Leisure Time Physical Activity and Sleep Predict Mortality in Men Irrespective of Background in Competitive Sports

This is a sel -a chi ed e sion o an o iginal a icle. This e sion may di e om he o iginal in pagina ion and ypog aphic de ails. Au ho (s): Ti le: Yea : Ve sion: Copy igh : Righ s: Righ s u l: Please ci e he o iginal e sion: CC BY-NC-ND 4.0 h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/ Leisu e Time Physical Ac i i y and Sleep P edic Mo ali y in Men I espec i e o Backg ound in Compe i i e Spo s © 2017 The Au ho (s). Published e sion Wennman, Heini; K onholm, E kki; Heinonen, Olli; Kujala, U ho; Kap io, Jaakko; Pa onen, Timo; Bäckmand, Heli; Sa na, Seppo; Bo odulin, Ka ja Wennman, H., K onholm, E., Heinonen, O., Kujala, U., Kap io, J., Pa onen, T., Bäckmand, H., Sa na, S., & Bo odulin, K. (2017). Leisu e Time Physical Ac i i y and Sleep P edic Mo ali y in Men I espec i e o Backg ound in Compe i i e Spo s. P og ess in P e en a i e Medicine, 2(6), A icle e0009. h ps://doi.o g/10.1097/pp9.0000000000000009 2017 Downloaded om h ps://jou nals.lww.com/p ogp e med by BhDM 5ePHKa 1zEoum1 Q N4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQ HD3bhnalqTQiPuWMPspR1 kCD17P2jh4nP Pi9Z9XCZsWXwx qE+VMbaw== on 06/13/2019 Downloaded om h ps://jou nals.lww.com/p ogp e med by BhDM 5ePHKa 1zEoum1 Q N4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQ HD3bhnalqTQiPuWMPspR1 kCD17P2jh4nP Pi9Z9XCZsWXwx qE+VMbaw== on 06/13/2019 1 PROGRESS IN PREVENTIVE MEDICINE OPEN ACCESS O iginal Resea ch Leisu e Time Physical Ac i i y and Sleep P edic Mo ali y in Men I espec i e o Backg ound in Compe i i e Spo s Heini Wennman, MSca; E kki K onholm, PhDa,b; Olli J. Heinonen, MDc; U ho M. Kujala, MDd; Jaakko Kap io, MDa,e,g; Timo Pa onen, MDa; Heli Bäckmand, PhD ; Seppo Sa na, PhDe; Ka ja Bo odulin, PhDa aDepa men o Public Heal h Solu ions, Na ional Ins i u e o Heal h and Wel a e, Helsinki, Finland; bUni o Resea ch, De elopmen , T aining and Se ice o Occupa ional Heal h, Finnish Ins i u e o Occupa ional Heal h, Helsinki, Finland; cDepa men o Heal h and Physical Ac i i y and Paa o Nu mi Cen e, Uni e si y o Tu ku, Tu ku, Finland; dFacul y o Spo and Heal h Sciences, Uni e si y o Jy äskylä, Uni e si y o Jy äskylä, Jy äskylä, Finland; eDepa men o Public Heal h, Uni e si y o Helsinki, Uni e si y o Helsinki, Helsinki, Finland; The Hospi al Dis ic o Helsinki and Uusimaa, Join Au ho i y Adminis a ion, HUS (Ci y o Helsinki), Finland; gIns i u e o Molecula Medicine FIMM, Uni e si y o Helsinki, Helsinki, Finland. Add ess ep in eques s o Heini Wennman, MSc, Na ional Ins i u e o Heal h and Wel a e, P.O. Box 30, FI-00271 Helsinki, Finland. E-mail add ess: heini.wennman@ hl. i (Wennman) ABSTRACT In oduc ion: Physical ac i i y and sleep a e closely ela ed beha io s wi h sugges ed syne gis ic in luence on ca dio ascula heal h. Physical ac i i y po en ially modi ies associa ions be ween sleep and mo ali y. Ou aim was o s udy he in e ela ionships be ween sleep, leisu e- ime physical ac i i y (LTPA), a his o y o spo s, and mo ali y. Me hods: A p ospec i e coho o o me eli e male a hle es (n = 1,028), and age- and egion-ma ched nona hle e men (n = 610) com- ple ed a heal h ques ionnai e in 1985. Thei mo ali y was ollowed up un il Decembe 31, 2011. Analyses included Cox p opo ional haza ds models wi h sleep du a ion and sleep quali y as main p edic o s o all-cause and ca dio ascula disease (CVD) mo ali y. Resul s: Sleep du a ion o sleep quali y we e no independen ly associa ed wi h mo ali y a e con olling o socioeconomic and li e- s yle ac o s. The in e ac ion be ween sleep du a ion and LTPA was signi ican , wi h highe isk o all-cause and CVD mo ali y i sleeping 6 hou s o less and no achie ing 450 me abolic equi alen minu es LTPA weekly, as compa ed wi h sleeping 6.5–8.5 hou s and achie ing 450 me abolic equi alen minu es o LTPA. Also, he ela i e excess isk due o in e ac ion be ween sho sleep and low LTPA was signi i- can o CVD mo ali y. Discussion: Signi ican in e ac ions be ween sleep du a ion and LTPA wi h ega d o mo ali y we e obse ed. In pa icula , sho sleep and low LTPA join ly p edic ed all-cause and CVD mo ali y i espec i e o a his o y o spo s. Findings sugges impo an syne gis ic associa ions o sho sleep and low LTPA wi h CVD mo ali y isk. Keywo ds: Sleep, Physical ac i i y, Mo ali y, Fo me a hle e, Ca dio ascula disease P elimina y esul s o he s udy ha e been p esen ed a he Ame ican Hea Associa ion Scien i ic Sessions, No embe 2015, O lando, Fla. PROGREVMED 2017; 2: e0009 Published online 1 No embe 2017 DOI:10.1097/pp9.0000000000000009 Copy igh © 2017 The Au ho (s). Published by Wol e s Kluwe on behal o he Eu opean Socie y o P e en i e Medicine. This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion-Non Comme cial-No De i a i es License 4.0 (CCBY-NC-ND), whe e i is pe missible o download and sha e he wo k p o ided i is p ope ly ci ed. The wo k canno be changed in any way o used comme cially wi hou pe mission om he jou nal. PROGPREVMED In oduc ion Bo h sho and long sleep du a ion a e sugges ed o associa e wi h highe all-cause mo ali y while he ela ionships wi h cause-spe- ci ic mo ali y ha e no conclusi ely been ag eed on.[1–3] The ole o con ounding beha io s and heal h pa ame e s in he sleep mo - ali y ela ionship need mo e e alua ion.[4] Sho sleep associa es wi h impai ed glucose me abolism and obesi y, low-g ade in lam- ma ion, and hype ension,[5,6] all possible mechanisms o highe mo ali y.[1,2,4,5] Fewe biological explana ions o he inc eased mo ali y isk o long sleepe s ha e been p esen ed.[1,7,8] Unde ly- ing disease and poo heal h o dis up ions in he ci cadian hy hm may pa icula ly be behind he isk o excessi e mo ali y o long sleep.[9,10] Whe e he mo ali y isk in sho and long sleepe s may be ela ed o and modi ied by di e en ac o s, physical ac i i y and mo e p ecisely he lack he eo , is one sugges ed con ounde in bo h long and sho sleepe s.[7,10,11] Physical ac i i y can be si u- a ed on ei he side o sleep on a causal pa hway o heal h and mo - ali y[4,12,13] and physical ac i i y and sleep a e sugges ed o ha e a syne gis ic e ec upon ca dio-me abolic isk ac o s.[14] Despi e impo an in e ela ionships be ween physical ac i i- y and sleep, he e a e only ew s udies ha ha e ocused on he in e ac ion be ween sleep and physical ac i i y in ela ion o mo - ali y. One such s udy was pe o med among U.S. men and wom- en (aged 50–72 yea s), and bo h he mul iplica i e and addi i e in e ac ion be ween sleep du a ion and mode a e- o- igo ous in ensi y physical ac i i y wi h mo ali y was examined. Howe e , in his s udy, physical ac i i y was no ound o in e ac wi h sleep in ela ion o mo ali y.[15] In ano he s udy consis ing o Swedish adul men and women aged 45–83 yea s, long sleep (> 8 hou s) p edic ed highe all-cause mo ali y only in hose wi h low physi- cal ac i i y.[11] Sho sleep (< 6 hou s) was ound o p edic mo ali- y a all le els o physical ac i i y.