A time-dependent bidirectional association between folate and lung cancer deaths among a national cohort.

The relationship between folate and the risk of cancer remains undetermined partially due to the dynamic changes in folate intakes at the population level caused by folic acid fortification implemented in the USA and other countries. To control for the interference from fortification, we assessed the relationship between folate and lung cancer death (LCD) risk among a national cohort established years before folic acid fortification. We followed up 14 528 adults aged 19 years or older who participated in the National Health and Nutrition Examination Survey (1988-1994) on average for 14 years. LCD's hazard ratios were estimated by the folate levels using Cox regressions. After 192 973 person-years (py) of follow-up, 233 LCDs were recorded. The LCD rates were 1.20/1000 py, 1.14/1000 py, and 1.38/1000 py for adults with low (1 st quarter), moderate (2 nd and 3 rd quarter), and high (4 th quarter) serum folate. In the first 10 years of follow-up, the adjusted hazard ratio was 2.87 (1.30-6.37) for adults with moderate, and 1.56 (0.58-4.23) for adults with high serum folate, compared to adults with low serum folate. For adults who survived longer than 10 years of follow-up, the hazard ratios were 0.45 (0.24-0.86) and 0.37 (0.16-0.87) respectively. No association was observed between LCD risk and red blood cell folate level. With minimized interference from folic acid fortification, we detected a time-dependent bidirectional association that supports the dual effects of folate on the carcinogenesis of lung cancer.