Residential radon exposure and brain cancer: an ecological study in a radon prone area (Galicia, Spain)

1 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 www.na u e.com/scien i ic epo s Residen ial adon exposu e and b ain cance : an ecological s udy in a adon p one a ea (Galicia, Spain) Albe o Ruano-Ra ina1,2,3, Nu ia A agonés2,4, Ka l T. Kelsey3, Mónica Pé ez-Ríos1,2, Ma ía Piñei o-Lamas2, Gonzalo López-Aben e2,4 & Juan M. Ba os-Dios1,2,5 We aimed o know i adon concen a ion is associa ed wi h municipal mo ali y due o b ain cance in Galicia, Spain. We designed an ecological s udy aking as s udy uni Galician municipali ies. To be included, municipali ies had o ha e a leas h ee adon measu emen s. We co ela ed adon concen a ions wi h municipal mo ali y due o hese malignan umo s du ing he pe iod 1999–2008. We calcula ed he ela i e isk o dying o b ain cance s o each municipali y and co ela ed his alue wi h municipal adon concen a ion using Spea man’s Rho. 251 municipali ies we e included, wi h close o 3,500 adon measu emen s and an a e age o 14 adon measu emen s a each municipali y. We obse ed a signi ican co ela ion be ween esiden ial adon wi h b ain cance mo ali y o males and emales and he in ensi y o he co ela ion was highe o emales. These esul s we e ein o ced when he analysis was es ic ed o municipali ies wi h mo e han 5 adon measu emen s: Spea man’s Rho 0.286 (p- alue < 0.001) and Spea man’s Rho 0.509 (p- alue < 0.001) o males and emales, espec i ely. These esul s sugges an associa ion be ween esiden ial adon and b ain cance mo ali y. Mo e esea ch using mo e obus epidemiological designs is needed o con i m hese indings. B ain cance s a e no equen . Age adjus ed incidence a es o Eu ope a e 6.3 and 4.7 cases pe 100,000 inhab- i an s o males and emales, espec i ely1, while in Spain hese a es a e 6.1 and 4.21. Despi e i s low incidence, hese umo s ha e a high mo ali y, wi h a 5-yea su i al o only 22%2. Di e ences on su i al a e minimum by sex. These umo s can be di ided acco ding o hei his ology, and hey can be classi ied as gliomas, neu onal umo s, poo ly di e en ia ed umo s, meningiomas and o he umo s. The e is poo in o ma ion ega ding he isk ac o s o b ain umo s. Due o i s low incidence i is di icul o assess he ole o en i onmen al and gene ic ac o s, and also he possible in e ac ions among hem. Some he ed- i a y synd omes inc ease he isk o hese umo s, such as such as neu o ib oma osis, Li-F aumeni synd ome, ube ous scle osis, Von Hippel-Lindau synd ome o e inoblas oma3, 4. One o he mos es ablished isk ac o s o b ain cance s is exposu e o ionizing adia ion, mainly om medi- cal imaging5. A ecen s udy has obse ed an associa ion be ween compu e ized omog aphies and b ain cance s6. Though some esea che s ha e sugges ed ha exposu e o elec omagne ic adia ion could be associa ed wi h b ain cance s, his ela ionship has no been con i med4. Some hing simila happened wi h he elec omagne ic adia ion coming om he use o cell phones, bu he INTERPHONE s udy did no ind a consis en associa ion7. Tobacco and alcohol consump ion ha e no been associa ed wi h hese cance s8. Radon is a na u ally occu ing gas ha is eleased by he bed ocks p esen in he ea h c us . I is odo less, colo less and as eless and i is se led in he U anium 238 decay chain. I has a hal -li e o 3.8 days and i is no a heal h p oblem by i sel . The eal p oblem comes om i s sho -li ed descendan s Polonium 214 and Polonium 218, which ha e a e y sho hal -li e. When hese elemen s disin eg a e, hey elease alpha pa icles ha impac cells loca ed in he lung epi helium and inally can cause lung cance 9. Radon was decla ed a human ca cinogen by he In e na ional Agency o Resea ch on Cance (IARC) in 198810. Radon accumula es indoo s in houses and 1Depa men o P e en i e Medicine and Public Heal h, Uni e si y o San iago de Compos ela, San iago, Spain. 