[11] 2017 2 Wennman e al. www.p ogp e med.com While he modi ying e ec o physical ac i i y on he sleep-mo - ali y associa ion needs cla i ying, physical ac i i y as such educes he isk o mo ali y.[16–18] Physical inac i i y, ha is, no mee ing he cu en guidelines o physical ac i i y accoun s o abou 9% o p ema u e dea hs wo ldwide.[17] Regula physical ac i i y imp o es ca dio espi a o y i ness le el and also a good ca dio espi a o y i ness educes he isk o mo ali y.[19,20] Mos compe i i e spo s equi e aining o a high ca dio espi a o y i ness le el, and lowe mo ali y in o me op-le el a hle es is pa ly due he good ca - dio espi a o y i ness hey ha e achie ed h ough a his o y o ig- o ous physical ac i i y.[21,22] Howe e , a aining su icien le els o physical ac i i y h oughou li e plays an impo an ole o heal h and mo ali y isk wi h inc easing age.[21,23,24] Remaining physical- ly ac i e o aking up physical ac i i y a olde ages (> 65 yea s) imp o es he odds o heal hy ageing and su i al.[25] Sleep and physical ac i i y a e closely ela ed beha io s bo h om ene gy expendi u e as well as a ime use poin o iew.[26,27] The ela ionship be ween physical ac i i y and sleep is likely bidi ec ional whe e bo h acu e and egula physical ac i i y can imp o e sleep,[12,28] and sleep du a ion and quali y may likewise in luence physical ac i i y beha io .[29] Physical ac i i y is hough o a ec sleep h ough mechanisms including luc ua ions in body empe a u e and glucose me abolism, au onomic ne ous sys em ac i i y, and changes in body composi ion, mood and ca dio espi- a o y i ness.[28–30] Howe e , among highly ac i e and well- ained a hle es, poo o insu icien sleep is su p isingly common and sleep pa e ns in his popula ion a e a he a ying.[31,32] A hle es seem o spend a he long ime in bed, bu hei ac ual sleep ime o en no mo e han 7 hou s, wi h bo h sleep du a ion and sleep quali y a ying in ela ion o spo s discipline, aining, and com- pe i ion schedules.[33,34] Among ac i e Mas e s a hle es Shepha d e al.[35] obse ed be e sleep quali y among he a hle es han he gene al popula ion. Ou aim was o s udy he in e ela ionships be ween a his o y o spo s, sleep, leisu e ime physical ac i i y (LTPA), and all-cause and ca dio ascula disease (CVD) mo ali y, in a popula ion coho en iched wi h indi iduals wi h a high physical ac i i y le el. Ou coho comp ises male o me eli e le el a hle es and age- and es- iden ial a ea–ma ched nona hle ic males. Conce ning his coho , we ha e lea ned ha ha ing a backg ound in compe i i e spo s inc eased he likelihood o a aining a heal hy li e s yle including physical ac i i y.[22,36,37] P e ious epo s in his coho also sugges lowe mo ali y and mo bidi y in o me a hle es han nona h- le es,[22,36,38] bu sleep o he associa ions be ween sleep and mo - ali y ha e so a no been s udied. Me hods The Finnish o me eli e a hle es’ coho A de ailed desc ip ion o s udy subjec s has been published p e- iously.[22] In b ie , he p ospec i e coho is based on o me eli e male a hle es (N = 2,448), ha is, hose who ep esen ed Finland a leas once in in e na ional o in e -coun y spo s compe i ions be ween he yea s 1920 and 1965. A hle es ep esen ed endu - ance spo s such as c oss-coun y skiing o long-dis ance unning, powe spo s such as boxing, w es ling, weigh li ing, and mixed spo s including, o example, ack and ield jumping, sho dis- ance unning, socce , ice hockey, and shoo ing.[22] The nona hle es (N = 1712) we e selec ed among Finnish men classi ied as com- ple ely heal hy (mili a y class A1, ully i o o dina y mili a y se - ice, which is obliga o y in Finland) a 20 yea s o age a he med- ical examina ion p eceding hei consc ip ion.[22] The nona hle es we e iden i ied om public a chi es o he egis e o men liable o mili a y se ice and ma ched on bi h coho and a ea o esi- dence wi h he a hle es.[22] In his s udy, a his o y o spo s e e s o he o me a hle e and nona hle e di ision. In 1985, a heal h ques ionnai e was simul aneously mailed o all a hle es and nona hle es ali e a ha ime poin (n = 2,528, 60.8% o all he iden i ied a hle es and nona hle es). In 1985, he esponse a es a ied be ween 77% and 90% o nona hle es and a hle es by spo , espec i ely. The ques ionnai e assessed demo- g aphic backg ound, an h opome ics, symp oms and diseases, heal h- ela ed ac o s, and heal h beha io s such as sleep and physical ac i i y. E hical app o al o his s udy (N° 173/13/03/2008) was p o id- ed by he HUCH E hics Commi ee, Biomedicum Helsinki 2 C P.O. Box 705, FI-00029 Helsinki, Finland on 20 h May, 2008. In o med consen was ob ained om all indi idual pa icipan s included in he s udy. Main exposu e a iables Sleep du a ion was assessed on he heal h ques ionnai e by ask- ing “How many hou s do you usually sleep pe 24 hou s?” wi h 9 esponse al e na i es (≤ 6 hou s, 6.5, 7, 7.5, 8, 8.5, 9, 9.5, and ≥10 hou s). Sleep du a ion was ca ego ized in o ≤ 6 hou s/d (sho sleep), 6.5–8.5 hou s/d (mid- ange sleep), and ≥ 9 hou s/d (long sleep). This ca ego iza ion was based upon p e ious li e a u e[10] and a uni a ia e Cox model whe e hose sleep du a ion ca ego ies ha we e ela ed wi h an inc eased isk o mo ali y as compa ed wi h 7 hou s sleep we e ecognized. Sleep quali y was epo ed on a ques ion abou usual sleep quali y wi h esponses “Well,” “Fai - ly well,” “Fai ly poo ly,” “Poo ly,” and “Canno say,” u he dicho o- mized in o “Good sleep” (including “ ai ly well” and “ ai ly poo ly”) and “Poo sleep.” Those epo ing “canno say” we e conside ed missing da a. The ques ions abou sleep du a ion and quali y ha e p e iously been used in he 1981 su ey o he Finnish Twin coho .[39–41] Th ee s uc u ed ques ions abou LTPA we e used o calcula e a physical ac i i y index, ep esen ing cumula i e leisu e me a- bolic equi alen (MET) hou s pe week.[37] A mul iple o he es - ing me abolic a e (1 MET e e ing o es ing me abolic a e) was assigned o he ype o LTPA epo ed and hen was he p oduc o in ensi y, du a ion, and equency calcula ed o a o al weekly LTPA. The me hod has been alida ed agains a de ailed physical ac i i y in e iew.[42] Following cu en physical ac i i y guide- lines,[18] subjec s we e g ouped acco ding o hei weekly amoun o LTPA in o hose wi h su icien (i.e., achie ing 450 MET minu es pe week) and insu icien LTPA (no -achie ing 450 MET minu es pe week). Co a ia es Occupa ional da a we e collec ed pa ly om he Cen al Popula- ion Regis y o Finland and pa ly om he 1985 ques ionnai e. Occupa ional g oups we e classi ied in o he ollowing main ca e- go ies: execu i es, whi e colla , blue colla , unskilled wo ke s, and a me s. Each pe son was classi ied acco ding o he occupa ion he had p ac iced o he longes pe iod.[22] P e ious s udies in his coho ha e epo ed on signi ican di e ences be ween a hle es and nona hle es in occupa ional s a us and signi ican associa- ions o occupa ional s a us, as a ma ke o socioeconomic s a us, wi h mo ali y.[22] Ma i al s a us was assessed on he ques ionnai e, and answe s we e g ouped in o he ca ego ies ma ied (including ma - ied, cohabi ing, ema ied) and no ma ied (including single, di o ced, and widowed), as p e iously in his coho .