Cancer incidence among Finnish male cobalt production workers in 1969-2013: a cohort study

RESEARCH ARTICLE Open Access Cance incidence among Finnish male cobal p oduc ion wo ke s in 1969–2013: a coho s udy Rii a Sauni 1* , Panu Oksa 2,3 , Jukka Ui i 2,3,4 , Asko Linna 5 , Raimo Ke ula 6 and Ee o Pukkala 7,8 Abs ac Backg ound: The e is inadequa e e idence o he ca cinogenici y o cobal and cobal compounds in humans. Consequen ly, he In e na ional Agency o Resea ch on Cance (IARC) has e alua ed cobal me al wi hou ungs en ca bide as possibly ca cinogenic o humans (G oup 2B). The aim o he s udy was o assess he isk o cance among wo ke s employed in a Finnish cobal plan since he beginning o p oduc ion in 1968. Me hods: The s udy coho consis ed o all males employed by he Finnish cobal plan o a leas a yea du ing 1968–2004. The ollow-up o cance was pe o med by s udying he iles o he Finnish Cance Regis y, using pe sonal iden i y codes as a key. The coho was di ided in o subcoho s by exposu e le els. S anda dised incidence a ios (SIRs) and 95% con idence in e als (95% CIs) we e calcula ed as a ios o he obse ed numbe s o cance cases and he numbe s expec ed on he basis o incidence a es in he popula ion o he same egion. Resul s: The ollow-up coho consis ed o 995 men wi h 26,083 pe son-yea s. Du ing he ollow-up pe iod, 92 cases o cance we e diagnosed (SIR 1.00, 95% CI 0.81–1.22), six o which we e lung cance cases (SIR 0.50; 95% CI 0.18–1.08). The only cance ype wi h inc eased incidence was ongue cance ( h ee cases, SIR 7.39; 95% CI 1.52–21.6). We obse ed no dose- esponse e ec ac oss he di e en exposu e le els and he incidence o any cance ype. Conclusions: The esul s sugges ha occupa ional exposu e o cobal is no associa ed wi h an inc eased o e all cance isk o lung cance isk among cobal wo ke s. Because o he small numbe o cance cases he esul s mus be in e p e ed wi h cau ion. Keywo ds: Cobal , Occupa ional exposu e, Incidence, Cance Backg ound Wo ke s may be exposed o cobal du ing he p oduc ion o cobal and cobal sal s, in he p oduc ion o alloys and ha d me al, d ying agen s, pigmen s and ca alys s, and du ing diamond polishing. Wo ld p oduc ion o e ined cobal has inc eased s eadily o e he las decade, due pa ly o new ope a ions and pa ly o a ne inc ease in p oduc ion by es ablished p oduce s [1]. Wo ld cobal mine p oduc ion in 2001 was 36,700 ons and he cobal e ine y p oduc ion 38,400 ons [2]. The espec i e numbe s in 2016 we e 123,000 ons and 91,300 ons [3]. In he Uni ed S a es, mo e han a million wo ke s a e po en ially exposed o cobal and i s compounds [4]. In Finland, abou 1500 wo ke s (0.1% o employed people) a e es ima ed o be exposed o cobal o cobal compounds a wo k [5]. Di e en epidemiological s udies p o ide ela i ely spa se and con adic o y da a on he ca cinogenici y o cobal o humans. A Swedish s udy e ospec i ely ollowed a coho o 3000 cobal wo ke s in 1951–1982 [6]. I ound a non-signi ican inc ease in mo ali y om lung cance in he whole coho (SMR 1.34, 95% CI 0.77–2.13). A signi ican excess mo ali y om lung cance was ound among wo ke s wi h o e en yea s o employmen who had died mo e han 20 yea s a e he end o exposu e (SMR 2.78, 95% CI 1.11–5.72). Ea ly indings om a F ench s udy in he 1980s sugges ed inc eased mo ali y om cance s o he achea, lung * Co espondence: [email p o ec ed] 1 Depa men o Occupa ional Sa e y and Heal h, Minis y o Social A ai s and Heal h, P.O. Box 33FI-00023 Go e nmen 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. Sauni e al. BMC Cance (2017) 17:340 DOI 10.1186/s12885-017-3333-2 and b onchus among cobal wo ke s (SMR 4.66; 95% CI 1.46–10.64) [7]. Howe e , he esul s we e based on only ou cases. In hei ollow-up s udy o he same popula- ion, Moulin e al. [8] could no con i m he p e ious esul s. A No wegian s udy in a nickel e ine y did no ind any inc ease in isk o lung cance om cobal exposu e [9]. The esul s o he s udies on he ca cinogenici y o co- bal a y acco ding o he kind o indus y in which he exposed wo ke