A Hospital-basd Case-control Study On Lifestyle Risk Factors Of Acute Myeloid Leukemia And Its WHO Subtypes

A Hospital-basd Case-control Study on Lifestyle Risk Factors of Acute Myeloid Leukemia and Its WHO SubtypesObjectives To analyse the effect of three major lifestyle risk factors(BMI,smoking and alcohol drinking) on AML(acute myeloid leukemia) and its WHO subtypes,as well as the impact of other environmental exposure upon this effect,both of which would facilitate the conduction of prevention for specific population at risk and the exploration of the disease mechanisms of AML.Materials and methods A hospital-based,1:2 paired case-control study design is employed in this study.All AML patients,who were Shanghai residents,newly diagnosed during August 2003～June 2007 in the 29 participating hospitals were supposed to be selected as cases.By the meantime of the enrollment for each case,2 controls with the same gender,age difference no larger than 5 years,Shanghai residence and without malignant neoplasms or ever diagnosed with lympho-hematopoietic diseases were selected from the same hospistal.All participants were included after they had signed an informed consent designed according to Helsinki Declaration.A double-blinded face-to-faced interview with a 17-page carefully developed questionnaire was immediately conducted by trained interviewers.The questionnaire covers information on demographics,family history,disease history, medical history,occupational history,lifestyle exposure and environmental exposure. Chi-square test,Mantel-Haenszel stratified analysis and conditional as well as unconditional logsitic regression models were conducted using SAS for analysis.Results A total of 722 effecient cases(among which 244 AML-RCA cases,186 AML-MD cases,5 t-AML cases and 269 AML-noc cases) and 1444 controls were included.Participation ratio reached 97.4%.Adjusted with recognized confoundings, compared with the BMI desirable group,the ORs and 95%CIs of underweight, overweight and obese upon AML-total were 1.00(0.72-1.38),0.87(0.70-1.09) and 0.67(0.46-0.97) respectively,with a negative trend among the risks(P_trend=0.04).For males,ORs and 95%CIs of the 3 BMI levels upon AML-total were 1.09(0.67-1.79), 0.93(0.69-1.25) and 0.65(0.38-1.09) respectively,while for females 0.94(0.59-1.50), 1.04(0.72-1.49) and 0.80(0.72-1.49) respectively.Significant trends were detected in neither of the genders(P_trend=0.31 both).There is a positive trend among BMI and APL risk in males(P_trend==0.03),but a negative trend among BMI and AML-noc (P_trend=0.01).Females with BMI＜18.5 who ever did house renovation have a significant elevated AML-total risk(OR=8.41,95%CI=2.06-34.34).Active smoking can elevate AML-total risk significantly(OR=1.28,95%CI=1.00-1.63),but without dose-response effect(P_trend=0.56).For males,ORs and 95%CIs of light,moderate and heavy smoking upon AML-total were 1.53(1.07-2.18),1.26(0.87-1.82) and 0.87(0.59-1.28) respectively,while for females 0.63(0.25-1.62),0.56(0.07-5.08) and 1.20(0.19-7.40) respectively.Significant dose-responese trends were detected in neither genders(P_trend=0.98 and 0.82 respectively).With regard to subtypes,male moderate smokers' AML-ETO risk was significantly elevated(OR=3.52. 95%CI=1.17-10.6).Male light smokers' AML-noc risk was also significantly elevated(OR=2.40,95%CI=1.22-4.71).Male light smokers who ever lived in a farm has a significantly elevated AML-total risk(OR=2.21,95%CI=1.06-4.61),which indicated a possible strengthened effect modification.Compared with male nonsmokers,male light,moderate and heavy smokers who ever used Chinese traditional medications had non-significantly reduced AML-total risks(OR=0.79, 95%CI=0.10-6.54;OR=0.64,95%CI=0.08-5.28;OR=0.86,95%CI=0.12-6.07),while smokers never used Chinses traditional medications had elevated AML-total risks (OR=1.45,95%CI=1.03-2.05;OR=1.20,95%CI=0.84-1.71;OR=0.87, 95%CI=0.60-1.27).Ever alcohol drinkers had an elevated AML-total risk(OR=1.53, 95%CI=1.17-2.01) and there was a positive dose-dependent trend among drinking dose and AML-total risk(P_trend=0.01).For males,light and heavy drinkers' ORs and 95%CIs were 1.35(0.92-1.97) and 1.18(0.82-1.71) respectively,while for females 2.30(0.68-7.86) and 1.50(0.21-10.9).Significant dose-dependent trends could be detected in neither genders(P_trend=0.16 and 0.23 respectively).Male light drinkers who ever used Chinese traditional medications had a non-significantly reduced AML-total risk(OR=0.60,95%CI=0.10-3.51),while male light and heavy drinkers who never used Chinese traditional medications had elevated AML-total risks (OR=1.37,95%CI=0.94-2.07;OR=1.25,95%CI=0.87-1.80 respectively).ORs and 95%CIs of beer,ricewine and hard liquor upon AML-total were 1.38(0.95-2.00), 1.47(1.03-2.09) and 1.29(0.93-1.80) respectively.ORs and 95%CIs of the three wine sub-categories for male light drinkers were 1.23(0.69-2.19),0.95(0.52-1.73) and 1.47(0.88-2.46),while for male heavy drinkers were 1.12(0.69-1.84),1.35(0.86-2.14) and 0.85(0.54-1.34) respectively.Regarding subtypes,total wine drinking and beer consumption had positive dose-dependent trends with AML-noc risk(P_trend=0.02 and 0.05 respectively).Male drinkers of hard liquor had a significantly elevated AML-noc risk(OR=3.84,95%CI=1.29-11.5).Conclusion AML-total risk is reduced by increased BMI,but the subtype APL's risk is significantly elevated by increased BMI.Active smoking increases AML risk, which shows a certain WHO subtype-specificity.Alcohol drinking increased AML risk too,and shows a dose-dependent trend.But the effect may not be WHO subtpe specific.Some of the environmental exposure characteristics may have a strengthening modifying effect on the AML risk of the three lifestyles.We may recognize certain high risk sub-populations for specific preventions.The use of Chinses traditional medications shows a potential protective effect on AML risks of both tobacco smoking and alcohol drinking.