[43] In one 3 Wennman e al. www.p ogp e med.com ques ion, ypical wo king hou s o he pa icipan we e assessed, and he in o ma ion was used o g oup subjec s as day-wo ke s and nigh - o shi -wo ke s (including all nigh -wo k, 2 shi - and 3 shi -wo k). Subjec s’ smoking s a us was de e mined based on a de ailed smoking his o y[44] and u he classi ied in o: ne e , occasion- al, o me , and cu en smoke s. Alcohol consump ion based on quan i y- equency measu es o be e age use was con e ed in o g ams o pu e alcohol pe mon h as p e iously epo ed.[45] The esponden s we e ca ego ized as abs aine s, ligh - o-mode a e use s (no mo e han 14 s anda d d inks a week, wi h 1 s anda d d ink con aining 12 g o alcohol) and hea y use s (on a e age mo e han 2 d inks a day).[46] Sel - epo ed da a on heigh (m) and weigh (kg) we e used o calcula e he body mass index in kg/m², u he ca ego ized in o no mal (< 25 kg/m²), o e weigh (25–29.9 kg/m²), and obese (≥ 30 kg/m²). Li e sa is ac ion was measu ed by a sco e including he i ems “in e es ingness o li e,” “li e happiness,” “li e easiness,” and “loneli- ness,” wi h a high o al sco e indica ing high dissa is ac ion.[47] Li e sa is ac ion has p e iously been shown o co ela e wi h dep es- sion in his coho .[48] Diu nal ype o ch ono ype was assessed by one sel - a ed ques ion, used in se e al Finnish da a se s[49] o igina ing in he Mo ningness-E eningness ques ionnai e by Ho ne and Os - be g[50] as ollows: “One hea s abou mo ning and e ening ype o people. Which one o hese do you conside you sel o be?” Response ca ego ies we e “De ini ely mo ning,” “Mo e mo n- ing- han-e ening,” “Mo e e ening- han-mo ning,” and “De ini ely e ening.” Sleep medica ion use was assessed as days o use wi hin he pas 12 mon hs. Those who epo ed no use o ha ing used du ing less han 10 days we e ca ego ized as “No sleep medica- ion,” whe eas mo e equen use was ca ego ized as “Sleep med- ica ion use.” Mos (14.5% poin s) o he 18.6% missing da a on sleep medica ion use we e eclassi ied as nonuse s as hey epo - ed o sleep well and o become ale in less han 20 minu es in he mo ning. The es wi h missing in o ma ion we e excluded as la e explained. Sel - epo ed and physician-diagnosed ch onic diseases in 1985, including co ona y hea disease (co ona y hea disease, angina pec o is), pulmona y disease (ch onic b onchi is, emphy- sema, as hma), diabe es, a h i is ( heuma oid a h i is, os eoa - h i is), Pa kinson’s disease, and cance we e dicho omized in o a heal h s a us a iable as ha ing a his o y o disease (wi h some o he men ioned diseases) o no . Da a on cance diagnoses (yes/no) we e ob ained om he Finnish Cance Regis y and om he hos- pi al discha ge epo s as In e na ional Classi ica ion o Diseases codes (ICD-8 om 1970 o 1985) o neoplasms o all o gans (ICD 140–239). Discha ge epo s a e held in a na ionwide egis e by he na ional boa d o heal h and include bo h p i a e hospi als and co e ed eco ds (discha ge da es) om he beginning o 1970 o he end o 1985. Mo ali y ollow-up da a All esiden s o Finland ha e a unique pe sonal iden i ica ion num- be , which was used o eco d linkages. Pe sonal iden i ie s and possible da es o emig a ion o dea h o he coho we e ob ained om he Popula ion Regis e o Finland. In o ma ion on dea h was ob ained using he na ional cause-o dea h egis e a S a is ics Finland un il Decembe 31, 2011. Causes o dea h we e classi- ied acco ding o he In e na ional Classi ica ion o Diseases ( he eigh h e sion in 1969–1986; he nin h e sion in 1987–1995; he 10 h e sion in 1996 and onwa ds). Inclusion c i e ia In he cu en s udy, c i e ia o inclusion we e as ollows: (1) hose ali e a he s a o he ollow-up on Janua y 1, 1985 and pa icipa - ing in he su ey (n = 2,141), (2) a ailabili y o da a on sleep du a- ion, sleep quali y, ch ono ype, LTPA, occupa ional s a us, ma i al s a us, shi wo k, BMI, smoking, alcohol consump ion, sleep med- ica ion, li e sa is ac ion, his o y o ch onic diseases, and mo ali y un il 31s Decembe 2011 (n = 1,782), (3) no deceased due o can- ce in 1985 o 1986 (n = 1,726). Fu he mo e we e hose esponding “canno say” o he sleep quali y ques ion (n = 12) and he shoo - e s (n = 76) among he o me a hle es excluded. Shoo e s ha e a di e en age dis ibu ion and shoo ing as a spo has a di e en na u e compa ed wi h he o he spo disciplines included.[22] S a is ical analysis The desc ip i e da a a e p esen ed as equencies and pe cen ages by his o y o spo s (ie, o me a hle es and nona hle es; Table1). Cox p opo ional haza ds model wi h age as he ime a iable and age a baseline in 1985 as he en y was used o es ima e c ude and adjus ed haza d a ios (HRs) and 95% con idence in e als (CIs) o o al (all causes) and ca dio ascula (CVD) mo ali y. Sleep du a ion and sleep quali y we e ega ded as he main explana o- y a iables. Full adjus men s o he models included his o y o spo s (in ull sample), occupa ional s a us, ma i al s a us, shi wo k, smoking, alcohol use, BMI, ch ono ype, sleep medica ion, and his o y o ch onic disease. In models o CVD mo ali y, dea h o any o he eason was ea ed as censo ing. The p opo ional haza ds assump ion o each exposu e was es ed o by including an in e ac ion wi h ime in c ude, uni a - ia e models and isually by Kaplan–Meie plo s. All included a i- ables, excep occupa ional s a us sa is ied he assump ion. Fo his eason, ex ended Cox p opo ional haza ds models including an in e ac ion be ween occupa ional s a us and ime we e calcula - ed. Tes ing o he di e ence in -2 LogLikelihood be ween he no - mal and he ex ended Cox eg ession models suppo ed ha he ex ended model pe o ms be e o all-cause mo ali y bu no o CVD mo ali y. The e o e, he esul s o all-cause mo ali y a e based on he ex ended Cox eg ession model, whe eas esul s o CVD mo ali y a e based on no mal Cox eg ession models. Fi s , we es ed c ude as well as adjus ed Cox p opo ional haz- a ds models wi h sleep du a ion and sleep quali y o all-cause and CVD mo ali y s a i ied by he his o y o spo s. Then, we es ed he associa ion be ween sleep du a ion and sleep quali y wi h all- cause and CVD mo ali y in he ull sample, and his o y o spo s was added among he co a ia es. In he ull sample, we es ed in e ac ions be ween a his o y o spo s, a his o y o ch onic dis- ease, and LTPA wi h sleep du a ion and sleep quali y, espec i ely, ega ding bo h all-cause and CVD mo ali y. No s a is ically signi - ican in e ac ion was ound be ween a his o y o spo s and sleep du a ion and sleep quali y, o be ween a his o y o ch onic dis- ease and sleep du a ion and sleep quali y, espec i ely. The e was a s a is ically signi ican in e ac ion be ween sleep du a ion and LTPA, and sleep quali y and LTPA ega ding all-cause mo ali y (P = 0.03 and P = 0.03, espec i ely). Fo CVD mo ali y, he in e ac ion be ween sleep du a ion and LTPA was signi ican (P = 0.01), bu he in e ac ion be ween sleep quali y and LTPA