s a e employed. Inc eased mo ali y om lung cance has been ound among wo ke s in he ha d me al indus y [10]. In addi ion o cobal , ha d me al also con ains ungs en ca bide. Acco ding o he In e na ional Agency o Resea ch on Cance (IARC) he ca cinogen- ici y o cobal me al wi h ungs en ca bide was e alua ed as p obably ca cinogenic o humans (G oup 2A), whe eas cobal me al wi hou ungs en ca bide was e alua ed as possibly ca cinogenic o humans (G oup 2B). Cobal sulpha e and o he soluble cobal (II) sal s we e also e alu- a ed as possibly ca cinogenic o humans (G oup 2B) [1]. In ea ly 2014, he Ne he lands Compe en Au ho i y (RIVM) no i ied i s in en ion ia he Regis y o In en o p epa e an EU-ha monised classi ica ion and labelling p oposal on ‘cobal me al and o he cobal compounds ( o be de e mined)’. Eu opean Commission is ac i ely igh ing agains occupa ional cance and aiming o amend he Ca cinogens and Mu agens Di ec i e [11]. I is impo an o ha e up- o-da e scien i ic e idence based da a as a backg ound o legisla ion. The pu pose o his s udy was o assess he isk o cance among wo ke s employed in a Finnish cobal plan , using he company’s employmen eco ds, expos- u e da a, and da a om he Finnish Cance Regis y. On he basis o his eliable da a, we aimed o add o he knowledge on he ca cinogenici y o cobal wi hou ungs en ca bide in occupa ional se ings. Me hods This is a e ospec i e coho s udy. The s udy coho was made up o all males employed o a leas one yea a he Kokkola cobal plan (F eepo Cobal Oy) du ing he pe iod 1968–2004. The coho o 1004 men was iden i ied om he company’s employmen eco ds. The co ec pe sonal iden i y codes (PICs), i al s a us and possible da es o emig a ion o dea h we e sea ched om he na ional Popula ion Regis e Cen e . Nine men (0.9%) we e no ound in he popula ion egis e and we e ex- cluded, lea ing 995 wo ke s in he inal coho (Table 1). Since 1967, all Finnish esiden s ha e had a unique PIC, which is used in all main egis e s in Finland. The PIC en- ables eliable au oma ic eco d linkage. We used PICs as keys in he ollow-up o cance h ough he iles o he popula ion-based na ionwide Finnish Cance Regis y. Follow-up began on he da e when he pe son had been wo king o one yea a he Kokkola cobal plan , and ended a emig a ion, dea h o on 31 Decembe 2013, whiche e came i s . We also calcula ed he cance isk as s a ing om he da e when a pe son had wo ked o i e yea s a he cobal plan . The numbe s o obse ed cases and pe son-yea s a isk we e coun ed, in i e-yea age g oups and i e-yea calenda pe iods. The expec ed numbe s o cases o o al cance and o speci ic cance ypes we e calcula ed by mul iplying he numbe o pe son-yea s in each s a um by he co esponding cance incidence among men in he cen al hospi al ca chmen a ea o Cen al Os obo hnia a ound he Kokkola cobal plan . Table 1 Numbe o male wo ke s (N) in cobal plan coho , and pe son-yea s du ing 1969–2013, by age and exposu e g oup ( o de ini ions, see Table 2). The numbe s in he N column in di e en age g oups e e o he age a he beginning o ollow-up N Pe son-yea s To al 995 26,083.2 High exposu e 380 11,254.9 Age (yea s) 15–29 269 1480.0 30–44 103 4243.6 45–59 7 3663.5 60–74 1 1736.5 75+ - 131.6 Mode a e exposu e 159 2823.4 Age (yea s) 15–29 122 563.4 30–44 35 1378.4 45–59 2 674.8 60–74 - 205.8 75+ - 1.0 Low exposu e 364 9966.7 Age (yea s) 15–29 275 1431.2 30–44 81 3903.3 45–59 8 2987.7 60–74 - 1522.6 75+ - 121.9 Va iable exposu e wi h peak exposu es 110 2174.1 Age (yea s) 15–29 94 475.9 30–44 16 960.6 45–59 - 551.2 60–74 - 182.8 75+ - 3.7 Sauni e al. BMC Cance (2017) 17:340 Page 2 o 8 To calcula e he s anda dised incidence a io (SIR), he obse ed numbe o cases was di ided by he expec ed numbe . The 95% con idence in e al (95% CI) o he SIR was based on he assump ion ha he numbe o obse ed cases ollowed a Poisson dis ibu ion. A non-comme cial so wa e p og am o Finnish Cance Regis y was used o s a is ical analysis. We call an SIR as “s a is ically signi ican ”i i s 95% con idence in e al does no