Prostate cancer-specific survival among warfarin users in the Finnish Randomized Study of Screening for Prostate Cancer

RESEARCH ARTICLE Open Access P os a e cance -speci ic su i al among wa a in use s in he Finnish Randomized S udy o Sc eening o P os a e Cance Pe e T. T. Kinnunen 1* , Teemu J. Mu ola 1,2 , Ki si Talala 3 , Kimmo Taa i 4 , Teu o L. J. Tammela 1,2 and Anssi Au inen 5 Abs ac Backg ound: Venous h omboembolic e en s (VTE) a e common in cance pa ien s and associa ed wi h highe mo ali y. In i o h ombosis and an icoagula ion migh be in ol ed in umo g ow h and p og ession. We s udied he associa ion o wa a in and o he an icoagulan use as an i h ombo ic medica ion and p os a e cance (PCa) dea h in men wi h he disease. Me hods: The s udy included 6,537 men diagnosed wi h PCa du ing 1995-2009. In o ma ion on an icoagulan use was ob ained om a na ional eimbu semen egis y. Cox eg ession wi h adjus men o age, PCa isk g oup, p ima y he apy and use o o he medica ion was pe o med o compa e isk o PCa dea h be ween wa a in use s wi h 1) men using o he ypes o an icoagulan s and 2) non-use s o an icoagulan s. Medica ion use was analyzed as a ime-dependen a iable o minimize immo al ime bias. Resul s: In o al, 728 men died om PCa du ing a median ollow-up o 9 yea s. Compa ed o an icoagulan non- use s, pos -diagnos ic use o wa a in was associa ed wi h an inc eased isk o PCa dea h (o e all HR 1.47, 95% CI 1. 13-1.93). Howe e , his was limi ed o low-dose, low-in ensi y use. O he wise, he isk was simila o an icoagulan non-use s. Addi ionally, we ound no isk di e ence be ween wa a in and o he ypes o an icoagulan s. P e-diagnos ic use o wa a in was no associa ed wi h he isk o PCa dea h. Conclusions: We ound no educ ion in isk o PCa dea h associa ed wi h wa a in use. Con e sely, he isk was inc eased in sho - e m use, which is p obably explained by a highe isk o h ombo ic e en s p omp ing wa a in use in pa ien s wi h e minal PCa. Keywo ds: P os a e cance , Wa a in, An icoagulan , Su i al, Coho Backg ound Venous h omboembolism (VTE) has been p oposed as p ognos ic ac o in p os a e cance (PCa). VTE is common in cance pa ien s and associa ed wi h poo p ognosis, isk o dea h is 8- old highe in cance pa ien wi h VTE [1–3]. This applies o PCa as well [4].VTE in PCa pa ien s has been associa ed wi h mo e han 6- old mo ali y in bo h symp oma ic and inciden al enous h omboembolic diseases [5]. Thus,an icoagulan d ugs could ha e an impac on PCa p ognosis by educing dea hs om VTE.Fu he mo e, an icoagulan s, including wa a in, ha e shown p omising an i umo p ope ies in i o [6–10] mainly in lung and b eas cance . P e ious ai ly small s udies on an icoagulan use and PCa su i al ha e epo ed di e ing esul s [11–13]. A po en ial explana ion could be con ounding by indica- ion; pa ien s wi h an ad anced cance a e a inc eased isk o VTE [14], hus mo e o en p esc ibed an icoagu- lan s compa ed o people wi hou cance . We conduc ed a e ospec i e, popula ion-based coho s udy o assess he associa ion o p e- and pos -diagnos ic use o wa a in and o he an icoagulan s wi h PCa su i al in he Finnish Randomized S udy o Sc eening o P os a e Cance (FinRSPC) [15]. * Co espondence: [email p o ec ed] 1 Uni e si y o Tampe e, Facul y o Medicine and Li e Sciences, Tampe e, Finland Full lis o au ho in o ma ion is a ailable a he end o he a icle © The Au ho (s). 2017 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Kinnunen e al. BMC Cance (2017) 17:585 DOI 10.1186/s12885-017-3579-8 Me hods S udy coho FinRSPC includes 80,458 men aged 55-67 yea s a base- line (i.e. a FinRSPC andomiza ion). A e exclusion o p e alen PCa cases, he men we e andomized du ing 1996-1999 ei he o PSA sc eening a ou -yea in e als ( he sc eening a m) o o no in e en ion ( he con ol a m). All men we e ollowed ia Finnish Cance Regis y, which co e s 99% o cance s diagnosed in Finland [16]. This s udy included 6,537 inciden PCa cases diagnosed du ing 1996-2013. Clinical in o ma ion included Gleason g ade and TNM s age (a ailable o 97.3% and 97.7% o he cases, espec i ely). PCa cases we e s a i ied in o low/in e media e- isk and high- isk g oups acco ding o he de ini ion o he Eu opean Associa ion o U ology (EAU) [17]. All M1 cases we e included in he high- isk g oup. In o ma ion on dea hs was ob ained om he S a is ics Finland, which assigns o icial causes o dea h based on manda o y dea h ce i ica es, co e ing all dea hs in Finland [18]. The accu acy o in o ma ion on PCa dea hs was asce ained by he FinRSPC cause o dea h