was no (P = 0.06). Acco ding o he aims o he s udy and he signi ican in e - ac ion be ween LTPA and sleep, subjec s we e g ouped by hei combina ion o sleep du a ion (3 ca ego ies) and LTPA le el (2 ca - ego ies) in he ollowing 6 g oups: sho sleep + insu icien LTPA; sho sleep + su icien LTPA; long sleep + insu icien LTPA; long sleep + su icien LTPA; mid- ange sleep + insu icien LTPA; and 4 Wennman e al. www.p ogp e med.com mid- ange sleep + su icien LTPA. Also, 4 g oups by combina ion o sleep quali y and LTPA le el was o med as ollows: poo sleep + insu icien LTPA; poo sleep + su icien LTPA; good sleep + insu i- cien LTPA; and good sleep + su icien LTPA. All-cause and CVD mo ali y isk in he combina ions o sleep and LTPA we e es ed, adjus ing as p e iously o a his o y o spo s, occupa ional s a us, ma i al s a us, shi wo k, smoking, alcohol use, BMI, and ch ono ype, sleep medica ion, and his o y o ch onic disease. The g oup o mid- ange sleep and su icien LTPA, o good sleep and su icien LTPA, espec i ely, was consid- e ed as he e e ence ca ego y. Finally, addi i e in e ac ion was es ima ed by he ela i e excess isk due o in e ac ion (RERI) wi h co esponding 95% CIs using an excel sp eadshee c ea ed by Knol and Vande Weele[51] a ailable as a supplemen o hei a icle. Fo he pu pose o he addi i e in e ac ion analyses we e sho sleep- e s wi h insu icien LTPA, sho sleepe s wi h su icien LTPA, and mid- ange and long sleepe s wi h insu icien LTPA compa ed wi h mid- ange and long sleepe s wi h su icien LTPA, adjus ing o all co a ia es as in p e ious models. S a is ical analyses we e pe - o med using SAS e sion 9.3. (SAS Ins i u e Inc., Ca y, N.C.), wi h signi icance a P < 0.05. Resul s Ou analyses included 1,638 (77.5%) o he ini ial 2141 o me a h- le es and nona hle es who pa icipa ed in he baseline su ey in 1985. Compa ed wi h he included, he excluded we e olde (mean age, 62 e sus 55 yea s; P < 0.0001) and mo e likely had a his o y o ch onic disease [odds a io (OR), 2.11; P < 0.0001], slep ei he ≤ 6 hou s (OR, 1.5; P = 0.02) o ≥ 9 hou s (OR, 1.38; P = 0.03) han 6.5–8.5 hou s, epo ed poo sleep (OR, 2.65; P < 0.0001), and did no achie e he LTPA ecommenda ions (OR, 1.58; P < 0.0001). Table1 p esen s he dis ibu ion in backg ound cha ac e is ics by his o y o spo s. A hle es we e on a e age olde han he nona h- le es wi h a mean age o 56 yea s, (SD, ±10 yea s) e sus nona hle es mean 53 yea s (SD, ±9 yea s). O he o me a hle es included in his s udy, 13% (n = 132) we e ca ego ized as endu ance a hle es, 34% (n = 348) as powe a hle es, and 53% (n = 548) as mixed spo s a h- le es. Mixed spo s a hle es had he lowes p e alence o long sleep wi h 9% as compa ed wi h 17% in endu ance a hle es and 14% in powe spo a hle es. The p e alence o sho sleep was 7% o all di e en spo s. Rega ding sleep opinion, 7% o he endu ance a h- le es epo ed poo sleep as compa ed wi h 9% in mixed and powe a hle es, espec i ely. Se en y pe cen age o he powe a hle es and 77% o endu ance and mixed spo s a hle es had su icien LTPA. Resul s o mo ali y in ela ion o sleep du a ion, sleep quali y, and LTPA in he ull sample a e p esen ed in Table2. The e was a c ude associa ion o sho sleep du a ion (HR, 1.33), poo sleep quali y (HR, 1.29), and insu icien LTPA (HR, 1.31), wi h all-cause mo ali y. Rega ding CVD mo ali y, only insu icien LTPA showed a s a is ically signi ican c ude associa ion (HR, 1.40). Adjus ing models o his o y o spo s and demog aphic and heal h- ela ed co a ia es, all independen associa ions o sleep and LTPA we e a enua ed and no longe signi ican (Table2). Table3 p esen s he esul s o he combina ions o sleep du a- ion and sleep quali y wi h LTPA ega ding all-cause mo ali y, espec i ely. A combina ion o sho sleep du a ion and insu - icien LTPA, as compa ed wi h mid- ange sleep and su icien LTPA, was associa ed wi h highe all-cause mo ali y, also a e ull adjus men o co a ia es. A combina ion o poo sleep quali y and insu icien LTPA compa ed wi h good sleep and su icien LTPA was also signi ican ly associa ed wi h all-cause mo ali y in c ude model, bu no longe in he ully adjus ed model. The anal- ysis o RERI showed a posi i e, bu nonsigni ican addi i e in e ac- ion be ween sho sleep and insu icien LTPA o all-cause mo - ali y (RERIsho sleep + insu icien LTPA = 0.36; 95% CI, ˗0.23 o 0.96) and a posi i e, bu nonsigni ican in e ac ion be ween sleep quali y and LTPA o all-cause mo ali y (RERIpoo sleep + insu icien LTPA = 0.27; 95% CI, ˗0.20 o 0.74). Table4 p esen s he esul s o he combina ions o sleep du a- ion and sleep quali y wi h LTPA ega ding CVD mo ali y, espec- i ely. Sho sleep and insu icien LTPA signi ican ly associa ed wi h CVD mo ali y h oughou he models. Also mid- ange sleep TABLE 1. Cha ac e is ics o he S udy Sample (N = 1638) by His o y o Spo s A hle es, 1,028 (63%) Nona hle es, 610 (37%) P alue o Chi-squa e es Occupa ional s a us Execu i es 272 (26) 71 (12) <0.0001 Cle ical wo ke s 419 (41) 164 (27) Skilled wo ke s 283 (28) 273 (45) Unskilled wo ke s 16 (2) 23 (4) Fa me s 38 (4) 79 (13) Ma i al s a us Ma ied o co-habi ing 883 (86) 508 (83) 0.15 Li ing alone 145 (14) 102 (17) Type o wo k Day-wo k 856 (83) 524 (86) 0.16 Nigh - o shi wo k 172 (17) 86 (14) Leisu e ime physical ac i i y Less han 450 MET minu es/wk 262 (25) 326 (53) <0.0001 A leas 450 MET minu es/wk 766 (75) 284 (47) Sleep du a ion sho sleep 73 (7) 63 (10) 0.07 Mid- ange sleep 835 (81) 480 (79) Long sleep 120 (12) 67 (11) Sleep quali y Good sleep 940 (91) 525 (86) 0.001 Poo sleep 88 (9) 85 (14) Ch ono ype Mo ning ype 398 (39) 227 (37) 0.01 Mo e mo ning ype 271 (26) 202 (33) Mo e e ening ype 276 (27) 129 (21) E ening ype 83 (8) 52 (9) Body mass index No mal < 25 419 (41) 216 (35) 0.10 O e weigh 25–29.9 490 (48) 315 (52) Obese ≥ 30 119 (12) 79 (13) Alcohol consump ion Abs aine s 75 (7) 69 (11) 0.02 Ligh o mode a e use s 793 (77) 458 (75) Hea y use s 160 (16) 83 (14) Smoking s a us Ne e 515 (50) 164 (27) <0.0001 Occasional 49 (5) 16 (3) Fo me 297 (29) 250 (41) Cu en 167 (16) 180 (30) Use o sleep medica ions No 967 (94) 581 (95) 0.31 Yes 61 (6) 29 (5) His o y o ch onic disease No ch onic disease 670 (65) 393 (64) 0.76 Ch onic disease 358 (35) 217 (36) Li e sa is ac ion (median, ange) 7 (4–20) 8 (4–20) 5 Wennman e al. www.p ogp e med.com and insu icien LTPA was signi ican ly associa ed wi h CVD mo - ali y un il ully adjus ed model. The combina ion o poo sleep and insu icien LTPA, as well as good sleep and insu icien LTPA showed a signi ican associa ion wi h CVD mo ali y in he i s s eps o modeling, bu no in he ully adjus ed model. Acco d- ing o he RERI, he e was a signi ican posi i e addi i e in e ac- ion be ween sho sleep and insu icien LTPA o CVD mo ali y (RERIsho sleep + insu icien LTPA = 1.12; 95% CI, 0.19–2.06). The addi i e in e ac ion be ween sleep quali y and insu icien LTPA o CVD mo ali y was nonsigni ican (RERIpoo sleep + insu icien LTPA = 0.45; 95% CI, ˗0.26 o 1.15). Sensi i i y analyses Fu he adjus ing he Cox p opo ional