include alue 1.0. Wo kplace exposu es Be ween 1966 and 1987, cobal powde was p oduced om py i e o e concen a e a he Kokkola cobal plan . A e 1987, cobal powde , ino ganic cobal , and nickel compounds ha e been p oduced using by-p oduc s o he me allu gic indus y as aw ma e ial (Fig. 1). In sulpha ising oas ing, dus in he ambien ai was ound o con ain 15–20% i on, 1% zinc, 0.4% cobal , and 0.2% nickel, whe eas in leaching building, he dus consis ed o me al sulphides and sulpha es. The highes exposu e le els o nickel (0.12 mg/m 3 ) we e measu ed in he chemical depa men du ing 1987–1999, o he wise exposu e le els ha e been ≤0.04 mg/m 3 . Cobal and nickel we e p esen as wa e -soluble sulpha es. In he e- duc ion plan and powde p oduc ion acili y, cobal is mainly in he o m o cobal powde and ine powde . In he chemical depa men , he cobal and nickel com- pounds we e mainly sulpha es, ca bona es, oxides, and hyd oxides. To al exposu e o dus , cobal , nickel, sulphu dioxide, hyd ogen sulphide, and ammonia has been egula ly moni o ed se e al imes a yea since 1966, as bo h s a- iona y measu emen s and pe sonal samples. The mean exposu e le el o o al dus was high in he sulpha ising oas ing depa men , a 8.5 mg/m 3 . The mean le els o cobal in he wo kplace ai in 1967–2003 a e p esen ed in Fig. 2. The me hods o measu ing wo kplace expo- su es ha e been desc ibed in de ail ea lie [12, 13]. Acco ding o biological moni o ing su eillance, he highes exposu e o cobal was in he educ ion and powde p oduc ion depa men . The highes u ina y con en o cobal was abou 16,000 nmol/l (le el o unexposed pe sons being <40 nmol/l). In he solu ion pu i ica ion and chemical depa men s, he u ina y cobal le els we e be ween 200 and 2000 nmol/l. Respi a- o s we e a ailable since he plan s a ed ope a ing, and became manda o y in he las en yea s in he powde p oduc ion and chemical depa men s. The biological moni o ing esul s show ha exposu e is s ill conside able, despi e he in ensi ied use o espi a o s [12, 13]. Exposu e g oups The coho was di ided in o subcoho s by exposu e le els, acco ding o he depa men in which hey had s a ed wo king du ing hei employmen a he plan (Tables 1, 2). The exposu e in di e en depa men s was classi ied acco ding o he indus ial hygienic measu e- men s (Fig. 2) and biological moni o ing. Exposu e in ac o y main enance was classi ied as a iable, because i includes e.g. epai s in di e en depa men s wi h possible peak exposu es when he machine y is ou o o de . Resul s The ollow-up coho consis ed o 995 men wi h 26,083 pe son-yea s (Table 1). The mean ollow-up o a pe son was hus 26.2 yea s. Cobal p ocess 1966 – 1987 Cobal p ocess 1987 – p esen Py i e o e concen a ion Cobal concen a es I on esidue I on esidue Coppe sulphide Zinc sulphide Coppe sulphide Zinc sulphide Nickel ammonium sulpha e Cobal and nickel solu ions Cobal powde Cobal ine powde Sulpha ising oas ing Leaching and solu ion pu i ica ion Reduc ion and powde p oduc ion Leaching and solu ion pu i ica ion Cobal ex ac ion Chemical p oduc ion Reduc ion and powde p oduc ion Cobal powde Cobal ine powde Cobal and nickel sulpha es, oxides, h y d oxides Fig. 1 Cobal p oduc ion p ocess in he Kokkola cobal plan in 1966 –1987 and 1987 –p esen Sauni e al. BMC Cance (2017) 17:340 Page 3 o 8 Du ing he ollow-up pe iod, 92 cases o cance we e diagnosed, and he expec ed numbe was 91.9 (SIR 1.00, 95% CI 0.81–1.22; Table 3). Fi e cases we e obse ed 5–15 yea s a e he beginning o ollow-up and 87 cases mo e han 15 yea s a e he beginning o ollow-up. A o al o 77 cases o cance we e diagnosed among men who had been wo king in he plan o o e i e yea s (SIR 1.08, 95% CI 0.85–1.34) (Table 3). The o e all cance incidence was no signi ican ly ele a ed in any o he exposu e g oups (Table 4). The SIR o lung cance among men wi h a leas one yea o employmen was 0.50 (95% CI 0.18–1.08) and among men wi h a leas i e yea s o employmen 0.52 (95%CI 0.17–1.22) (Table 3). None o he exposu e g oup-speci ic SIRs o lung cance was signi