2CIBER de Epidemiología y Salud Pública, CIBERESP, San iago, Spain. 3Depa men o Epidemiology, B own School o Public Heal h, P o idence, Rhode, Island, USA. 4En i onmen al and Cance Epidemiology Uni , Na ional Cen e o Epidemiology, Ca los III Ins i u e o Public Heal h, Mad id, Spain. 5P e en i e Medicine Uni , San iago de Compos ela Clinic Uni e si y Hospi al, San iago, Spain. Co espondence and eques s o ma e ials should be add essed o A.R. (email: [email p o ec ed]) Recei ed: 5 Janua y 2017 Accep ed: 5 May 2017 Published: xx xx xxxx OPEN www.na u e.com/scien i ic epo s/ 2 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 wo kplaces, and i can en y om he soil mainly h ough c acks and join s and i is also highly soluble in wa e . I is accep ed ha people li ing in places wi h high adon concen a ions ha e a highe isk o lung cance 9. Since b ain cance is associa ed wi h ionizing adia ion11, i is biologically plausible ha esiden ial adon exposu e migh inc ease i s isk. O gans apa he lungs may ecei e low doses om adon exposu e, as s a ed in he WHO epo on indoo adon9. Ne e heless, e y ew s udies ha e assessed he ela ionship be ween adon and b ain umo s. Radia ion dose ecei ed by he b ain due o adon and i s byp oduc s is much lowe han ha ecei ed by o he o gans12. A e one yea o an a e age exposu e o 200 Bq/m3 by inhala ion, he b ain would ecei e be ween 0.06 and 0.15 mS whe eas he lung in he same si ua ion would ecei e be ween 35.8 and 159 mS . Since 200 Bq/m3 is app oxima ely equi alen o 10 mSV pe yea (one compu ed omog aphy), he b ain would ecei e a dose equi alen o 20 compu ed omog aphies du ing 20 yea s. I has been p oposed ha mac- ophages migh phagocy e small solid pa icles in he lungs om adon descendan s ha could each he b ain h ough he blood12. The exis ing li e a u e is con o e sial and mos s udies ha e been pe o med in mine s. In gene al adul popula ion, a Danish s udy obse ed a s a is ically signi ican associa ion be ween esiden ial adon and b ain cance s13 bu a s udy based on he Cance P e en ion S udy-II showed no associa ion be ween mean coun y adon le els and b ain cance mo ali y14. Galicia, he s udy a ea, has been de ined p e iously as a adon p one egion. App oxima ely 20% o all dwell- ings ha e adon concen a ions abo e he En i onmen al P o ec ion Agency ac ion le el (148 Bq/m3)15–17. The high adon concen a ions ound in Galicia acili a e he analysis o possible associa ions be ween adon and o he cance s di e en han lung cance 18 because i is easie o ha e municipali ies wi h high adon concen a- ions and co ela e i hese municipali ies ha e highe o lowe b ain cance mo ali y. The aim o he p esen s udy is o analyze i he e is a co ela ion be ween esiden ial adon le els and b ain cance mo ali y a a municipal le el in Galician popula ion. A subanalysis will be pe o med o know i his co ela ion di e s by sex. Me hods Se ing and s udy design. The s udy has been pe o med in Galicia, No hwes o Spain. Galicia limi s o he No h and Wes wi h he A lan ic Ocean, o he Sou h wi h Po ugal and o he Eas wi h mainland Spain. A ound hal o he popula ion li es in u al a eas, mainly in de ached houses. We designed an ecological s udy aimed o analyze he co ela ion be ween esiden ial adon and isk a ios o b ain cance mo ali y o Galician municipali ies. We es ablished as inclusion c i e ion ha each municipali y had o ha e a leas 3 esiden ial adon measu emen s. Radon measu emen s. Radon measu emen s we e ob ained