haza ds models o li e sa - is ac ion as a p oxy o dep ession had a mino impac on he HRs o he in e p e a ional ou come o he models. A small pe cen age o pe sons in he mid- ange sleep and su icien LTPA g oup, as well as in he good sleep and su icien LTPA g oup epo ed sleep medica ion use (4.8% and 3.4%, espec i ely). Excluding hose epo ing use o sleep medica ion om he e e ence g oups, espec i ely, esul ed in i ually no changes in HRs o in e p e a- ional ou come o he models. Finally, no signi ican c ude associ- a ions we e ound o sleep du a ion o sleep quali y o he com- bina ions o sleep du a ion and insu icien LTPA and sleep quali y and insu icien LTPA wi h cance mo ali y ( esul s no shown). Discussion A ew signi ican mul iplica i e and addi i e in e ac ions be ween sleep du a ion and insu icien LTPA ega ding all-cause and CVD mo ali y we e ound. The main inding was ha sho sleepe s (6 hou s o less) wi h insu icien LTPA (less han 450 MET minu es pe week) had an inc eased all-cause and CVD mo ali y compa ed wi h hose wi h mid- ange sleep and su icien LTPA. No o he sig- ni ican associa ions we e ound. The in e ac ion be ween sleep and physical ac i i y o heal h has been ecognized,[52–54] bu he ew esul s ega ding in e ela- ionships be ween sleep, physical ac i i y, and mo ali y a e no cohe en and lea e some unanswe ed ques ions. Fi s o all, he classi ica ion o physical ac i i y in he s udies has di e ed. Bella ia e al.[11] who epo ed on a signi ican in e ac ion be ween physical ac i i y and sleep in ela ion o bo h all-cause and CVD mo ali- y classi ied daily physical ac i i y o he pa icipan s in o hi ds, whe e physical ac i i y ep esen ed he usual o al amoun o occu- pa ional, domes ic, leisu e ime, and walking o biking du ing he p e ious yea . The associa ion be ween sleep du a ion and mo al- i y was U-shaped in he wo lowes hi ds, bu no in he highes . TABLE 2. Haza d Ra ios and 95% CIs o Sleep Du a ion, Sleep Quali y, and Leisu e Time Physical Ac i i y P edic ing All-Cause and Ca dio ascula Mo ali y C ude Model 1 Model 2 Model 3 All-cause mo ali y (824 dea hs/1,638) Sho sleep 1.33 (1.05–1.68) 1.23 (0.95–1.59) 1.21 (0.93–1.57) 1.16 (0.89–1.50) Long sleep 0.92 (0.75–1.13) 0.89 (0.73–1.10) 0.90 (0.73–1.10) 0.85 (0.69–1.05) Poo sleep 1.29 (1.05–1.59) 1.16 (0.92–1.46) 1.14 (0.91–1.44) 0.99 (0.78–1.26) Insu icien LTPA 1.31 (1.14–1.50) 1.28 (1.10–1.48) 1.27 (1.09–1.48) 1.12 (0.96–1.31) CVD mo ali y (391 CVD cases/1,638) Sho sleep 1.33 (0.95–1.86) 1.22 (0.84–1.78) 1.15 (0.79–1.68) 1.13 (0.78–1.65) Long sleep 0.91 (0.68–1.23) 0.88 (0.65–1.18) 0.86 (0.64–1.16) 0.79 (0.59–1.07) Poo sleep 1.31 (0.97–1.77) 1.19 (0.85–1.65) 1.17 (0.84–1.63) 1.00 (0.71–1.41) Insu icien LTPA 1.44 (1.17–1.76) 1.42 (1.15–1.76) 1.42 (1.14–1.77) 1.21 (0.97–1.51) Model 1: sleep a iables, his o y o spo s and physical ac i i y; Model 2: Model 1 + socioeconomic a iables; Model 3: Model 2 + o he li es yles, sleep medica ion, and ch onic disease. Models 2 and 3 o all-cause mo ali y a e ex ended Cox models including an in e ac ion be ween occupa ional s a us and ime. Re e ence ca ego ies o sleep and LTPA we e ollowing: mid- ange sleep, good sleep, su icien LTPA. S a is ically signi ican (p<0.05) haza d a ios and CIs a e bolded. TABLE 3. Haza d Ra ios and 95% Con idence In e als o All-Cause Mo ali y by Sleep Du a ion and Leisu e Time Physical Ac i i y G oups and Sleep Quali y and Leisu e Time Physical Ac i i y G oups, Respec i ely C ude Model 1 Model 2 Fully Adjus ed Model Sleep du a ion and LTPA, e e ence: mid- ange sleep and su icien LTPA (387/861) Sho sleep and insu icien LTPA (38/56) 2.23 (1.59–3.11) 2.12 (1.51–2.98) 1.93 (1.37–2.71) 1.49 (1.05–2.11) Sho sleep and su icien LTPA (42/80) 1.12 (0.81–1.55) 1.10 (0.80–1.52) 1.13 (0.82–1.56) 1.04 (0.75–1.45) Long sleep and insu icien LTPA (47/78) 0.96 (0.71–1.30) 0.93 (0.68–1.26) 0.95 (0.70–1.30) 0.80 (0.58–1.11) Long sleep and su icien LTPA (68/109) 1.05 (0.81–1.36) 1.05 (0.81–1.36) 1.03 (0.80–1.34) 0.99 (0.76–1.29) Mid- ange sleep and insu icien LTPA (242/454) 1.34 (1.14–1.58) 1.30 (1.10–1.53) 1.29 (1.09–1.53) 1.15 (0.97–1.37) Sleep quali y and LTPA, e e ence: good sleep and su icien LTPA (446/950) Poo sleep and insu icien LTPA (51/73) 2.05 (1.53–2.74) 1.94 (1.44–2.61) 1.84 (1.36–2.49) 1.29 (0.94–1.77) Poo sleep and su icien LTPA (51/100) 1.05 (0.79–1.41) 1.03 (0.77–1.38) 1.04 (0.77–1.39) 0.94 (0.69–1.27) Good sleep and insu icien LTPA (276/515) 1.23 (1.06–1.43) 1.19 (1.02–1.39) 1.20 (1.03–1.41) 1.08 (0.92–1.27) Model 1: his o y o spo s; model 2: model 1 + socioeconomic a iables; ully adjus ed model: model 2 + o he li es yles, sleep medica ion, and ch onic disease. Models 2 and he ully adjus ed model o all-cause mo ali y a e ex ended Cox models including an in e ac ion be ween occupa ional s a us and ime. Numbe o cases and numbe o men in ha ca ego y a baseline a e gi en in pa en heses. S a is ically signi ican (p<0.05) haza d a ios and CIs a e bolded. 6 Wennman e al. www.p ogp e med.com Xiao e al.[15] who did no ind any signi ican in e ac ion assessed physical ac i i y as he ypical equency o mode a e- o- igo ous in ensi y physical ac i i y du ing he pas 10 yea s. A cu o o high physical ac i i y was se a 1 hou o mode a e- o- igo ous in en- si y physical ac i i y pe week. This is less han he ecommended weekly physical ac i i y dose o adul s ha we e used as a c i e i- on o su icien LTPA in he cu en s udy. Howe e , i mus be kep in mind ha ou measu e o physical ac i i y did no include occu- pa ional o commu ing physical ac i i y, which can lead o lowe p opo ion o hose mee ing he su icien physical ac i i y le el. No only sleep du a ion, bu also sleep quali y in ela ion o mo ali y wa an conside a ion. We also s udied sleep quali y bu did no ind any s a is ically signi ican in e ac ion be ween poo sleep quali y and insu icien LTPA wi h all-cause o CVD mo ali y. Nei he he s udy o Bella ia e al.[11] o Xiao e al.[15] examined sleep quali y. Sleep du a ion and sleep quali y may be o e lapping, ye di e en cha ac e is ics o sleep[5,10] and he solu ion o poo sleep is no always a change in sleep du a ion.[10,55] Epidemiological s udies ha e shown signi ican associa ions be ween measu es o sleep quali y and ca dio ascula heal h.[56] Howe e , in adul Finn- ish win pai s,[40] and among Ame ican men and women,[57] sleep du a ion a he han sleep quali y showed a signi ican associa ion wi h mo ali y. Poo sleep can be ini ia ed by and ela ed o se e al ac o s,[58] one being ch ono ype.[59] In his cu en coho o o me a hle es, B oms e al.[49] p e iously obse ed e ening ypes o ha e a highe mo ali y han mo ning ypes. We included ch ono ype as a co a ia e in he Cox models, bu ou esul did no show any independen e ec o ch ono ype in he models. Ou indings add o he spa se li e a u e on he opic and sugges ha LTPA modi ies he associa ion pa icula ly be ween sho sleep and ca dio ascula mo ali y whe e he coexis ence o sho sleep and insu icien LTPA con ibu e signi ican ly o he excess CVD mo - ali y isk. Sho sleep associa es wi h impai ed glucose me abolism and obesi y, low-g ade in lamma ion, and hype ension,[6,56] all ha make up possible mechanisms o he ela ionship be ween sho sleep and CVD mo ali y.