ican ly di - e en om 1.0 (Table 4). Th ee o he lung cance cases we e in he age g oup 45–59 yea s and h ee in he age g oup 60–74 yea s. The age speci ic SIRs o lung cance did no di e signi ican ly om hose o he e e ence popula ion (age g oup 45–59 yea s: SIR 0.75, 95% CI 0.15–2.18 and age g oup 60–74 yea s: SIR 0.42, 95% CI 0.09–1.24). The incidence o ongue cance was signi ican ly in- c eased (SIR 7.39; 95% CI 1.52–21.6) (Table 3). The age o he pa ien s a he ime o diagnosis was 39, 52 and 70 yea s. The age g oup speci ic SIRs o ongue cance did no di e signi ican ly om hose o he e e ence popula ion (age g oup 30–44 yea s: SIR 16.35, 95% CI 0.41–91.08; age g oup 45–59 yea s: SIR 8.02, 95% CI 0.20–44.66 and age g oup 60–74 yea s: SIR 16.23, 95% CI 0.41–90.40). All h ee cases we e smoke s. One o hem had wo ked o 20.5 yea s in he cobal plan ( a iable exposu e g oup), one had wo ked he e o 36.1 yea s (low exposu e g oup) and he hi d o 15.8 yea s (high exposu e g oup). The e was an excess o bladde cance cases in he low exposu e g oup (SIR 3.07; 95% CI 1.12–6.67) (Table 4). One bladde cance pa ien was a non-smoke , ou we e smoke s and he smoking s a us o ou pa ien s was unknown. The SIRs o s omach cance , la ynx cance and hy oid cance exceeded 2.0 bu we e based on small numbe s o cases and we e no s a is ically signi ican (Table 3). Discussion Ea ly indings om a F ench s udy in he 1980s sug- ges ed ha cobal wo ke s may be a an inc eased isk o cance s o he achea, lung and b onchus [7]. La e indings indica ed ha occupa ional exposu e o cobal in he ambien ai does no inc ease ei he he o al cance isk o he isk o lung cance [8]. The esul s o he p esen s udy a e in line wi h he no-inc ease indings. In he p esen s udy o a coho o 995 men wo king in a cobal plan , wi h a mean ollow-up o 26.2 yea s, we obse ed no ele a ed o e all cance incidence o ele a ed incidence o lung cance . Ou coho consis ed o all employees who had been wo king in he in eg a ed p oduc ion uni o he Kokkola cobal plan du ing 1969–2004. Iden i ica ion o coho membe s and ollow-up o dea hs and emig a- ion we e comple e o he pe iod o his s udy. The comple eness o cance egis a ion in Finland is a leas 99% [14], and he compu e ised eco d linkage p oced- u e p ecise [15]. The e o e, echnical incomple eness does no cause bias in he esul s. Exposu e assessmen The mean le els o cobal in he ambien ai in his plan we e gene ally a he le el o he cu en occupa ional exposu e limi in Finland (0.02 mg/m 3 ). Du ing he i s yea s o cobal p oduc ion, he cobal le els may ha e been conside ably highe , o e 1 mg/m 3 , especially in he oas ing depa men . The Finnish occupa ional limi s we e also o en exceeded in he educ ion and powde p oduc ion depa men s. Cobal exposu e was moni o ed wi h excep ional accu acy om he ea ly days o he s udied plan . The wo king his o y o he pa ien s was e i ied om he egis e s, which minimises ecall and in o ma ion bias. This was no he case in p e ious epidemiological s udies, which did no epo ei he he exposu e le els o cobal o simul aneous co-exposu es [7]. O he o he simul aneous co-exposu es, nickel is con- side ed as ca cinogenic. The e has been nickel exposu e Fig. 2 Mean cobal exposu e le els a Kokkola cobal plan in 1968–2014 Table 2 Exposu e g oups acco ding o depa men s De ini ion Depa men s Va iable exposu e wi h peak exposu es Fac o y main enance Low exposu e Leaching and solu ion pu i ica ion Mode a e exposu e Chemical depa men , es plan High exposu e Sulpha ising oas ing, educ ion and powde p oduc ion Sauni e al. BMC Cance (2017) 17:340 Page 4 o 8 in leaching, solu ion, pu i ica ion and chemical depa - men s. In a ecen s udy on cance isk in a Finnish nickel e ine y he inc eased lung cance and sinonasal cance isk was ound in he mos nickel-exposed wo k si e whe e he nickel concen a ion had been ≥0.2 mg/m 3 [16]. The nickel exposu e le els ha e been much lowe in he cobal plan han in he nickel e ine y, and in he p esen s udy he e we e no sinonasal cance s and he isk