om he Galician Radon Map and om con- ols belonging o wo case-con ol s udies pe o med p e iously17, 19. The Galician Radon Map has been de el- oped by he Uni e si y o San iago de Compos ela o iden i y hose a eas wi h he highes adon concen a ions in Galicia. Pa icipan s in he Galician Radon Map we e selec ed om he gene al popula ion h ough a s a i ied andom sampling weigh ed by he popula ion o each municipali y16. The sampling was designed so ha a leas wo adon measu emen s we e a ailable in hose municipali ies less popula ed. Con ols om wo hospi al-based case-con ol s udies we e also included in he p esen esea ch. This inclusion allowed us o inc ease he sample size in mo e han 600 adon measu emen s. Con ols om one case-con ol s udy17 we e ec ui ed om wo Galician hospi als among people esiding in municipali ies co e ed by hose hospi als. The second case-con ol s udy had a simila design19, wi h he di e ence ha i ec ui ed con ols li ing in he ca chmen a eas o ou di e en Galician hospi als. We did no include lung cance cases since adon is associa ed wi h lung cance and adon concen a ion in cases could bias ou adon measu emen s owa ds a highe median. Fo all measu emen s, adon de ec o s we e placed o a minimum o h ee mon hs and a maximum o six. We used alpha- ack de ec o s (Radosys Inc, Budapes , Hunga y). Radon de ices we e placed by a adon ech- nician o by pa icipan s in he main bed oom and, in some cases, in he li ing oom o he measu ed dwelling. De ec o s we e away om doo s, windows o elec ic de ices, and a a dis ance o he loo be ween 60 and 180 cm. When he measu emen pe iod inished, he de ec o s we e sen by he pa icipan s in a specially sealed en elope o he Galician Radon Labo a o y, loca ed a he San iago de Compos ela Clinic Uni e si y Hospi al. The Galician Radon Labo a o y egula ly akes pa in quali y con ol p ocedu es o ganized by he Uni e si y o Can ab ia and he Nuclea Sa e y Council o Spain wi h excellen esul s20, 21. In some dwellings, we included double measu emen s o check he eliabili y o ou p ocesses. All adon measu emen s we e adjus ed aking in o accoun seasonal a ia ions. B ain cance mo ali y. We calcula ed s anda dized mo ali y a ios (SMR) o b ain cance o all Galician municipali ies o a 10-yea pe iod (1999–2008) using da a supplied by he Na ional S a is ics Ins i u e. We used all Spanish dea h en ies o he pe iod conside ed co esponding o b ain cance (In e na ional Classi ica ion o Diseases[ICD], 9 h e ision code 191, ICD-10 code C71) b oken down by sex, age g oup (18 g oups), i e-yea pe iod (1999–2003, 2004–2008) and municipali y. The municipal popula ions, also b oken down by sex and age g oup, we e d awn om he 2001 census and 2006 municipal oll. These yea s co espond o he midway poin s o he wo quinquennia ha comp ise he s udy pe iod. Pe son-yea s o each i e-yea pe iod we e es ima ed by mul iplying hese popula ions by 5. Fi s ly, s anda dized mo ali y a ios (SMR) we e compu ed as he a io o obse ed e sus expec ed dea hs. Expec ed cases we e compu ed aking Spanish b ain cance mo ali y a es as e e ence, b oken down by sex, age and quinquenium. In a second s ep, smoo hed municipal ela i e isks (RR) we e calcula ed using he condi ional au o eg essi e spa ial model de eloped by Besag, Yo k and Mollié (BYM)22. I is a spa ial Poisson model, wi h obse ed cases as he dependen a iable, expec ed cases as o se , and wo ypes o andom-e ec s e ms ha www.na u e.com/scien i ic epo s/ 3 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 ake he ollowing in o accoun : (a) he e ec s which a y in a s uc u ed manne in space (municipal con igu- i y); and, (b) a componen ha models he e ec s which a