[1,2,5] Also, physical ac i i y modi ies ca - diome abolic isk ac o s, and low physical ac i i y is an es ablished isk ac o o CVD mo ali y.[17,20] Thus, i is no su p ising o ind he combina ion o sho sleep and insu icien LTPA o p edic a highe CVD mo ali y. Simila o us, Xiao e al.[15] also s udied he addi i e in e ac ion be ween sleep and physical ac i i y, bu hey did no ind any signi ican addi i e in e ac ion. Whe e we chose o include he long sleepe s in he e e ence g oup o ou addi i e in e ac ion analyses, Xiao e al. excluded he long sleepe s om hose analyses. Con a y o p e ious li e a u e, we did no see any clea U-shaped associa ion be ween sleep du a ion and mo ali y in ou models. The wo ends o he U-shaped associa ion be ween sleep du a ion and mo ali y a e p oposed o be explained by di e en mechanisms.[1,8,10] I has, o example, been sugges ed ha heal h s a us modi ies he associa ions be ween sleep and mo ali y, pa icula ly long sleep and mo ali y.[4,9] In he cu en s udy, long sleep was no ela ed o inc eased mo ali y nei he in c ude o adjus ed models and he e we e no s a is ically signi ican in e ac- ions be ween his o y o disease and sleep wi h ega d o mo ali- y. The su i al o he heal hies un il baseline measu emen s may ha e a ec ed he esul in he cu en s udy, especially ega ding long sleep. Bella ia e al.[11] who obse ed ha long sleep associ- a e wi h mo ali y only among subjec s wi h low physical ac i i y, could no con ol o dep ession. We assessed li e sa is ac ion, a co ela e o dep ession,[48] bu did no ind i o impac he nonsig- ni ican associa ions be ween long sleep and mo ali y. Some di e ences in he dis ibu ion o sleep du a ion and sleep quali y and le el o LTPA be ween o me a hle es and non- a hle es we e obse ed in his s udy. The o me a hle e men mo e o en had mid- ange han sho sleep as compa ed wi h he nona hle es, and su icien LTPA was as common in sho as in long sleepe s among he o me a hle es. As a compa ison, in he coho o Bella ia e al.,[11] high physical ac i i y was epo ed o a subs an ially highe pe cen age o sho han long sleepe s (53% and 16%, espec i ely). In hei e iew, D i e and Taylo [28] sugges ed ha a hle ic indi iduals sleep be e han nona hle ic coun e pa s whe he hey a e aining o no . Ca dio espi a o y i ness ha de elops h ough egula physical ac i i y and exe - cise is sugges ed as one mechanism linking physical ac i i y wi h good sleep.[12,28,29] Many o me a hle es sus ain good i ness le els a e hei ac i e ca ee mainly by emaining physically ac i e.[22] Ne e heless, in he cu en s udy, he associa ion be ween sleep du a ion o sleep quali y and mo ali y was no di e en be ween a hle es and nona hle es, and sleep was no an independen p e- dic o o mo ali y in ei he g oup. The occu ence o diseases and heal h issues di e be ween he o me a hle es and nona hle es, and also o some ex en be ween a hle es om di e en spo s.[23,60] Gene ic ac o s ha migh unde lie he selec ion o spo s and di e en aining eg- imens, bu also he backg ound o igo ous physical ac i i y included in aining o high-le el compe i i e spo s may explain some o he heal h di e ences be ween a hle es and nona hle es and also be ween di e en a hle es, espec i ely.[60] The longe li e TABLE 4. Haza d Ra ios and 95% Con idence In e als o Ca dio ascula Mo ali y by Sleep Du a ion and Leisu e Time Physical Ac i i y G oups and Sleep Quali y and Leisu e Time Physical Ac i i y G oups, Respec i ely C ude Model 1 Model 2 Fully adjus ed model Sleep du a ion and LTPA, e e ence: mid- ange sleep and su icien LTPA (180/861) Sho sleep and insu icien LTPA (23/56) 2.90 (1.87–4.48) 2.79 (1.79–4.36) 2.57 (1.64–4.02) 1.98 (1.25–3.12) Sho sleep and su icien LTPA (15/80) 0.85 (0.50–1.45) 0.84 (0.50–1.43) 0.81 (0.47–1.38) 0.75 (0.44–1.30) Long sleep and insu icien LTPA (23/78) 0.99 (0.64–1.54) 0.97 (0.62–1.51) 0.95 (0.60–1.49) 0.75 (0.47–1.18) Long sleep and su icien LTPA (31/109) 1.04 (0.71–1.53) 1.04 (0.71–1.52) 1.03 (0.70–1.51) 0.96 (0.65–1.42) Mid- ange sleep and insu icien LTPA (119/454) 1.41 (1.18–1.78) 1.37 (1.08–1.75) 1.38 (1.08–1.76) 1.18 (0.92–1.52) Sleep quali y and LTPA, e e ence: good sleep and su icien LTPA (204/950) Poo sleep and insu icien LTPA (27/73) 2.37 (1.58–3.54) 2.28 (1.50–3.46) 2.17 (1.42–3.31) 1.46 (0.94–2.27) Poo sleep and su icien LTPA (22/100) 0.99 (0.64–1.54) 0.98 (0.63–1.52) 0.97 (0.62–1.51) 0.87 (0.55–1.39) Good sleep and insu icien LTPA (138/515) 1.33 (1.07–1.65) 1.30 (1.04–1.63) 1.31 (1.04–1.66) 1.14 (0.90–1.44) Model 1: his o y o spo s; model 2: Model 1 + socioeconomic a iables; ully adjus ed model: Model 2 + o he li es yles, sleep medica ion, and ch onic disease. Numbe o cases and numbe o men in ha ca ego y a baseline a e gi en in pa en heses. S a is ically signi ican (p<0.05) haza d a ios and CIs a e bolded. 7 Wennman e al. www.p ogp e med.com expec ancy in o me a hle es is o a high deg ee explained by a dec eased CVD mo ali y and lowe cance isk in his popula ion, as epo ed ea lie .[22] We did no obse e any in e ac ions be ween sleep, LTPA, and cance mo ali y ( esul s no shown), and no did Bella ia e al.[11] in hei s udy. F equen sno ing was epo ed by 45% in he sho sleep and insu icien LTPA g oup, bu i is unlike- ly hough ha mo ali y isk is due o undiagnosed sleep apnoea. Sno ing was no independen ly ela ed o mo ali y and excluding 5% o he o al sample wi h equen sno ing and a BMI ≥ 30 kg/m2 did no change he ou come o he inal analyses. Limi a ions o he s udy One weakness o ou s udy is he use o sel - epo ed measu es o sleep and physical ac i i y ha include he possibili y o epo - ing bias. Howe e , a he ime o he baseline ques ionnai e, a ail- abili y and use o objec i e accele ome e s was limi ed. Sleep du a ion was measu ed wi h one ca ego ical ques ion, and i is he e o e no possible o analyze a ue con inuous measu e o sleep du a ion. Fu he mo e, we do no know how he pa ici- pan s, especially he a hle es slep when hey we e ac i e. Thus, i emains unknown whe he o no he e ha e been changes in hei sleep du a ion a e hei ac i e ca ee , and i so, o wha ex en . This kind o ansi ion has also no , o ou knowledge, been s udied elsewhe e. Whe e almos hal o hose wi h sho sleep and insu icien LTPA we e o me a hle es, hese subjec s can ha e had insu icien sleep al eady in hei ac i e days, caus- ing hem o e en ually s op being physically ac i e. Ano he sce- na io is ha hey ha e become sho sleepe s only la e in li e, as a consequence o he a hle ic e i emen . Findings suppo ing he la e heo y come om he Ae obics Cen e coho and sugges ha a decline in ca dio espi a o y i ness ha mimics a educed physical ac i i y le el p edisposes o sho e sleep and poo e quali y o sleep.[61] I has also been obse ed ha dec easing he amoun o daily physical ac i i y in ac i e a hle es has a nega i e impac on sleep.