o lung cance was low. SIR The p e ious epidemiological s udies on he ca cinogen- ici y o cobal used he s anda dised mo ali y a io (SMR) as a measu e o cance isk, compa ing he numbe o dea hs in he coho wi h he expec ed num- be o dea hs, calcula ed om he mo ali y a es o he gene al popula ion [7, 8, 10]. As he majo i y o cance pa ien s die om non-cance causes o dea h, he SIR is a mo e sensi i e way o analysing he e ec s o cobal han he SMR. O e all cance The o e all cance incidence was no ele a ed in he p esen s udy. We used he incidence a es o he popula ion o Cen al Os obo hnia Finland as he main e e ence, because incidence a es a y geog aphically. Table 3 Obse ed (Obs) and expec ed (Exp) numbe s o cance cases and s anda dised incidence a ios (SIR) wi h 95% con idence in e als (CI) among male wo ke s a he Kokkola cobal plan du ing 1969–2013, by si e and du a ion o employmen . Only p ima y si es wi h ≥2 obse ed o expec ed cance cases included P ima y si e Employmen >1 yea Employmen >5 yea s Obs Exp SIR 95% CI Obs Exp SIR 95% CI All si es 92 91.9 1.00 0.81–1.22 77 71.6 1.08 0.85–1.34 Tongue 3 0.41 7.39 1.52–21.6 3 0.30 10.0 2.06–29.2 Oesophagus 2 1.15 1.74 0.21–6.28 2 0.89 2.24 0.27–8.08 S omach 7 3.47 2.01 0.81–4.15 5 2.73 1.83 0.59–4.26 Colon 4 4.36 0.92 0.25–2.34 4 3.46 1.16 0.32–2.96 Rec um, ec osigmoid 4 3.80 1.05 0.29–2.69 3 2.93 1.03 0.21–2.99 Panc eas 2 3.44 0.58 0.07–2.09 1 2.69 0.37 0.01–2.07 La ynx, epiglo is 2 0.81 2.45 0.30–8.86 2 0.65 3.09 0.37–11.2 Lung, achea 6 12.0 0.50 0.18–1.08 5 9.55 0.52 0.17–1.22 Melanoma o he skin 1 3.29 0.30 0.01–1.69 1 2.53 0.39 0.01–2.20 Skin, non-melanoma 3 2.78 1.08 0.22–3.15 3 2.22 1.35 0.28–3.94 Basal cell ca cinoma o he skin 18 19.2 0.94 0.56–1.48 12 15.1 0.80 0.41–1.38 P os a e 33 24.4 1.35 0.93–1.89 26 19.4 1.34 0.87–1.96 Kidney 2 3.81 0.52 0.06–1.89 2 3.00 0.67 0.08–2.40 Bladde and u ina y ac 9 4.79 1.88 0.86–3.56 6 3.76 1.60 0.59–3.47 B ain, cen al ne ous sys em 2 2.82 0.71 0.09–2.56 2 2.07 0.97 0.12–3.49 Thy oid gland 2 0.99 2.01 0.24–7.26 2 0.72 2.78 0.34–10.0 Non-Hodgkin lymphoma 3 4.44 0.68 0.14–1.97 3 3.42 0.88 0.18–2.56 Leukaemia 3 2.11 1.42 0.29–4.15 3 1.58 1.90 0.39–5.54 Table 4 Obse ed (Obs) and expec ed (Exp) numbe s o cance cases (all si es, lung and nasal cance ) and s anda dised incidence a ios (SIR) wi h 95% CI du ing 1967–2011 among wo ke s in he Kokkola cobal plan du ing 1969–2013wi h employmen o >1 yea , by exposu e g oup Exposu e g oup All si es Lung, achea Tongue Bladde Obs Exp SIR 95% CI Obs Exp SIR 95% CI Obs Exp SIR 95% CI Obs Exp SIR 95% CI Va iable 7 5.02 1.39 0.56–2.87 - 0.55 0.00 0.00–6.68 1 0.04 26.4 0.67–14.0 - 0.25 0.00 0.00–15.0 Low 42 37.8 1.11 0.80–1.50 2 4.89 0.41 0.05–1.47 1 0.15 6.48 0.16–36.1 6 1.96 3.07 1.12–6.67 Mode a e 4 6.02 0.66 0.18–1.70 - 0.66 0.00 0.00–5.56 - 0.06 0.00 0.00 - 0.30 0.00 0.00–12.2 High 39 43.6 0.90 0.64–1.22 4 5.95 0.67 0.18–1.72 1 0.16 6.12 0.15–34.1 3 2.32 1.30 0.27–3.78 To al 92 91.9 1.00 0.81–1.22 6 12.0 0.50 0.18–1.08 3 0.41 7.39 1.52–21.6 9 4.79 1.88 0.86–3.56 Sauni e al. BMC Cance (2017) 17:340 Page 5 o 8 Cance incidence also a ies acco ding o socioeconomic posi ion. The majo i y o he wo ke s in he depa men s o he Kokkola cobal plan we e skilled blue-colla wo ke s, whe eas he e e ence g oup included all socio- economic g oups. I is known ha he cance incidence o blue-colla wo ke s in mos cance ypes is close o he popula ion a e age [17], and hence, he e e ence a es used in ou calcula ions o expec ed numbe s o cases should be alid. Heal hy wo ke e ec is a phenomenon ini ially obse ed in s udies o occupa ional diseases: wo ke s usually exhibi lowe o e all dea h a es han he gene al popula ion because he se e ely ill and ch onically disabled a e o dina ily excluded om employmen [18]. The heal hy wo ke e ec migh also a ec cance inci- dence in he i s yea s a e employmen bu he e ec is much smalle han i would be in a s udy on mo ali y. In he Kokkola cobal plan he e was no selec ion o wo ke s because o possible cance isk. No ma