commi - ee wi h an excellen conco dance be ween o icial causes o dea h and he cause o dea h commi ee assignmen s o PCa (kappa 0.95) [15]. Only dea hs wi h PCa (ICD-10 code C61) as he p ima y cause o dea h we e ega ded as PCa dea hs. A sensi i i y analysis wi h PCa as an in e media e cause o dea h was also pe o med. In o ma ion on an icoagulan usage In o de o ob ain in o ma ion on an icoagulan d ug pu chases du ing 1995-2009, he s udy coho was linked o a na ional medica ion eimbu semen da abase main ained by he Finnish Social Insu ance Ins i u ion (SII) using he unique pe sonal iden i ica ion code as he key. As a pa o he na ional heal h insu ance ha co e s all Finnish ci izens, SII p o ides eimbu semen s o pu chases o physician-p esc ibed d ugs [19].The e- imbu semen is 50-100 % depending on he indica ion and se e i y o he condi ion. In Finland, an icoagulan d ugs a e a ailable only h ough physicians’p esc ip- ion, hus all an icoagulan pu chases in ou pa ien se - ing a e egis e ed by he da abase. D ugs used du ing hospi al inpa ien pe iods a e no egis e ed. All 13 an icoagulan d ugs used in ou pa ien se ing du ing he s udy pe iod we e iden i ied based on hei ATC codes (Addi ional ile 1: Table S1). Addi ionally, we ob ained in o ma ion on choles e ol-lowe ing d ugs, an idiabe ic and an ihype ensi e d ugs, aspi in and o he NSAIDs and alpha-blocke s. We also collec ed in o ma ion on p ima y he apy o PCa ( adical p os a- ec omy, ex e nal beam adia ion he apy (EBRT), ho - monal he apy o ac i e su eillance/wa ch ul wai ing). O he medica ion se ed as a p oxy o co-mo bidi ies, as hey may in luence su i al [20–25]. The in o ma ion om he Na ional Ca e Regis e o Heal h Ca e main ained by he Na ional Ins i u e o Heal h and Wel a e included p o ided diagnoses o condi ions se ing as indica ions o an icoagulan use: a ial ib illa ion (ICD-10: I48), enous h omboembol- ism (all eco ded I82 diagnoses in he s udy popula ion), pulmona y embolism (ICD-10: I26.0, I26.9) and h om- bocy osis. The egis e co e s all o Finland and egis e s all diagnoses made du ing in- and ou -pa ien hospi al isi s du ing 1996-2014, bu does no co e diagnoses om p ima y ca e [26]. Addi ionally, we s a i ied he analysis by Cha lson Como bidi y Sco e [27] calcula ed on basis o he egis e ed diagnoses o subg oup analysis. S a is ical analysis Cox p opo ional haza ds eg ession was used o calcu- la e haza d a ios (HR) and 95% con idence in e als (CIs) o PCa dea h. Follow-up s a ed a PCa diagnosis and con inued un il dea h, emig a ion om Finland o Janua y 1 s , 2015, whiche e was i s . Time me ic was yea s and mon hs since he PCa diagnosis. We used wo di e en model adjus men s. The i s Cox eg ession model was adjus ed o age only and he second o age, EAU umo isk g oup and o he medica ions. The analysis was pe o med sepa a ely o p e-diagnos ic and pos -diagnos ic use o wa a in and addi ionally o o he ypes o an icoagulan s. Sepa a e compa isons we e pe o med be ween all an icoagulan use s and non-use s o es ima e he o e all e ec o an icoagulan usage, and be ween wa a in use s and men using non-wa a in an icoagulan s o es ima e speci ic e ec s o wa a in wi h simul aneous con ol o con ounding by indica ion. Fo each man in he s udy coho , he o al annual amoun o medica ion pu chases was calcula ed o each calenda yea , and hen sepa a ely o p e- and pos - diagnos ic use. S anda diza ion o doses be ween di e - en an icoagulan s was pe o med by di iding he annual o al millig am amoun wi h he s anda d De ined Daily Dose (DDD) as lis ed by he WHO [28]. Each yea wi h egis e ed an icoagulan pu chases, ega dless o he amoun , was conside ed a yea o usage. The in ensi y o he use was calcula ed by di iding he cumula i e annual doses wi h he numbe o yea s o usage. P e-diagnos ic medica ion use was analyzed as a ime- independen a iable, and cumula i e p e-diagnos ic an icoagulan usage beginning om 1995 was s a i ied by e iles. Pos -diagnos ic use was analyzed as a ime- dependen a iable; he usage s a us and cumula i e use we e upda ed sepa a ely o each yea o ollow-up, beginning om he yea o diagnosis. Kinnunen e al. BMC Cance (2017) 17:585 Page 2 o 9 All men we e ca ego ized as non-use s un il he po en ial i s an icoagulan pu chase. A e he i s pu chase, he s a us was changed in o a use , which was main ained o each yea wi h eco ded pu chases. Men who discon inued he pu chases du ing he ollow-up we e kep as e e -use s. In analysis o wa a in use s compa ed o non-wa a in use s, men we e ca ego ized as wa a in use