y among municipali ies in an uns uc u ed manne (municipal he e ogenei y). The c i e ion o con igui y used was adjacency o municipal bounda ies. The model akes he ollowing o m: λ λα ∼ =++ OPoE hb () log( ) iii iii whe e: λi is he ela i e isk in a ea i; Oi is he numbe o dea hs in a ea i; Po is Poisson dis ibu ion; α is he in e cep ; Ei a e he expec ed numbe o cases; hi is he municipal he e ogenei y e m; and bi is he spa ial e m. In eg a ed nes ed Laplace app oxima ions (INLAs)23 we e used as a ool o Bayesian in e ence. Fo his pu pose, we used R-INLA. Municipal ela i e isks (RR) ha e been independen ly es ima ed in bo h men and women. S a is ical analysis. We pe o med a desc ip i e analysis (uni a ia e) on esiden ial adon measu emen s and also on b ain cance mo ali y RR b oken down by municipali y. Fo each municipali y wi h mo e han 3 adon measu emen s we ob ained he numbe o o al measu emen s and he median esiden ial adon. We ob ained he mo ali y isk o b ain cance sepa a ely by males and emales. The desc ip i e analysis was ollowed by a bi a ia e analysis whe e we used Spea man co ela ions o associa e median municipal adon concen a ion wi h b ain cance mo ali y RR o males and emales, espec i ely. We also displayed g aphically he co ela- ions compa ing ela i e isks wi h esiden ial adon le els. We also pe o med a sensi i i y analysis including only hose municipali ies ha ing mo e han 5 adon measu emen s. We conside ed s a is ical signi icance when p- alue was lowe han 0.05. The analysis was pe o med wi h IBM SPSS 20 (IBM, A monk, NY, USA). Resul s This s udy has included 3,498 adon measu emen s in 251 Galician municipali ies, an a e age o 14 measu e- men s pe municipali y. Six y- ou municipali ies we e no included (18.7% o he o al) because hey had less han 3 adon measu emen s. 50.6% o he included municipali ies had adon le els (geome ic mean) abo e WHO ecommenda ion (100 Bq/m3), 25.9% had geome ic means abo e EPA ac ion le els (148 Bq/m3) and 13.5% had geome ic means abo e 200 Bq/m3. A desc ip ion o adon concen a ions by municipali y can be obse ed in Table1. Rega ding mo ali y due o b ain cance du ing he pe iod 1999–2008, he e we e 949 o e all dea hs among males and 758 among emales in he included municipali ies. The ange o dea hs due o b ain cance (maximum-minimum) o males was 0–99 and 0–107 o emales. Fo males, mos municipali ies (136, 54.1%) had a low ela i e isk o dea h o hese cance s (be ween 1 and 1.04). Only 23 (9.2%) o municipali ies had a el- a i e isk o dea h o males highe han 1.10. Fo emales, he isk o dea h o hese cance s was lowe compa ed Va iables N (%) Numbe o adon measu emen s by municipali y 3–4 78 (31.1) 5–14 120 (47.8) 15–29 32 (12.7) 30–49 10 (4.0) ≥50 11 (4.4) Radon concen a ion by municipali y (Geome ic mean Bq/m3) <100 124 (49.4) 100–147 62 (24.7) 148–199 31 (12.3) 200–249 20 (8.0) ≥250 14 (5.6) Rela i e Risk o b ain cance in males by municipali y <1, 0 8 (3.2) 1, 0–1, 04 136 (54.1) 1, 05–1, 09 84 (33.5) ≥1, 10 23 (9.2) Rela i e Risk o b ain cance in emales by municipali y <1, 0 151 (60.2) 1, 0–1, 04 53 (21.1) 1, 05–1, 09 29 (11.5) ≥1, 10 18 (7.1) Table 1. Cha ac e is ics o esiden ial adon measu emen s and isk o b ain cance in he included municipali ies. www.na u e.com/scien i ic epo s/ 4 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 o males. Only 18 municipali ies (7.1%) had a ela i e isk o dea h highe han 1.10. This in o ma ion is shown in Table1. Rega ding he co ela ion be ween esiden ial adon and b ain cance mo ali y, i was s a is ically signi i- can o males and emales, o emales he associa ion obse ed was also mode a e. The Spea man’s Rho was 0.164 (p- alue 