ke s o es s we e used in p e-employmen heal h examina ions o he plan o exclude indi iduals ha could be in isk o cance . We could ollow he wo ke s also a e he end o employmen . Thus, i hey had o lea e wo k because o heal h easons, hey we e s ill included in ou coho . The e o e, we sugges ha heal hy wo ke e ec did no play a ma ked ole in his s udy. A dose- esponse ela ionship be ween exposu e le el o cobal and cance incidence would ha e suppo ed he idea o cobal as a causa i e agen o he cance s. Howe e , he o e all cance incidence did no inc ease acco ding o he cobal exposu e le el. Lung cance In he p esen s udy, he incidence o lung cance cases dec eased by 50%, and e en mo e among he coho membe s who had been wo king in he same depa - men o mo e han i e yea s. A F ench s udy [7] ound an inc eased SMR o lung cance among cobal wo ke s, bu a ollow-up s udy o he same coho wi h an ex ended obse a ion pe iod could no e i y he p e ious esul s [8]. In ha F ench s udy, he exposing compounds we e qui e compa able o hose in ou s udy, bu he exposu e le els we e no known. Tongue cance The o al s udy g oup con ained h ee cases o ongue cance . The SIR o ongue cance was signi ican ly highe in he o al g oup (7.39, 95% CI 1.52–21.60). In di e en exposu e o age g oups he inc ease was no s a is ically signi ican . To ou knowledge, no p e ious s udies ha e associa ed cobal exposu e wi h ongue cance . Acco ding o he li e a u e, he mos signi ican isk ac o s o o al cance a e smoking and alcohol use [19, 20]. All h ee cases o ongue cance in ou coho we e smoke s. We do no ha e da a on hei alcohol consump ion. In e ac ion o di e en ca cinogenic me als could be a hypo he ical explana ion o he excess o ongue cance s. Howe e , none o he ongue cance pa ien s had wo ked in he chemical depa men whe e he nickel exposu e was highes , bu obacco smoke con ains also di e en me als including aluminium, cadmium, ch omium, nickel, lead, me cu y, selenium, anadium, manganese and zinc [21], which may ha e syne gis ic e ec s wi h cobal ex- posu e. Me als a e hough o p omo e cance by a num- be o common mechanisms. In e ms o di ec damage o DNA, mos me als a e only weakly mu agenic; howe e , many a e s ong co-ca cinogens, p omo ing a syne gis ic e ec in he p esence o o he cance -causing agen s [22]. Because he small numbe o ongue cases he e, he excess may be explained by chance alone. Cance o u ina y bladde Nine cases we e bladde cance , which is nea ly wice he expec ed numbe , bu he di e ence was no s a is i- cally signi ican . Howe e , he lowes exposu e g oup had a s a is ically signi ican h ee- old excess o u ina y bladde cance cases. All in all, 2/3 o he cases we e om he lowes exposu e ca ego y. The mos impo an isk ac o o bladde cance is obacco smoking [23]. Only one bladde cance pa ien was a non-smoke ou o i e cases wi h known smoking s a us. Occupa ionally, a oma ic amines a e known o cause bladde cance as an occupa ional disease [24]. No one in his s udy g oup was known o ha e been exposed o a oma ic amines. Smoking The mos impo an sou ce o incompa abili y be ween he esul s de i ed om di e en coho s is usually con ounding due o smoking, which has no been con olled o in mos s udies and which may lead o a bias in a di e en di ec ion, depending on whe he smoking in he coho is lowe o highe han in he e e ence popula ion. Acco ding o da a ga he ed in 2000, he p e alence o cu en smoke s among he em- ployees o he Kokkola cobal plan was 31.8% [13]. The p e alence o daily smoke s in he male popula ion in he p o ince o Cen al Os obo hnia a ound he cobal plan in 1990–2005 a ied om 18% in he highes educa ional class o 25% in he lowes [25]. These pe cen ages may be an unde es ima ion o smoke s, possibly due o he small sample size. Du ing his same ime pe iod, he p e alence o daily smoke s in he whole coun y was abou 30% [25]. Mo eo e , daily smoke s and cu en smoke s may ep esen di e en g oups. Hence, he explana ion o he SIR 0.5 o lung cance in ou coho is appa en ly no a lowe smoking