s each yea wi h eco ded wa a in pu chases, e en i hey had used o he ypes o an icoag- ulan s. Only o yea s wi h eco ded an icoagulan pu chases wi hou wa a in use, we e hey eco ded as non-wa a in use s. Long- e m impac o iming o pos -diagnos ic an i- coagulan use was e alua ed in lag- ime analysis, whe e an icoagulan exposu e was lagged 1 o 3 yea s o wa d om he ac ual yea o usage i.e. i s e ec was igno ed o ha du a ion om he i s exposu e. Compe ing isks analysis using Fine and G ay eg ession me hod wi h non-PCa dea h and majo h omboembolic diseases as he compe ing cause o dea h was pe o med. All s a is ical analyses we e ca ied ou using IBM SPSS S a is ics 22.0. All s a is ical es s a e wo-sided. Resul s Popula ion cha ac e is ics Du ing he median ollow-up o 9 yea s a e PCa diag- nosis 2,296 men died, o whom 728 om PCa (Table 1). The median ollow-up om he diagnosis o PCa dea h was 3.9 yea s among men wi h no pos -diagnos ic an i- coagulan use, 4.8 yea s among wa a in use s and 5.8 yea s in use s o o he an icoagulan s. High-g ade cance s we e almos equally dis ibu ed in hese sub- ca ego ies. Dis ibu ion o backg ound cha ac e is ics a e p esen ed in Table 1. Risk o PCa dea h by p e-diagnos ic use o wa a in and o he an icoagulan s In gene al, when compa ed o an icoagulan non-use s, he e was no clea associa ion be ween p e-diagnos ic use o wa a in and PCa dea h in ei he age-adjus ed o mul i a iable-adjus ed analysis (mul i a iable-adjus ed HR 1.15, 95% CI 0.88-1.49). No s a is ically signi ican isk ends by cumula i e amoun o du a ion o use we e obse ed (Table 2). Howe e , p e-diagnos ic use o wa a in o 5 yea s o mo e was associa ed wi h a bo - de line signi ican isk inc ease in he mul i a iable- adjus ed model (HR 1.49, 95% CI 0.97-2.28). When use s o o he an icoagulan s we e used as he e e ence g oup, wa a in use was no associa ed wi h PCa dea h. No isk ends by cumula i e use we e obse - e ei he (Table 2). Risk o PCa dea h by pos -diagnos ic an icoagulan use In he analysis o pos -diagnos ic use o wa a in, he o e all isk o PCa dea h was signi ican ly highe among wa a in use s in compa ison o an icoagulan non-use s (mul i a iable-adjus ed HR 1.47, 95% CI 1.13-1.93) (Table 3). The isk associa ion was s on- ges in low-dose usage (<200 DDD) (mul i a iable-ad- jus ed HR 2.50, 95% CI 1.86-3.36). Sho - e m (1 yea o less)/low-in ensi y use (<128 DDD/yea ) o wa a in was simila ly associa ed wi h an inc eased isk o PCa dea h. As he cumula i e amoun s in- c eased and he du a ion o usage ex ended o 2 yea s o mo e, he isk inc ease was a enua ed and was no longe s a is ically signi ican . The isk PCa dea h was simila o wa a in and o he an icoagulan s (mul i a iable-adjus ed HR 1.01, 95% CI 0.71-1.44). In he s a i ied analysis, sho - e m use o wa a in was associa ed wi h an inc eased isk o PCa dea h, bu again he isk inc ease was smalle in high- in ensi y use (Table 3). Lag- ime analysis In o de o e alua e long- e m e ec s o wa a in use and minimize bias due o h ombo ic e en s igge ing he p esc ip ion, we pe o med an analysis lagging wa a in use by one, wo and h ee yea s, i.e. ela ing i o e en s a leas 1-3 yea s la e han he exposu e o allow o la ency. The isk inc ease compa ed o an i- coagulan non-use s was dec eased al eady in he ana- lysis lagged by one yea , bu emained non-signi ican ly ele a ed (Table 3). The isk inc ease no ed in low-dose usage was ound in he 1-yea lagged analysis, bu a e 2-yea lag- ime i was only bo de line signi ican ( wo- yea lag- ime o ≤200 DDD HR 1.34, 95% CI 0.99-1.83). In gene al, ex ending he lag- ime o wo o h ee yea s did no subs an ially al e he esul s compa ed o one- yea lag- ime analysis. Lag- ime analysis o wa a in use ela i e o usage o o he an icoagulan d ugs was no subs an ially al e ed compa ed wi h he non-lagged esul s. No isk inc ease among sho - e m use s o dec eased isk among high- in ensi y use s was ound in he lagged analysis (Table 3). Subg oup analyses No clea isk modi ica ion by any backg ound a iable was p esen be ween p e-diagnos ic wa a in use and isk PCa dea h (Fig. 1). The same applied also o pos -diagnos ic use and isk o PCa dea h (Fig. 2). E ec modi ica ion was no ound when compa ing wa a in usage o o he ypes o an icoagulan s in p e- and pos -diagnos ic se ing (Addi ional ile 2: Figu e S1 and Addi ional ile 3: Figu e S2) Kinnunen e al. BMC Cance (2017) 17:585 Page 3 o 9 Table 1 Popula ion cha ac e is is o he s udy popula ion acco ding o p e- and pos -diagnos ic an icoagulan usage s a us P e-diagnos ic