.88 and 95%CI = 1.83-1.94 in those with 4 points; P for trend < 0.001]. This trend was more prominent in men than in women and those with healthy habits (no smoking, no alcohol consumption, and a low body mass index) than in those with unhealthy habits.

CONCLUSION Repeatedly elevated GGT levels were associated with an increased risk of incident digestive cancer in a dose-responsive manner, particularly in men and those with healthy habits. Repeated GGT measurements may be a good biomarker of incident digestive cancer and could help physicians identify high-risk populations.