[62] O he ac o s such as heal h s a us o socioeconomic s a us a e also p one o change o e ime. In addi ion, he e a e many impo an bidi ec ional associa ions be ween heal h and beha - io s and a change in weigh can, o example, ha e majo e ec s on sleep quali y and quan i y as well as nume ous heal h indices. We a e limi ed by he ac ha we do no know wha changes in he a hle es’ and nona hle es’ li e ha e led o he epo ed sleep in 1985. The ole o heal h s a us was con olled o by excluding ea ly dea hs du ing he ollow-up, and models we e adjus ed o by he known his o y o disease. The in e ac ion be ween his o y o disease and sleep wi h mo ali y was also examined bu ound nonsigni ican . Rega ding socioeconomic s a us, each pe son was classi ied acco ding o he occupa ion he had p ac iced o he lon- ges pe iod un il 1985,[22] hus pa ly con olling o he change in occupa ional s a us aken place be o e he baseline measu emen . Fu he mo e, ma i al s a us a baseline was also aken in o accoun . I also needs o be acknowledged ha he sample comp ised only men and hus he s udy lea es a lack o knowledge ega d- ing women a hle es’ sleep and mo ali y. The e may also be some selec ion bias ela ed o he o ming o he coho . As s a ed, ou coho includes o me op-le el a hle es who a e a selec ed g oup wi h good heal h a baseline.[22,60] Fu he mo e, he nona hle es also ep esen he mos heal hy and i popula ion as de e mined in he medical examina ion p eceding hei consc ip ion a age 20. I is no easy o ind ma ching subjec s o wo ld class a hle es, bu in his coho , he compa abili y o he nona hle es o he a h- le es can be conside ed good.[22] E en i he esul s a e no di ec ly gene alizable o he a e age popula ion, hey highligh he impo - ance o join beha io s such as physical ac i i y and sleep o CVD isk, e en in such a selec ed popula ion. Conclusions In his unique coho including o me eli e a hle es and nona h- le e men, we obse ed a signi ican in e ac ion be ween sleep and LTPA wi h mo ali y, pa icula ly CVD mo ali y isk. Highe isk was obse ed o men ha ing sho sleep and insu icien LTPA e en a e adjus ing o a his o y o spo s, and a a ie y o o he demog aphic, beha io al, and heal h- ela ed ac o s. The esul s also sugges ha on an addi i e scale, he combina ion o sho sleep and insu icien LTPA is associa ed wi h a la ge isk han he sum o he es ima ed isk o sho sleep o LTPA alone. I can be concluded ha a o me spo s ca ee does no ully p o ec wi h ega d o mo ali y isk, and he nega i e impac o sho sleep a he a ies acco ding o LTPA le el. Fu u e s udies a e s ill needed o eplica e he indings o an in e ac ion be ween sleep and physical ac i i y o mo ali y, ye in di e en popula ions. The CVD isk p o ile including bo h beha - io al and heal h indices o sho sleepe s wi h low LTPA, need o be de e mined u he o p o ide mo e ools o heal h p omo ion and p e en i e wo k. I is especially impo an also o in es iga e how changes in sleep du a ion o physical ac i i y le els and how di e - en ypes o physical ac i i y a ec he isk p o ile. Disclosu e The Fo me eli e a hle e coho s udy has been unded by he Minis y o Educa ion and Cul u e, he Juho Vainio Founda ion, he Finnish Hea Resea ch Founda ion, Paa o Nu mi Founda ion, he Swedish Cul u al Founda ion in Finland, he Finnish Cul u al Founda ion, and by a g an om Medical Socie y o Finland, Finska Läka esällskape . J.K. is suppo ed by he Academy o Finland (g an #263278). This s udy has u he been suppo ed by he Juho Vainio Founda ion, he Finnish Cul u al Founda ion, and by a g an om Medical Socie y o Finland, Finska Läka esällskape . The au ho s ha e no inancial in e es o decla e in ela ion o he con en o his a icle. The A icle P ocessing Cha ge was paid o by he au ho s. Re e ences [1] Gallicchio L, Kalesan B. Sleep du a ion and mo ali y: a sys ema ic e iew and me a-analysis. J Sleep Res. 2009;18:148–158. [2] Cappuccio FP, Coope D, D’Elia L, e al. Sleep du a ion p edic s ca dio ascu- la ou comes: a sys ema ic e iew and me a-analysis o p ospec i e s udies. Eu Hea J. 2011;32:1484–1492. [3] Liu TZ, Xu C, Ro a M, e al. Sleep du a ion and isk o all-cause mo ali y: a lexible, non-linea , me a- eg ession o 40 p ospec i e coho s udies. Sleep Med Re . 2017;32:28–36. [4] Ku ina LM, McClin ock MK, Chen JH, e al. Sleep du a ion and all-cause mo ali y: a c i ical e iew o measu emen and associa ions. Ann Epidemiol. 2013;23:361–370. [5] G andne MA, Pa el NP, Geh man PR, e al. P oblems associa ed wi h sho sleep: b idging he gap be ween labo a o y and epidemiological s udies. Sleep Med Re . 2010;14:239–247. [6] Knu son KL. Sleep du a ion and ca diome abolic isk: a e iew o he epide- miologic e idence. Bes P ac Res Clin Endoc inol Me ab. 2010;24:731–743. [7] Cappuccio FP, D’Elia L, S azzullo P, e al. Sleep du a ion and all-cause mo - ali y: a sys ema ic e iew and me a-analysis o p ospec i e s udies. Sleep. 2010;33:585–592. [8] Knu son KL, Tu ek FW. The U-shaped associa ion be ween sleep and heal h: he 2 peaks do no mean he same hing. Sleep. 2006;29:878–879. [9] Magee CA, Holliday EG, A ia J, e al. In es iga ion o he ela ionship be- ween sleep du a ion, all-cause mo ali y, and p eexis ing disease. Sleep Med. 2013;14:591–596. [10] G andne MA, D ummond SP. Who a e he long sleepe s? Towa ds an unde - s anding o he mo ali y ela ionship. Sleep Med Re . 2007;11:341–360. [11] Bella ia A, Åke s ed T, Bo ai M, e al. Sleep du a ion and su i al pe cen iles ac oss ca ego ies o physical ac i i y. Am J Epidemiol. 2014;179:484–491. 8 Wennman e al. www.p ogp e med.com [12] K edlow MA, Capozzoli MC, Hea on BA, e al. The e ec s o physical ac i i y on sleep: a me a-analy ic e iew. J Beha Med. 2015;38:427–449. [13] Pepin V, Bouce a S, Chan-Thim E, e al. The ole o sleep and physical ac i i y on he isk o ca dio ascula disease. Cu en Ca dio ascula Risk Repo s. 2014;8(12):413–422. [14] Buman MP, Winkle EA, Ku ka JM, e al. Realloca ing ime o sleep, seden a y beha io s, o ac i e beha io s: associa ions wi h ca dio ascula disease isk bioma ke s, NHANES 2005-2006. Am J Epidemiol. 2014;179:323–334. [15] Xiao Q, Keadle SK, Hollenbeck AR, e al. Sleep du a ion and o al and cause-speci ic mo ali y in a la ge US coho : in e ela ionships wi h phys- ical ac i i y, seden a y beha io , and body mass index. Am J Epidemiol. 2014;180:997–1006. [16] Sami z G, Egge M, Zwahlen M. Domains o physical ac i i y and all-cause mo ali y: sys ema ic e iew and dose- esponse me a-analysis o coho s ud- ies. In J Epidemiol. 2011;40:1382–1400. [17] Lee IM, Shi oma EJ, Lobelo F, e al.; Lance Physical Ac i i y Se ies Wo k- ing G oup. E ec o physical inac i i y on majo non-communicable diseas- es wo ldwide: an analysis o bu den o disease and li e expec ancy. Lance . 