s a us Pos -diagnos ic s a us No an icoagula ion Wa a in O he an icoagulan s No an icoagula ion Wa a in O he an icoagulan s n o men in he s udy popula ion 5601 570 366 4485 1074 978 Mean age a diagnosis 67 70 70 67 68 67 PCa dea hs 624 (11.1%) 65 (11.4%) 39 (10.7%) 509 (11.3%) 121 (11.3%) 98 (10.0%) Median age a dea h (yea s) 73 76** 76** 72 75** 74** Median ollow-up om diagnosis o PCa dea h (yea s) 4.5 4.5 4.5 3.9 4.8 5.8 EAU p os a e cance isk-g oup b High-g ade 1661 (29.6%) 181 (31.8%) 109 (29.8%) 1353 (30.2%) 330 (30.7%) 268 (27.4%) Low-/In e media e-g ade 3940 (70.3%) 389 (68.2%) 257 (70.2%) 3132 (69.8%) 744 (69.3%) 710 (72.6%) P ima y he apy a Radical p os a ec omy 1575 (28.1%) 38 (6.7%)** 39 (10.7%)** 1141 (25.4%) 203 (18.9%)** 308 (31.5%)** EBRT 2019 (36.0%) 255 (44.7%) 149 (40.7%) 1647 (36.7%) 444 (41.3%) 332 (33.9%) Ho monal ea men 2197 (39.2%) 303 (53.2%)** 169 (46.2%)** 1814 (40.4%) 490 (45.6%)** 365 (37.3%)** Ac i e su eillance o wa ch ul wai ing 957 (17.1%) 113 (19,8%) 82 (22.4%) 796 (17.7%) 195 (18,2%) 161 (16.5%) Use o o he medica ion S a in use s 2443 (43.6%) 346 (60.7%)** 270 (73.8%)** 1825 (40.7%) 607 (56.5%)** 627 (64.1%)** An i-diabe ic d ug use s 1024 (18.3%) 158 (27.7%)** 92 (25.1%) 780 (17.4%) 277 (25.8%)** 217 (22.2%)** An i-hype ensi e d ug use s 3868 (69.1%) 540 (94.7%)** 340 (92.9%)** 2966 (66.1%) 998 (92.9%)** 784 (80.2%)** NSAID use s 4806 (85.8%) 487 (85.4%) 343 (93.7%)** 3800 (84.7%) 935 (87.1%) 901 (92.1%)** Alpha-blocke use s 2509 (44.8%) 320 (56.1%)** 203 (55.5%)** 2017 (45.0%) 536 (49.9%) 479 (49.0%) Aspi in use s 623 (11.1%) 91 (16.0%) 174 (47.5%)** 393 (8.8%) 151 (14.1%)** 344 (35.2%)** Reco ded diagnoses o : A ial ib illa ion 404 (7.2%) 303 (53.2%)** 27 (7.4%) 129 (2.9%) 571 (53.2%)** 34 (3.5%) Th ombo ic ac o s* 198 (3.5%) 93 (16.3%)** 36 (9.8%) 88 (2.0%) 169 (15.7%)** 70 (7.2%)** Cha lson Como bidi y Sco e 0 3469 (61.9%) 226 (39.6%)** 127 (34.7%)** 2955 (65.9%) 449 (41.8%)** 418 (42.7%)** 1 1093 (19.5%) 147 (25.8%) 113 (30.9%) 822 (18.3%) 272 (25.3%)** 259 (26.5%)** 2 1039 (18.6%) 197 (34.6%)** 126 (34.4%)** 708 (15.8%) 353 (32.9%)** 301 (30.8 %)** ** P o di e ence compa ed o non-use s < 0.001. Calcula ed wi h Chi-squa e es a P ima y ea men modali ies no mu ually exclusi e b Ca ego ized acco ding o c i e ia o he Eu opean Associa ion o U ology (EAU) as o 2015 Kinnunen e al. BMC Cance (2017) 17:585 Page 4 o 9 Sensi i i y analysis As pa o sensi i i y analyses, use s o any an icoagu- lan s including wa a in we e compa ed o non-use s. The e was a bo de line signi ican isk inc ease among p e-diagnos ic use s o an icoagulan s (mul i a iable-ad- jus ed HR 1.19, 95% CI 0.96-1.48) (Addi ional ile 4: Table S2). In analysis o pos -diagnos ic use, he isk was inc eased (HR 1.58, 95% CI 1.27-1.97), bu no appa en isk ends by cumula i e amoun o du a ion we e obse ed. In he lag- ime analyses, no s a is ically signi ican isk inc ease was p esen (Addi ional ile 4: Table S3). No clea e ec modi ica ion was obse ed o p e- no pos -diagnos ic use o all an icoagulan s (Addi ional ile 5: Figu e S3 and Addi ional ile 6: Figu e S4). Compe ing isk analysis In a Fine-G ay eg ession analysis wi h non-PCa dea hs as compe ing cause o dea h, no isk di e ence was obse ed be ween use s o wa a in and o he an icoagu- lan s (HR 0.95, 95% CI 0.74-1.24). The indings we e simila when dea hs caused by pulmona y embolism o s oke we e analyzed as he compe ing causes o dea h (HR 0.95, 95% CI 0.73-1.23). Discussion Sho - e m and low-dose wa a in use a e he diagnosis was associa ed wi h inc eased isk o PCa dea h compa ed o an icoagulan non-use s. Howe e , in analyses allowing o lag- ime, he isk was a enua ed a e one yea and disappea ed in 2-yea o 3-yea lag- ime analysis. This sugges s ha he isk inc ease occu s only o a sho ime pe iod a e s a ing d ug use. We ound no s a is ically signi ican isk di e ence be ween use s o wa a in and o he ypes o an icoagulan s. Addi ionally, we ound no associa ion be ween p e-diagnos ic use o wa a in and PCa su i al. Al hough he indica ion o an icoagulan ea men did no modi y he isk associa ion, i is likely ha he sho - e m inc eased isk is explained by h ombo ic e en s as indica ions o an icoagula ion. In epidemiological li e a u e his phenomenon whe e pha maceu ical d ug is p esc ibed o ea ly mani es a ion o ye un-diagnosed disease is e e ed o as ‘p o opa hic bias’[29]. P e ious s udies on wa a in and PCa mo ali y a e spa se. To da e, h ee published s udies ha e assessed he associa ion be ween PCa su i al and