Controlling for high-density lipoprotein cholesterol does not affect the magnitude of the relationship between alcohol and coronary heart disease

BackgroundThisstudytestedthehypothesisthatmoderatealcoholintakeexertsitscardioprotectiveeffectmainlythroughanincreaseintheserumlevelofhigh-densitylipoproteincholesterol.MethodsandResultsIntheCohortofNorway(CONOR)study,149729adultparticipants,recruitedfrom1994to2003,werefollowedbylinkagetotheCauseofDeathRegistryuntil2006.Atrecruitment,questionnairedataonalcoholintakewerecollected,andtheconcentrationofhigh-densitylipoproteincholesterolinserumwasmeasured.UsingCoxregression,wefoundthattheadjustedhazardratioformenfordyingfromcoronaryheartdiseasewas0.52(95%confidenceinterval,0.39-0.69)whenconsumingalcoholmorethanonceaweekcomparedwithneverorrarely.Theratiochangedonlyslightly,to0.55(0.41-0.73),aftertheregressionmodelincludedtheserumlevelofhigh-densitycholesterol.Forwomen,thecorrespondinghazardratioswere0.62(0.32-1.23)and0.68(0.34-1.34),respectively.ConclusionsAlcoholintakeisrelatedtoareducedriskofdeathfromcoronaryheartdiseaseinthefollow-upofalarge,population-basedNorwegiancohortstudywithextensivecontrolforconfoundingfactors.Ourfindingssuggestthattheserumlevelofhigh-densitycholesterolisnotanimportantintermediatevariableinthepossiblecausalpathwaybetweenmoderatealcoholintakeandcoronaryheartdisease.