0.009) o males and 0.433 (p- alue < 0.001) o emales. When he analysis was es ic ed o hose municipali ies ha ing a leas 5 o mo e adon measu emen s he s a is ical signi icance pe sis s and he s eng h o he associa ion inc eases (measu ed h ough he Spea man’s Rho). Fo males, Spea man’s Rho was 0.286 (p- alue < 0.001) and o emales Spea man’s Rho was 0.509 (p < 0.001). The sca e plo showing he co - ela ion be ween esiden ial adon and b ain cance mo ali y can be obse ed in Fig.1A and B. Fo bo h sexes he co ela ion is clea . I we analyze only hose municipali ies wi h 10 o mo e adon measu emen s (84 o he 251, 26.6% o he o al included), he co ela ion pe sis s. Fo males, Spea man’s Rho is 0.314 (p = 0.004) and o emales Spea man’s Rho is 0.642 (p- alue < 0.0001). Table2 shows he 15 municipali ies wi h he highes esiden ial adon concen a ions in Galicia wi h hei el- a i e isk (and pos e io p obabili y ha RR > 1) o mo ali y o b ain cance s. I can be obse ed ha o males, all municipali ies ha e ela i e isks abo e one and 5 municipali ies ha e RR abo e pe cen ile 75. Fo emales, 6 municipali ies ha e ela i e isks abo e pe cen ile 75 o all ela i e isks o he municipali ies included in he anal- ysis (n = 251), hough he emaining 9 municipali ies had ela i e isks below 1 (bu e y close o 1). Discussion The e is e y li le in o ma ion a ailable on he possible ela ionship be ween esiden ial adon and b ain cance . In he p esen s udy, we ha e obse ed a s a is ically signi ican co ela ion be ween esiden ial adon exposu e a a municipal le el and mo ali y due o b ain cance o males and emales. The associa ion appea s o be sligh ly highe o emales. The co ela ion inc eases when he analysis is es ic ed o municipali ies ha ing 5 o mo e adon measu emen s and is clea ly iden i iable in he sca e plo s. This is he second s udy pe o med in Eu ope on his opic and he i s pe o med in a high- isk adon a ea. The possible ela ionship be ween indoo adon and b ain cance in adul popula ion has been s udied in h ee di e en in es iga ions. The s udy by Tu ne e al. had a coho design using he cance p e en ion S udy II coho 14. The s udy analyzed b ain umo s mo ali y, and adon measu emen s we e es ima ed assuming he Figu e 1. Sca e plo showing he co ela ion be ween municipal adon concen a ion and ela i e isk o mo ali y o b ain cance in (A) males, and (B) emales. www.na u e.com/scien i ic epo s/ 5 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 median o each coun y whe e each pa icipan esided. In his ecological s udy including 2,232 cases, each coun y had an a e age o 10 adon measu emen s, and no e ec was obse ed o esiden ial adon exposu e, wi h a Haza d Ra io o HR 0.98 pe 100 Bq/m3 (95% CI 0.83–1.15). O no e, Tu ne e al. used coun ies as he geo- g aphical uni s o s udy, which ha e much highe a ea and popula ion han Galician municipali ies. The e o e, in ela i e e ms, he p esen s udy has a highe numbe o adon measu emen s o cha ac e ize municipali ies. Fu he mo e, while in he Tu ne e al. s udy he a e age mean concen a ion o adon was 53.5 Bq/m3, in ou s udy only 43 o he 251 s udied municipali ies (17.1%) had esiden ial adon concen a ions below he Tu ne ’s a e age. A second s udy was pe o med in Denma k by B äune e al., whe e he au ho s obse ed a s a is ically signi ican e ec o 1.44 (95% CI 1.07–1.93) o each 100 Bq/m3 o inc emen in esiden ial adon concen a ion13. In his s udy, adon concen a ion was es ima ed in each pa icipan esidence, using a alida ed me hod. The emaining s udy, published in 1993 and pe o med in Maine (USA), analyzed he p esence o adon in wa e and did no ind any associa ion wi h b ain umo s24 using an ecological design. The