use o i amin K an agonis s [11–13]. Tagalakis e al. ound an inc eased isk o PCa dea h associa ed wi h one-yea o e e -use o wa a in in a e PCa diagnosis [11]. Ou s udy is consis en wi h an inc eased isk associa ed wi h sho - e m use. In con as o Tagalakis e al., we we e able o e alua e cumula i e amoun s and in ensi y o use and ound ha high-dose o long- e m use a e no associa ed wi h isk inc ease. In a coho s udy including 12,186 men wi h PCa, bu no in o ma ion on s age no e idence was ound o an associa ion be ween PCa dea h and p e-diagnos ic use o wa a in du ing a mean ollow-up o 3.7 yea s [12]. In a sensi i i y analysis pos -diagnos ic use was associa ed wi h an inc eased isk o PCa dea h. E alua ion o cumu- la i e amoun s, du a ion no in ensi y o use o pos - Table 2 P e-diagnos ic use o wa a in compa ed o an icoagulan non-use s and wa a in usage in compa ison wi h o he an icoagulan d ugs s a i ied by De ined Daily Doses (DDD), du a ion and in ensi y o usage n o dea hs Age-adjus ed Mul i a iable-adjus ed Wa a in compa ed o non-use s None 624 Re Re Any 65 1.11 (0.85-1.44) 1.15 (0.88-1.49) Amoun o wa a in use ≤200 DDD 24 0.98 (0.65-1.48) 0.97 (0.64-1.47) 201-796 DDD 18 1.09 (0.68-1.74) 1.19 (0.74-1.91) >796 DDD 23 1.31 (0.86-1.99) 1.37 (0.90-2.09) Du a ion o wa a in use ≤1 yea 20 0.91 (0.53-1.56) 0.72 (0.42-1.24) 2-4 yea s 22 0.96 (0.57-1.61) 0.87 (0.52-1.47) 5 o mo e yea s 23 1.33 (0.80-2.23) 1.16 (0.69-1.94) In ensi y o wa a in use ≤114 DDD/yea 21 0.95 (0.56-1.62) 0.79 (0.47-1.35) 115-200 DDD/yea 24 1.09 (0.65-1.81) 0.94 (0.57-1.57) >200 DDD/yea 20 1.10 (0.64-1.89) 0.96 (0.56-1.65) Age-adjus ed and mul i a iable-adjus ed haza d a ios (95% CI) ela ed o all PCa dea hs Kinnunen e al. BMC Cance (2017) 17:585 Page 5 o 9 diagnos ic use was no possible. Conco dan ly, we ob- se ed no isk inc ease o p e-diagnos ic wa a in use. Compa ed o he p e ious s udy, we had highe o PCa dea hs and we e able o e alua e cumula i e amoun o pos -diagnos ic use. We ound a isk inc ease only o some subg oups. A hi d s udy o Pa k e al. [13] was limi ed o 247 pa- ien s wi h me as a ic PCa ecei ing doce axel chemo- he apy, and included only 17 LMWH use s and 12 wa a in use s. LMWH was associa ed wi h an imp o ed su i al, whe eas wa a in was no . In ou la ge and mo e comp ehensi e s udy, we ound no e idence o imp o ed su i al among men using o he ypes o an icoagulan s. In i o s udies ha e sugges ed ha he coagula ion cascade and h ombocy es migh be in ol ed in umo g ow h and p og ession, and ha an icoagulan d ugs could in heo y imp o e p ognosis [6–10]. Howe e , in epidemiologic s udies, con ounding by indica ion dilu es his po en ial e ec , as ad anced cance is associa ed wi h an inc eased isk o h ombosis and consequen ly wi h likelihood o ini ia ing an icoagu- lan ea men . We con olled o such bias by com- pa ing use s o di e en ypes o an icoagulan s and pe o ming a compe ing isk analysis. We ound no signi ican isk di e ence, which does no suppo he pu a i e bene icial e ec o wa a in o any o he an icoagulan s. Table 3 Pos -diagnos ic use o wa a in compa ed o an icoagulan non-use s and sepe a ely o use s o o he ypes o an icoagu- lan s s a i ied by De ined Daily Doses (DDD), du a ion and in ensi y o usage n o dea hs Age-adjus ed Mul i a iable-adjus ed 1-yea lag- ime 2-yea lag- ime 3-yea lag- ime Wa a in compa ed o non-use s None 509 Re Re Re Re Re Any 121 1.46 (1.12-1.90) 1.47 (1.13-1.93) 1.12 (0.85-1.48) 1.12 (0.85-1.48) 1.08 (0.83-1.41) Amoun o wa a in use ≤200 DDD 63 2.47 (1.84-3.31) 2.50 (1.86-3.36) 1.45 (1.04-2.02) 1.34 (0.99-1.83) 1.26 (0.94-1.70) 200-667 DDD 32 0.87 (0.53-1.42) 0.88 (0.54-1.46) 0.85 (0.52-1.41) 0.76 (0.44-1.32) 0.76 (0.43-1.36) >667 DDD 26 0.97 (0.56-1.68) 1.02 (0.59-1.78) 1.05 (0.60-1.83) 1.15 (0.67-1.97) 1.04 (0.58-1.85) Du a ion o wa a in use ≤1 yea 56 2.03 (1.47-2.81) 2.04 (1.47-2.83) 1.26 (0.89-1.78) 1.25 (0.91-1.71) 1.15 (0.85-1.57) 2-4 yea s 44 1.28 (0.88-1.87) 1.30 (0.89-1.90) 1.11 (0.73-1.69) 0.87 (0.53-1.42) 0.93 (0.56-1.54) 5 o mo e yea s 21 0.95 (0.49-1.85) 1.05 (0.54-2.04) 1.06 (0.56-1.99) 1.37 (0.76-2.47) 1.22 (0.66-2.26) In ensi y o wa a in use ≤128 DDD/yea 45 1.92 (1.34-2.74) 1.91 (1.34-2.74) 1.31 (0.91-1.88) 1.35 (0.98-1.86) 1.21 (0.89-1.66) 128-200 DDD/yea 51 1.74 (1.24-2.45) 1.77 (1.25-2.50) 1.13 (0.74-1.74) 0.93 (0.57-1.51) 0.94 (0.57-1.56) >200 DDD/yea 25 0.73 (0.40-1.33) 0.78 (0.43-1.43) 0.99 (0.58-1.69) 0.97 (0.56-1.69) 0.99 (0.56-1.77) Wa a in compa ed o o he an icoagulan d ugs Non-wa a in an icoagulan use s 98 Re Re Re