p e alence as compa ed o he e e ence popula ion. Sauni e al. BMC Cance (2017) 17:340 Page 6 o 8 Di e en e ec s o di e en compounds An assessmen o cobal and i s compounds equi es a clea dis inc ion be ween di e en compounds and needs o ake in o accoun he di e en mechanisms in- ol ed. The e is e idence ha soluble cobal (II) ca ions exe geno oxic and ca cinogenic ac i i y bo h in i o and in i o in expe imen al sys ems [26, 27]. The e is also e idence ha ha d me al pa icles ha con ain ungs en ca bide, in addi ion o cobal , exe geno oxic and ca cinogenic ac i i y in i o and in human s udies [28]. Howe e , he conclusion o IARC (2006) is ha he e idence is limi ed in humans o inc eased isk o lung cance in case o cobal wi h ungs en ca bide and inad- equa e in case o cobal wi hou ungs en ca bide [1]. Fo ca cinogenici y o cobal oxides and o he com- pounds he e is insu icien in o ma ion [28]. Conclusions The esul s sugges ha occupa ional exposu e o cobal is no associa ed wi h an inc eased o e all cance isk o lung cance isk among cobal wo ke s. Unexpec edly we ound a signi ican inc ease in he incidence o ongue cance . The e a e no p e ious da a om ei he animal o human s udies o suppo an associa ion be ween cobal and ongue cance . Because o he small numbe o cance cases he esul s mus be in e p e ed wi h cau ion. Abb e ia ions CI: Con idence in e al; IARC: In e na ional Agency o Resea ch on Cance ; SIR: S anda dised incidence a io Acknowledgemen s We acknowledge Alice Leh inen o p o essional edi ing o g amma o he manusc ip . Funding This esea ch ecei ed unding o da a collec ion om F eepo Cobal . The analysis, in e p e a ion o da a and w i ing was pe o med independen ly. A ailabili y o da a and ma e ials Da a can be made a ailable wi h he pe mission o Kokkola cobal plan and Finnish Ins i u e o Occupa ional Heal h. Au ho s’con ibu ions RS, PO and JU p epa ed he s udy plan and he a icle, AL and RK ex ac ed he da a om he iles o he company’s employmen eco ds, and EP pa icipa ed in s udy design and pe o med he s a is ical analyses. All au ho s ead and app o ed he inal manusc ip . Compe ing in e es s RK has been employed by Boliden since 2006 and AL 1983–2005. Boliden and he Kokkola Cobal Plan a e loca ed in he same Kokkola Indus ial Pa k a ea. Boliden has p o ided occupa ional heal h se ices o he Kokkola Cobal Plan un il end o 2016. Consen o publica ion No applicable. E hics app o al Na ional Ins i u e o Heal h and Wel a e, Finland. 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 Depa men o Occupa ional Sa e y and Heal h, Minis y o Social A ai s and Heal h, P.O. Box 33FI-00023 Go e nmen Tampe e, Finland. 2 Clinic o Occupa ional Medicine, Tampe e Uni e si y Hospi al, Tampe e, Finland. 3 Finnish Ins i u e o Occupa ional Heal h, Tampe e, Finland. 4 The Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland. 5 Heal h Cen e o Kokkola own, Kokkola, Finland. 6 Occupa ional Heal h, Boliden, Kokkola, Finland. 7 Facul y o Social Sciences, Uni e si y o Tampe e, Tampe e, Finland. 8 Finnish Cance Regis y, Ins i u e o S a is ical and Epidemiological Cance Resea ch, Helsinki, Finland. Recei ed: 23 May 2016 Accep ed: 9 May 2017 Re e ences 1. In e na ional Agency o Resea ch on Cance . Cobal in Ha d Me als and Cobal Sul a e, Gallium A senide, Indium Phosphide and Vanadium Pen oxide. IARC Monog E al Ca cinog Risks Hum. 2006;86:37–155. 2. Shedd KB. Mine als yea book: cobal . Res on: US geological su ey; 2001. p. 20.1–20.18. h p://mine als.usgs.go /mine als/pubs/commodi y/cobal /index. h ml. Accessed 4 Ma 2017. 3. Shedd KB. Mine als yea book: cobal . Res on: US geological su ey; 2014. p. 19.10. h p://mine als.usgs.go /mine als/pubs/commodi y/cobal /index.h ml. Accessed 4 Ma 2017. 4. Na ional Ins i u e o Heal h. Cobal . h ps://www.cdc.go /niosh/ opics/ cobal /. Accessed 4 Ma 2017. 5. Vainio H, Liesi uo i J, Leh ola M, Loueka i K, Engs öm K, Kauppinen T, Ku ppa K, Riipinen H, Sa olainen K, Tossa ainen A. Chemicals and wo k (in Finnish). Helsinki: Finnish Ins i u e o Occupa ional Heal h; 2005. 