e is mo e published li e a u e in mine s analyzing adon exposu e and b ain cance s, hough he esul s a e no consis en . Resul s om he F ench Mine s’ Coho 25–27 sugges ed an associa ion wi h adon exposu e, while esul s om he Ge man Mine s’ coho (Wismu Mining Company) did no 28, 29. O he mine s’ s udies30, 31 ha e no obse ed any associa ion while o he obse ed a close o signi ican associa ion32. Only one o hese s udies has eached s a is ical signi icance, wi h obse ed o expec ed cases a io o 2.00 (95% CI 1.09–3.35)27. Mos o hese s udies had a poo adjus men o co a ia es and gi en he low equency o b ain umo s, he num- be o obse ed cases was e y low in mos o hem, causing imp ecise es ima ions. Fu he mo e, hough some s udies adjus ed by o he ca cinogens ha can be p esen in mines (such as gamma ays, a senic o dus s), esidual con ounding could no be excluded. The e is biological plausibili y o suppo an associa ion be ween esiden ial adon and b ain cance . Though he e ec i e dose eaching he b ain om adon exposu e is much lowe han ha ecei ed by he lungs, i has been es ima ed ha b ain would ecei e a dose be ween 0.06 and 0.15 mS a e one yea o an a e age exposu e by inhala ion o 200 Bq/m3. In he s udied geog aphical a ea, 34 municipali ies ha e a geome ic mean o esi- den ial adon concen a ion highe han 200 Bq/m3. I we conside ha Galician popula ion ends o li e in he same house o decades19, we could es ima e ha pa o ou popula ion is ecei ing impo an amoun s o adia- ion due o esiden ial adon exposu e and he e o e could p omo e b ain cance . A s ong ela ionship be ween esiden ial adon and lung cance has al eady been obse ed in ou egion in e e and ne e -smoke s17, 19. Few s udies ha e assessed he biological e ec o adon in he b ain. The e is a esea ch on how low adon concen a- ions could ac i a e supe oxide dismu ase in a s bu high le els o long exposu es inac i a e i 33. Anyway, his obse a ion does no explain a possible ca cinogenic e ec . The highe co ela ion obse ed o women compa ed o men migh be due o a mo e in ense exposu e o adon among emales. Olde coho s o Galician women ha e been p edominan ly housewi es and women ha e s a ed wo king ou o hei homes in he las wo decades. This means ha olde women we e mo e exposed o esiden ial adon han hei male coun e pa s due o his mo e in ense exposu e a home. A highe suscep ibili y o women o alpha adia ion canno be dis ega ded ei he . Ou esul s should be in e p e ed aking in o accoun some limi a ions. Rega ding b ain cance da a, since he e is no a popula ion-based cance incidence egis y in Galicia, mo ali y is he only a ailable sou ce o in o - ma ion. While due o i s low su i al (5-yea ela i e su i al is app oxima ely 20% o malignan umo s)34, b ain cance mo ali y can be conside ed a good es ima o o incidence da a, a limi a ion o mo ali y in o ma ion is Municipali y Median adon concen a ion (Bq/m3)RR in menaPP RR >1cRR in womenbPP RR > 1c O Bolo 450 (n = 3) 1.02 0.56 0.96 0.39 A Mezqui a 429 (n = 3) 1.01 0.52 0.94 0.35 Pad enda 363 (n = 5) 1.04 0.61 0.98 0.45 A Pob a de T i es 341 (n = 3) 1.03 0.60 0.98 0.43 Esgos 313 (n = 3) 1.04 0.61 0.99 0.47 Cenlle 312 (n = 5) 1.03 0.60 1.02 0.55 Fo nelos de Mon es 309 (n = 3) 1.06 0.68 1.07 0.69 Samos 292 (n = 3) 1.02 0.57 0.95 0.34 Cualed o 289 (n = 4) 1.00 0.51 0.98 0.45 Mos 281 (n = 20) 1.11 0.79 1.22 0.91 Sobe 271 (n = 3) 1.03 0.61 0.99 0.47 Campo Lamei o 264 (n = 3) 1.08 0.73 1.06 0.68 Salceda de Caselas 255 (n = 5) 1.15 0.86 1.12 0.78 Meaño 251 (n = 8) 1.12 0.80 1.08 0.69 Bueu 247 (n = 8) 1.10 0.73 0.98 0.45 Table 2. Municipali ies wi h he anked by esiden ial adon concen a ion and ela i e isk o b ain cance in males and