Re Re Wa a in use s 121 1.13 (0.79-1.61) 1.01 (0.71-1.44) 0.93 (0.64-1.35) 1.02 (0.70-1.48) 0.93 (0.65-1.33) Amoun o wa a in use ≤200 DDD 63 1.85 (1.25-2.74) 1.63 (1.10-2.42) 1.15 (0.75-1.77) 1.19 (0.78-1.80) 1.06 (0.71-1.58) 200-667 DDD 32 0.65 (0.37-1.14) 0.58 (0.33-1.01) 0.68 (0.38-1.20) 0.67 (0.36-1.25) 0.64 (0.34-1.21) >667 DDD 26 0.72 (0.39-1.33) 0.67 (0.36-1.23) 0.83 (0.45-1.55) 1.02 (0.56-1.87) 0.87 (0.46-1.64) Du a ion o wa a in use ≤1 yea 56 1.52 (1.00-2.31) 1.33 (0.88-2.02) 1.00 (0.64-1.55) 1.11 (0.73-1.68) 0.97 (0.65-1.45) 2-4 yea s 44 0.96 (0.61-1.52) 0.85 (0.54-1.34) 0.88 (0.53-1.46) 0.77 (0.44-1.35) 0.78 (0.44-1.38) 5 o mo e yea s 21 0.71 (0.35-1.46) 0.69 (0.34-1.40) 0.84 (0.42-1.67) 1.22 (0.64-2.33) 1.02 (0.53-2.00) In ensi y o wa a in use ≤128 DDD/yea 45 1.44 (0.93-2.24) 1.25 (0.80-1.95) 1.04 (0.66-1.64) 1.20 (0.78-1.83) 1.02 (0.68-1.53) 128-200 DDD/yea 51 1.31 (0.85-2.01) 1.16 (0.75-1.78) 0.90 (0.54-1.49) 0.82 (0.47-1.44) 0.79 (0.45-1.40) >200 DDD/yea 25 0.55 (0.29-1.05) 0.51 (0.27-0.98) 0.79 (0.43-1.44) 0.86 (0.46-1.60) 0.83 (0.44-1.57) Age-adjus ed, mul i a iable-adjus ed and lag- ime haza d a ios (95% CI) ela ed o all PCa dea hs. S a is ically signi ican esul s a e bolded Kinnunen e al. BMC Cance (2017) 17:585 Page 6 o 9 Themains eng hso hiss udya e hela ge popula ion-based coho as well as comp ehensi e and ou abili y o use de ailed, na ionally comp ehen- si e egis e -based in o ma ion on an icoagulan use una ec ed by ecall bias. Being able o s udy la ge popula ions h ough na ional egis ies allows us o es ima e e en ela i ely uncommonly used d ugs such as an icoagulan s as cance isk ac o . Ou s udy in- cluded subs an ially mo e wa a in use s han in p e- ious s udies. We we e able o analyze pos -diagnos ic use o wa a in mo e comp ehensi ely han be o e. We also pe o med lag- ime analyses o es ima e he isk associa ions allowing o la ency and emo ing e ec s limi ed o he immedia e pe iod ollowing i s subsc ip ion. Fu he mo e, we had longe ollow-up han in p e ious s udies. Ou s udy also has some limi a ions. We we e no able o adjus ou analysis o smoking, li e-s yle ac o s o BMI which may be associa ed wi h isk o PCa dea h [30–33]. Ne e heless, we we e able o adjus o Cha l- son Como bidi y Index and his enabled adjus men o co-exis ing mo bidi ies. Fu he mo e, ou s udy was no andomized, and is hus p one o esidual con ounding. Conclusion In a popula ion-based se ing, wa a in o o he ypes o an icoagulan s a e no associa ed wi h imp o ed PCa p ognosis. Con e sely, in sho - e m use isk o PCa dea h was inc eased, which is mos likely due o h om- bosis caused by an ad anced cance , as he isk inc ease was no obse ed in long e m. Fig. 1 P e-diagnos ic subg oup analysis o wa a in usage compa ed o an icoagulan non-use s Fig. 2 Pos -diagnos ic subg oup analysis o wa a in usage compa ed o an icoagulan non-use s Kinnunen e al. BMC Cance (2017) 17:585 Page 7 o 9 Addi ional iles Addi ional ile 1: Table S1. ATC codes o an icoagulan d ugs included in he s udy. (DOCX 207 kb) Addi ional ile 2: Figu e S1. P e-diagnos ic subg oup analysis be ween use s o wa a in and o he an icoagulan d ugs. (DOCX 224 kb) Addi ional ile 3: Figu e S2. Pos -diagnos ic subg oup analysis be ween use s o wa a in and use s o o he an icoagulan d ugs. (DOCX 224 kb) Addi ional ile 4: Table S2. P e-diagnos ic analysis o combined an icoagulan d ug usage s a i ied by numbe o De ined Daily Doses (DDD), du a ion and in ensi y o usage. Age-adjus ed and mul i a iable-adjus ed haza d a ios (95% CI) ela ed o all PCa dea hs. Table S3. Pos -diagnos ic esul s o combined an icoagulan usage. Age-adjus ed, mul i a iable-adjus ed and lag- ime haza d a ios (95% CI) ela ed o all PCa dea hs. (DOCX 405 kb) Addi ional ile 5: Figu e S3. Subg oup analysis o p e-diagnos ic combined an icoagulan usage compa ed o non-use s. (DOCX 224 kb) Addi ional ile 6: Figu e S4. Subg oup analysis o combined an icoagulan usage compa ed o non-use s in pos -diagnos ic se ing. (DOCX 224 kb) Abb e ia ions VTE: Venous h omboembolism; PCa: P os a e cance ; FinRSPC: he Finnish Randomized S udy o Sc eening o P os a e Cance ; EAU: he Eu opean Associa ion o U ology; SII: (Finnish) Social Insu ance Ins i u ion; EBRT: ex e nal beam adia ion he apy; HR: Haza d a io; CI: Con idence in e al; DDD: De ined Daily Dose Acknowledgemen s No applicable. Funding Suppo ed by non- es ic ed compe i i e esea ch unding om he Pi kanmaa Hospi al Dis ic 150640 o TJ Mu ola and non- es ic ed esea ch g an om Pi kanmaa Cance Socie y o PTT