6. Högs ed C, Alexande sson R. Mo ali y s udy among wo ke s om he ha d me al indus y (in Swedish), A be e och Hälsa; 1990. p. 211–20. 7. Mu JM, Moulin JJ, Cha uye -Seine a MP, La i e J. A coho mo ali y s udy among cobal and sodium wo ke s in an elec ochemical plan . Am J Ind Med. 1987;11:75–81. 8. Moulin JJ, Wild P, Romazini S, Las a gues G, Pel ie A, Bozec C, e al. Lung cance isk on ha d-me al wo ke s. Am J Epidemiol. 1998;148:241–8. 9. G ims ud TK, Be ge SR, Haldo sen T, Ande sen A. Can lung cance isk among nickel e ine y wo ke s be explained by occupa ional exposu es o he han nickel? Epidemiology. 2005;16:146–54. 10. Las a gues G, Wild P, Moulin JJ, Hammon B, Rosmo duc B, Rondeau du Noye C, e al. Lung cance mo ali y o a F ench coho o ha d-me al wo ke s. Am J Ind Med. 1994;26:585–95. 11. Eu opean Commission. Communica ion om he commission o he Eu opean Pa liamen , he Council, he Eu opean Economic and Social Commi ee and he Commi ee o Regions. Sa e and Heal hie Wo k o All - Mode nisa ion o he EU Occupa ional Sa e y and Heal h Legisla ion and Policy. COM(2017) 12 Final. B ussels; 2017. 12. Linna A, Oksa P, Palm oos P, Ro o P, Laippala P, Ui i J. Respi a o y heal h o cobal p oduc ion wo ke s. Am J Ind Med. 2003;44(2):124–32. 13. Linna A, Oksa P, G ounds oem K, Halkosaa i M, Palm oos P, Huikko S, e al. Exposu e o cobal in he p oduc ion o cobal and cobal compounds and i s e ec on he hea . Occup En i on Med. 2004;61(11):877–85. 14. 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;33:365–9. 15. Pukkala E. Biobanks and egis e s in epidemiological esea ch on cance . In: Dillne J, edi o . Me hods in biobanking. Me hods in molecula biology. To owa: Humana P ess. 2011;675:127–64. 16. Pa ela M, Ui i J, Pukkala E. Cance incidence among coppe smel ing and nickel e ining wo ke s in Finland. Am J Ind Med. 2017;60:87–95. 17. Pukkala E. Cance isk by social class and occupa ion. A su ey o 109,000 cance cases among Finns o wo king age. Con ibu ions o epidemiology and bios a is ics 7. Basel: Ka ge ; 1995. 18. Las J. A dic iona y o epidemiology. 3 d ed. Ox o d: Ox o d Uni e si y P ess; 1995. 19. Bagna di V, Blangia do M, La Vecchia C, Co ao G. A me a-analysis o alcohol d inking and cance isk. B J Cance . 2001;85:1700–5. Sauni e al. BMC Cance (2017) 17:340 Page 7 o 8 20. Mayne ST, Mo se DE, Winn DM. Cance s o he o al ca i y and pha ynx. In: Scho en eld D, F aumeni J JF, edi o s. Cance epidemiology and p e en ion. New Yo k (US): Ox o d Uni e si y P ess; 2006. p. 674–96. 21. Be nha d D, Rossmann A, Wick G. Me als in ciga e e smoke. IUBMB Li e. 57(12):805–9. 22. Du ham TR, Snow ET. Me al ions and ca cinogenesis. In: Bignold LP, edi o . Cance : cell s uc u es, ca cinogens and genomic ins abili y. Swi ze land: Bi khäuse Ve lag; 2006. p. 97–130. 23. Cumbe ba ch MG, Ro a M, Ca o JW, La Vecchia C. The ole o obacco smoke in bladde and kidney ca cinogenesis: a compa ison o exposu es and me a-analysis o incidence and mo ali y isks. Eu U ol. 2016;70(3):458– 66. doi: 10.1016/j.eu u o.2015.06.042. 24. Schul e PA, Ringen K, Hems ee GP, Al ek use EB, Gullen WH, Tille S, e al. Risk ac o s o bladde cance in a coho exposed o a oma ic amines. Cance . 1986;58(9):2156–62. 25. Helako pi S, Lai alainen E, Abse z P, To ppa J, Uu ela A, Puska P. Heal h beha iou and heal h among Finnish adul s in he Finnish egions in 1978– 2005 (In Finnish, wi h English abs ac ). Na ional Public Heal h Ins i u e. 2007;B15:14. 26. Buche JR, Hailey JR, Royc o JR, Haseman JK, Sills RC, G umbein SL, e al. Inhala ion oxici y and ca cinogenici y s udies o cobal sul a e. Toxicol Sci. 1999;49:56–67. 27. Bo m PJ, D oscoll K. Pa icles, in lamma ion and espi a o y ac ca cinogenesis. Toxicol Le . 1996;88:109–13. 28. Lison D, De Boeck M, Ve ougs ae e V, Ki sch-Volde s M. Upda e on he geno oxici y and ca cinogenici y o cobal compounds. Occup En i on Med. 2001;58:619–25. • 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: Sauni e al. BMC Cance (2017) 17:340 Page 8 o 8