emales. aIn bold hose municipali ies wi h ela i e isk o b ain cance abo e pe cen ile 75 (1.07). bIn bold hose municipali ies wi h ela i e isk o b ain cance abo e pe cen ile 75 (1.02). cPos e io p obabili y ha RR >1. www.na u e.com/scien i ic epo s/ 6 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 ha i didn’ allow us o analyze b ain cance by sub ype, and we canno exclude ha adon exposu e may ha e a pa icula impac in some mo phological ypes. As ega ds o he me hodology, an ob ious limi a ion o his ecological s udy is ha causal associa ions canno be es ablished. We canno dis ega d ha some unmeasu ed o unknown isk ac o s o b ain umo s jeopa dize he associa ion obse ed. In any case, we do no conside ha esiden ial adon is associa ed wi h o he isk ac- o s o b ain umo s, as has been obse ed in o he s udies. Radon exposu e measu emen may also be a conce n. In he p esen s udy, we ha e a limi ed numbe o adon measu emen s in some municipali ies, bu some o hem ha e a small popula ion (in many cases below 500 inhabi an s), and 3 measu emen s could be conside ed an accep able numbe o es ima e indoo adon in such small a eas. Al oge he , we belie e ha he a ailable adon measu emen s a e adequa e o cha ac e ize a e age esiden ial adon, as has been done in o he simila s udies14. A u he limi a ion is he use o SMR, which has long been c i icized since i can be misused and misin e - p e ed, due o he in insic di icul y o compa ison among SMRs, because he weigh ing o each SMR depends on he age s uc u e o he s udy popula ion35, 36. Though some au ho s ha e p oposed special solu ions37, he use o SMRs has become widesp ead due o i s easy in e p e abili y and because when iewed as a weigh ed a e - age o he a ios o age-speci ic a es o coho and s anda d popula ion, he weigh s minimize he a iance o weigh ed a e age36. In p ac ical e ms, he SMR end o be less sensi i e o nume ical ins abili ies in one o wo o he age-speci ic a es. Some p oblems o he SMR a e ex ensi e o he use o age-speci ic a es, since in bo h cases some in o ma ion is los . In ou expe ience and wi h Galician da a, he inclusion o he age componen (wi h pe son-yea s as o se ) in he BYM models o smoo hing RRs does no al e he esul s subs an ially. Finally, conce ning s a is ical modelling, we used he BYM model o ob ain a mo ali y indica o which a oids he p oblem o he small a ea analysis, allowing us o add ess he co ela ion s udy, hough we canno disca d comple ely ha he spa ial dependence was no he same h oughou he e i o y o Galicia. The model has o conceal SMRs wi h esiden ial adon concen a ion a a municipal le el. The e a e no o he ele an sou ces o ionizing adia ion exposu e apa om adon in he s udy se ing. Radon exposu e comp ises app oxima ely 50% o all li e ime exposu e o ionizing adia ion, ollowed by medical exposu e (15–20%). Gi en ha he s udy a ea is a adon p one-a ea we could say ha adon exposu e p obably accoun s o mo e han a 50%. Since Galician popula ion is co e ed by he same heal h p o ide (Galician Heal h Se ice), which is uni e sal and ee, we do no expec ele an di e ences in ionizing exposu e coming om medical de ices o he popula ion esiding in each o he included municipali ies. The p esen s udy has also some ad an ages. The i s one is ha Galicia is a adon p one a ea, as has been published be o e15, 16. In ac , o all he a ailable s udies on his opic, Galicia is he a ea which has he highes esiden ial adon le els. This cha ac e is ic makes easie o analyze he e ec o adon in b ain cance mo ali y. The second ad an age o his wo k is ha Galician popula ion ends o li e o a long ime in he same dwelling and he e o e he po en ial la ency pe iod o adon