2012;380:219–229. [18] Wo ld Heal h O ganiza ion, ed. Global Recommenda ions on Physical Ac i i y o Heal h. Gene a, Swi ze land: Wo ld Heal h O ganiza ion; 2010. [19] Blai SN, Cheng Y, Holde JS. Is physical ac i i y o physical i ness mo e impo an in de ining heal h bene i s? Med Sci Spo s Exe c. 2001;33(6 Sup- pl):S379–S399; discussion S419. [20] Mye s J, McAuley P, La ie CJ, e al. Physical ac i i y and ca dio espi a o y i ness as majo ma ke s o ca dio ascula isk: hei independen and in e - wo en impo ance o heal h s a us. P og Ca dio asc Dis. 2015;57:306–314. [21] Pa enba ge RS J , Lee IM. A na u al his o y o a hle icism, heal h and lon- ge i y. J Spo s Sci. 1998;16:S31–S45. [22] Sa na S, Sahi T, Kosken uo M, e al. Inc eased li e expec ancy o wo ld class male a hle es. Med Sci Spo s Exe c. 1993;25:237–244. [23] Laine MK, E iksson JG, Kujala UM, e al. Fo me male eli e a hle es ha e be - e me abolic heal h in la e li e han hei con ols. Scand J Med Sci Spo s. 2016;26:284–290. [24] Ba is a C, Soa es JM. A e o me eli e a hle es mo e p o ec ed agains me a- bolic synd ome? J Ca diol. 2013;61:440–445. [25] Almeida OP, Khan KM, Hankey GJ, e al. 150 minu es o igo ous physical ac i i y pe week p edic s su i al and success ul ageing: a popula ion-based 11-yea longi udinal s udy o 12 201 olde Aus alian men. B J Spo s Med. 2014;48:220–225. [26] Tudo -Locke C, Leona di C, Johnson WD, e al. Time spen in physical ac i i- y and seden a y beha io s on he wo king day: he Ame ican ime use su ey. J Occup En i on Med. 2011;53:1382–1387. [27] T emblay MS, Colley RC, Saunde s TJ, e al. Physiological and heal h implica- ions o a seden a y li es yle. Appl Physiol Nu Me ab. 2010;35:725–740. [28] D i e HS, Taylo SR. Exe cise and sleep. Sleep Med Re . 2000;4:387–402. [29] Chennaoui M, A nal PJ, Sau e F, e al. Sleep and exe cise: a ecip ocal issue? Sleep Med Re . 2015;20:59–72. [30] Uchida S, Shioda K, Mo i a Y, e al. Exe cise e ec s on sleep physiology. F on Neu ol. 2012;3:48. [31] Fullaga HH, Sko ski S, Du ield R, e al. Sleep and a hle ic pe o mance: he e ec s o sleep loss on exe cise pe o mance, and physiological and cogni i e esponses o exe cise. Spo s Med. 2015;45:161–186. [32] Halson SL. Sleep in eli e a hle es and nu i ional in e en ions o enhance sleep. Spo s Med. 2014;44:S13–S23. [33] Las ella M, Roach GD, Halson SL, e al. Sleep/wake beha iou s o eli e a hle es om indi idual and eam spo s. Eu J Spo Sci. 2015;15:94–100. [34] Sa gen C, Las ella M, Halson SL, e al. The impac o aining schedules on he sleep and a igue o eli e a hle es. Ch onobiol In . 2014;31:1160–1168. [35] Shepha d RJ, Ka anagh T, Me ens DJ, e al. Pe sonal heal h bene i s o Mas- e s a hle ics compe i ion. B J Spo s Med. 1995;29:35–40. [36] Fogelholm M, Kap io J, Sa na S. Heal hy li es yles o o me Finnish wo ld class a hle es. Med Sci Spo s Exe c. 1994;26:224–229. [37] Bäckmand H, Kujala U, Sa na S, e al. Fo me a hle es’ heal h- ela ed li e- s yle beha iou s and sel - a ed heal h in la e adul hood. In J Spo s Med. 2010;31:751–758. [38] Ke unen JA, Kujala UM, Kap io J, e al. All-cause and disease-speci ic mo al- i y among male, o me eli e a hle es: an a e age 50-yea ollow-up. B J Spo s Med. 2015;49:893–897. [39] Hublin C, Pa inen M, Kosken uo M, e al. Gene ic ac o s in e olu ion o sleep leng h—a longi udinal win s udy in Finnish adul s. J Sleep Res. 2013;22:513–518. [40] Hublin C, Pa inen M, Kosken uo M, e al. Sleep and mo ali y: a popula- ion-based 22-yea ollow-up s udy. Sleep. 2007;30:1245–1253. [41] Pa inen M, Kap io J, Kosken uo M, e al. Gene ic and en i onmen al de e - mina ion o human sleep. Sleep. 1983;6:179–185. [42] Walle K, Kap io J, Kujala UM. Associa ions be ween long- e m physical ac- i i y, wais ci cum e ence and weigh gain: a 30-yea longi udinal win s udy. In J Obes (Lond). 2008;32:353–361. [43] Bäckmand H, Kap io J, Kujala U, e al. In luence o physical ac i i y on dep es- sion and anxie y o o me eli e a hle es. In J Spo s Med. 2003;24:609–619. [44] Kap io J, Kosken uo M. A p ospec i e s udy o psychological and socioeco- nomic cha ac e is ics, heal h beha io and mo bidi y in ciga e e smoke s p io o qui ing compa ed o pe sis en smoke s and non-smoke s. J Clin Epidemiol. 1988;41:139–150. [45] Romano K, Rose RJ, Kap io J, e al. Sel - epo ed alcohol use: a longi udinal s udy o 12,994 adul s. Alcohol Alcohol Suppl. 1987;1:619–623. [46] Jä enpää T, Rinne JO, Kosken uo M, e al. Binge d inking in midli e and demen ia isk. Epidemiology. 2005;16:766–771. [47] Koi umaa-Honkanen H, Honkanen R, Viinamäki H, e al. Sel - epo ed li e sa is ac ion and 20-yea mo ali y in heal hy Finnish adul s. Am J Epidemiol. 2000;152:983–991. [48] Bäckmand H, Kap io J, Kujala U, e al. Pe sonali y and mood o o me eli e male a hle es—a desc ip i e s udy. In J Spo s Med. 2001;22:215–221. [49] B oms U, Pi käniemi J, Bäckmand H, e al. Long- e m consis ency o diu - nal- ype p e e ences among men. Ch onobiol In . 2014;31:182–188. [50] Ho ne JA, Os be g O. A sel -assessmen ques ionnai e o de e mine mo ning- ness-e eningness in human ci cadian hy hms. In J Ch onobiol. 1976;4:97– 110. [51] Knol MJ, Vande Weele TJ. Recommenda ions o p esen ing analyses o e ec modi ica ion and in e ac ion. In J Epidemiol. 2012;41:514–520. [52] A kinson G, Da enne D. Rela ionships be ween sleep, physical ac i i y and human heal h. Physiol Beha . 2007;90:229–235. [53] Ding D, Roge s K, Macni en R, e al. Re isi ing li es yle isk index assessmen in a la ge Aus alian sample: should seden a y beha io and sleep be included as addi ional isk ac o s? P e Med. 2014;60:102–106. [54] Zuo H, Shi Z, Yuan B, e al. In e ac ion be ween physical ac i i y and sleep du a ion in ela ion o insulin esis ance among non-diabe ic Chinese adul s. BMC Public Heal h. 2012;12:247. [55] Buysse DJ, G uns ein R, Ho ne J, e al. Can an imp o emen in sleep posi i ely impac on heal h? Sleep Med Re . 2010;14:405–410. [56] S -Onge MP, G andne MA, B own D, e al.; Ame ican Hea Associa ion Obesi y, Beha io Change, Diabe es, and Nu i ion Commi ees o he Coun- cil on Li es yle and Ca diome abolic Heal h; Council on Ca dio ascula Disease in he Young; Council on Clinical Ca diology; and S oke Council. Sleep du a ion and quali y: impac on li es yle beha io s and ca diome abolic heal h: a scien i ic s a emen om he Ame ican Hea Associa ion. Ci cula- ion. 2016;134:e367–e386. [57] K ipke DF, Ga inkel L, Winga d DL, e al. Mo ali y associa ed wi h sleep du a ion and insomnia. A ch Gen Psychia y. 2002;59:131–136. [58] Ohayon MM. Epidemiology o insomnia: wha we know and wha we s ill need o lea n. Sleep Med Re . 2002;6:97–111. [59] Ba clay NL, Eley TC, Buysse DJ, e al. Diu nal p e e ence and sleep quali y: same genes? A s udy o young adul wins. Ch onobiol In . 2010;27:278–296. [60] Kujala UM, Ma i P, Kap io J, e al. Occu ence o ch onic disease in o me op-le el a hle es. P edominance o bene i s, isks o selec ion e ec s? Spo s Med. 2003;33:553–561. [61] Dishman RK, Sui X, Chu ch TS, e al. Decline in ca dio espi a o y i ness and odds o inciden sleep complain s. Med Sci Spo s Exe c. 2015;47:960– 966. [62] Hague JF, Gilbe SS, Bu gess HJ, e al. A seden a y day: e ec s on subsequen sleep and body empe a u es in ained a hle es. Physiol Beha . 2003;78:261–267.