Kinnunen. Academy o Finland (g an 132385 and 260 931) and Cance O ganisa ions o Finland o A Au inen, Compe i i e Resea ch unding o Pi kanmaa Hospi al Dis ic o TLJ Tammela. No Funde o ganiza ion pa icipa ed in any pa o he s udy design o collec ion, analysis o in e p e a ion o da a o w i ing he manusc ip . A ailabili y o da a and ma e ials Pe mission o use he en i e da a has been g an ed pe sonally o his speci ic s udy, hus we a e no allowed o make he en i e da a publicly a ailable wi hou pe mission. Limi ed da a (i.e a iables used o he analyses) can be ob ained upon easonable eques . Each Adminis a o and In o ma ion Commissione o he used egis ies p ocessed ou eques o use he da a and g an ed a pe sonal pe mission. I desi ed, simila pe missions can be applied om he Finnish Social Insu ance Ins i u ion (a ailable a : h p://www.kela. i/web/en/ esea ch) and he Na ional Ins i u e o Heal h and Wel a e (a ailable a : h ps://www. hl. i/en/web/ hl i- en/s a is ics/in o ma ion- o - esea che s). Au ho s' con ibu ions S udy Concep : KPTT, MTJ, AA. S udy Design: KPTT, MTJ, AA. Da a Acquisi ion: KPTT, MTJ, TK, TK, TTLJ, AA. Quali y Con ol o Da a: KPTT, MTJ, TK, TK, TTLJ, AA. All au ho s ha e ead and app o ed he manusc ip . E hics app o al and consen o pa icipa e The s udy has been app o ed by he E hical Commi ee o Pi kanmaa Hospi al Dis ic (Commi ee’s e e ence numbe R10167). This s udy is based on egis e da a collec ed ou inely o o he pu poses. Thus no in o med consen is needed based on in e na ional p ac ices. Consen o publica ion No applicable. Compe ing in e es s We decla e he ollowing compe ing in e es s: PTT Kinnunen: none, TJ Mu ola: Paid consul an o As ellas and Janssen- Cilag, lec u e ees om As ellas, Janssen-Cilag, Abb ie and MSD, K Talala: none, K Taa i: Consul ing ee om Abb ie, esea ch unding om Medi a ion, a el suppo om As ellas, and O ion, TLJ Tammela: Paid consul an o As ellas, , O ion Pha ma and Jansen-Cilag, A Au inen: lec u e ee om MSD, paid consul an o Epid Resea ch Inc. Publishe ’sNo e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a ilia ions. Au ho de ails 1 Uni e si y o Tampe e, Facul y o Medicine and Li e Sciences, Tampe e, Finland. 2 Depa men o U ology, Tampe e Uni e si y Hospi al, Tampe e, Finland. 3 Finnish Cance Regis y, Helsinki, Finland. 4 Depa men o U ology, Uni e si y o Helsinki and Helsinki Uni e si y Hospi al, Helsinki, Finland. 5 Uni e si y o Tampe e, Facul y o Social Sciences, Tampe e, Finland. Recei ed: 10 Ap il 2017 Accep ed: 22 Augus 2017 Re e ences 1. Lyman GH. Venous h omboembolism in he pa ien wi h cance : ocus on bu den o disease and bene i s o h ombop ophylaxis. Cance . 2011 Ap 1;117(7):1334–49. 2. P andoni P, Lensing AW, P ins MR. The na u al his o y o deep- ein h ombosis. Semin Th omb Hemos . 1997;23(2):185–8. 3. Sø ensen HT, Mellemkjae L, Olsen JH, Ba on JA. P ognosis o cance s associa ed wi h enous h omboembolism. N Engl J Med. 2000 Dec 21;343(25):1846–50. 4. Van Hemel ijck M, Adol sson J, Ga mo H, Bill-Axelson A, B a O, Ingelsson E, e al. Risk o h omboembolic diseases in men wi h p os a e cance : esul s om he popula ion-based PCBaSe Sweden. Lance Oncol. 2010 May;11(5):450–8. 5. Cha u edi S, Sidana S, Elson P, Kho ana AA, McC ae KR. Symp oma ic and inciden al enous h omboembolic disease a e bo h associa ed wi h mo ali y in pa ien s wi h p os a e cance . PLoS ONE. 2014;9(8):e94048. 6. Maa B. Selec i e mac ophage inhibi ion abolishes wa a in-induced educ ion o me as asis. B J Cance . 1980 Feb;41(2):313–6. 7. B own JM. A s udy o he mechanism by which an icoagula ion wi h wa a in inhibi s blood-bo ne me as ases. Cance Res. 1973 Jun;33(6):1217–24. 8. Bobek V, Boubelik M, Ko a ik J, Tal yno O. Inhibi ion o adhesion b eas cance cells by an icoagulan d ugs and cime idine. Neoplasma. 2003;50(2):148–51. 9. McCulloch P, Geo ge WD. Wa a in inhibi ion o me as asis: he ole o an icoagula ion. B J Su g. 1987 Oc ;74(10):879–83. 10. Mousa SA. An icoagulan s in h ombosis and cance : he missing link. Expe Re An icance The . 2002 Ap ;2(2):227–33. 11. Tagalakis V, Tamim H, Blos ein M, Hanley JA, Kahn SR. Risk o p os a e cance dea h in long- e m use s o wa a in: a popula ion-based case- con ol s udy. Cance Causes Con ol. 2013 Jun;24(6):1079–85. 12. O'Ro ke MA, Mu ay LJ, Hughes CM, Can well MM, Ca dwell CR. The e ec o wa a in he apy on b eas , colo ec al, lung, and p os a e cance su i al: a popula ion-based coho s udy using he Clinical P ac ice Resea ch Da alink. Cance Causes Con ol. 2015 Ma ;26(3):355–66. 13. Pa k JC, P a z CF, Tes aye A, B odsky RA, An ona akis ES. The e ec o he apeu ic an icoagula ion on o e all su i al in men ecei ing i s -line doce axel chemo he apy o me as a ic cas a ion- esis an p os a e cance . Clin Geni ou in Cance . 2015 Feb;13(1):32–8. 14. Ahlb ech J, Dickmann B, Ay C, Dunkle D, Thale J, Schmidinge M, e al. Tumo g ade is associa ed wi h enous h omboembolism in pa ien s wi h cance : esul s om he Vienna Cance and Th ombosis S udy. J Clin Oncol. 2012 No 1;30(31):3870–5. 15. Kilpeläinen TP, Tammela TL, Malila N, Hakama M, San i H, Mää änen L, e al. P os a e cance mo ali y in he Finnish andomized sc eening ial. J Na l Cance Ins . 2013 May 15;105(10):719–25. Kinnunen e al. BMC Cance (2017) 17:585 Page 8 o 9 16. Teppo L, Pukkala E, Leh onen M. Da a quali y and quali y con ol o a popula ion-based cance egis y. Expe ience in Finland. Ac a Oncol. 1994;33(4):365–9. 17. EAU isk g oup s a i ica ion, Table 4.2.2. A ailable a : h ps://u oweb.o g/ wp-con en /uploads/EAU-Guidelines-P os a e-Cance -2016.pd 18. The S a is ical Finland. A ailable a h p://www.s a . i/index_en.h ml 19. Ma ikainen J, Rajaniemi S. D ug eimbu semen sys ems in EU Membe S a es, Iceland and No way. Helsinki: The Social Insu ance Ins i u ion, Finland, Social secu i y and heal h epo s. 2002;54 A ailable a : h ps://helda.helsinki. i/handle/10138/13 20. Flaha an EM, Benne K, Sha p L, Ba on TI. A coho s udy in es iga ing aspi in use and su i al in men wi h p os a e cance . Ann Oncol. 2014 Jan;25(1):154–9. 21. Holmes S, G i i h EJ, Mus o G, Minuk GY. An ihype ensi e medica ions and su i al in pa ien s wi h cance : a popula ion-based e ospec i e coho s udy. Cance epidemiol. 2013 Dec;37(6):881–5. 22. Tan P, Wei S, Yang L, Tang Z, Cao D, Liu L, e al. The e ec o s a ins on p os a e cance ecu ence and mo ali y a e de ini i e he apy: a sys ema ic e iew and me a-analysis. Sci ep. 2016;6:29106. 23. Vei onmäki T, Mu ola TJ, Mää änen L, Taa i K, S enman U, Tammela TLJ, e al. Use o non-s e oidal an i-in lamma o y d ugs and p os a e cance su i al in he Finnish p os a e cance sc eening ial. P os a e. 2015 Sep;75(13):1394–402. 24. Kjellman A, F iis S, G ana h F, Gus a sson O, So ensen HT, Ak e O. T ea men wi h inas e ide and p os a e cance su i al. Scand J U ol. 2013 Aug;47(4):265–71. 25. Randazzo M, Bea ice J, Hube A, G obholz R, Manka L, Wyle SF, e al. In luence o me o min use on PSA alues, ee- o- o al PSA, p os a e cance incidence and g ade and o e all su i al in a p ospec i e sc eening ial (ERSPC Aa au). Wo ld J U ol. 2015 Aug;33(8):1189–96. 26. The Na ional Ca e Regis e o Heal h Ca e main ained by he Na ional Ins i u e o Heal h and Wel a e. In o ma ion o easea che s a ailable a : h ps://www. hl. i/en/web/ hl i-en/s a is ics/in o ma ion- o - esea che s 27. Cha lson ME, Pompei P, Ales KL, MacKenzie CR. A new me hod o classi ying p ognos ic como bidi y in longi udinal s udies: de elopmen and alida ion. J Ch onic Dis. 1987;40(5):373–83. 28. WHO ATC/DDD index 2016. A ailable a : h p://www.whocc.no/a c_ddd_ index/. 29. Ho wi z RI, Feins ein AR. The p oblem o "p o opa hic bias" in case-con ol s udies. Am J Med. 1980;68(2):255–8. 30. Islami F, Mo ei a DM, Bo e a P, F eedland SJ. A sys ema ic e iew and me a-analysis o obacco use and p os a e cance mo ali y and incidence in p ospec i e coho s udies. Eu U ol. 2014 Dec;66(6):1054–64. 31. Dicke man BA, Ma k SC, Kosken uo M, Pukkala E, Mucci LA, Kap io J. Alcohol in ake, d inking pa e ns, and p os a e cance isk and mo ali y: a 30-yea p ospec i e coho s udy o Finnish wins. Cance Causes Con ol. 2016 Sep;27(9):1049–58. 32. Da ies NM, Gaun TR, Lewis SJ, Holly J, Dono an JL, Hamdy FC, e al. The e ec s o heigh and BMI on p os a e cance incidence and mo ali y: a Mendelian andomiza ion s udy in 20,848 cases and 20,214 con ols om he PRACTICAL conso ium. Cance Causes Con ol. 2015 No ;26(11):1603–16. 33. Fowke JH, McLe an DF, Gup a PC, He J, Shu X, Ramadas K, e al. Associa ions o body mass index, smoking, and alcohol consump ion wi h p os a e cance mo ali y in he Asia Coho Conso ium. Am J Epidemiol. 2015 Sep 1;182(5):381–3. • We accep p e-submission inqui ies • Ou selec o ool helps you o ind he mos ele an jou nal • We p o ide ound he clock cus ome suppo • Con enien online submission • Tho ough pee e iew • Inclusion in PubMed and all majo indexing se ices • Maximum isibili y o you esea ch Submi you manusc ip a www.biomedcen al.com/submi Submi you nex manusc ip o BioMed Cen al and we will help you a e e y s ep: Kinnunen e al. BMC Cance (2017) 17:585 Page 9 o 9