exposu e could be ul illed mo e easily when compa ed o o he geog aphical se ings whe e people has mo e mobili y. A hi d ad an age is ha we a e using eal adon measu emen s ins ead o es ima ions o adon concen a ion, as has been done in o he in es iga ions13. This means ha he cha ac e iza ion o adon exposu e is mo e eliable in he p esen s udy. A las ad an age is he use o BYM model, which is a sui able ool o ob aining smoo hed indica o s in spa ial epidemiology. In ou con ex , i has been used o ob ain a mo ali y indica o which bypasses he p oblem o he “small a ea” analysis, allowing us o add ess he co ela ion s udy. To conclude, his s udy sugges s ha a possible associa ion be ween esiden ial adon and mo ali y due o b ain cance migh exis . This associa ion is biologically plausible due o he ole o ionizing adia ion om o he sou ces on he onse o hese malignan umo s. To ha e a deepe knowledge o his ela ionship, i is impo an o pe o m well-designed case-con ol s udies, p e e ably in adon p one a eas o shed mo e ligh on his associ- a ion. Heal h au ho i ies should p omo e hese s udies and aise public awa eness on he al eady known heal h isks o indoo adon exposu e. Re e ences 1. Fe lay J. e al. GLOBOCAN 2012 1.0, Cance Incidence and Mo ali y Wo ldwide: IARC Cance Base No. 11 [In e ne ]. (In e na ional Agency o Resea ch on Cance ). 2. De Angelis, R. e al. Cance su i al in Eu ope 1999–2007 by coun y and age: esul s o EUROCARE–5-a popula ion-based s udy. Lance Oncol. 15, 23–34 (2014). 3. Range , A. M., Pa el, Y. K., Chaudha y, N. & Anan ha, R. V. Familial synd omes associa ed wi h in ac anial umou s: a e iew. Childs Ne . Sys . ChNS O . J. In . Soc. Pedia . Neu osu g. 30, 47–64 (2014). 4. Goodenbe ge , M. L. & Jenkins, R. B. Gene ics o adul glioma. Cance Gene 205, 613–621 (2012). 5. Bondy, M. L. e al. B ain umo epidemiology: consensus om he B ain Tumo Epidemiology Conso ium. Cance 113, 1953–1968 (2008). 6. Ma hews, J. D. e al. Cance isk in 680,000 people exposed o compu ed omog aphy scans in childhood o adolescence: da a linkage s udy o 11 million Aus alians. BMJ 346, 2360 (2013). 7. In e phone S udy G oup. B ain umou isk in ela ion o mobile elephone use: esul s o he INTERPHONE in e na ional case- con ol s udy. In . J. Epidemiol. 39, 675–694 (2010). 8. U.S. Depa men o Heal h and Human Se ices. The Heal h Consequences o Smoking—50 Yea s o P og ess: A Repo o he Su geon Gene al. (Cen e s o Disease Con ol and P e en ion, Na ional Cen e o Ch onic Disease P e en ion and Heal h P omo ion, O ice on Smoking and Heal h 2014). 9. WHO handbook on indoo adon: a public heal h pe spec i e. h p://apps.who.in /i is/bi s eam/10665/44149/1/9789241547673_eng. pd Da e o access: 01/04/2017 (Wo ld Heal h O ganiza ion 2009). 10. Man made mine al ib es and adon:… iews and expe s opinions o an IARC Wo king G oup on he E alua ion o Ca cinogenic Risks o Humans, which me in Lyon 16–23 June 1987 (1988). 11. IARC monog aphs on he e alua ion o ca cinogenic isks o humans, olume 100 D, adia ion: his publica ion ep esen s he iews and expe opinions o an IARC Wo king G oup on he E alua ion o Ca cinogenic Risks o Humans, which me in Lyon, 02–09 June 2009 (IARC, 2012). www.na u e.com/scien i ic epo s/ 7 Scien i ic RepoR s | 7: 3595 | DOI:10.1038/s41598-017-03938-9 12. Kendall, G. M. & Smi h, T. J. Doses o o gans and issues om adon and i s decay p oduc s. J. Radiol. P o . O . J. Soc. Radiol. P o . 22, 389–406 (2002). 13. B äune , E. V. e al. Residen ial adon and b ain umou incidence in a Danish coho . PloS One 8, e74435 (2013). 14. Tu ne , M. C. e al. Radon and non espi a o y mo ali y in he Ame ican Cance Socie y coho . Am. J. Epidemiol 176, 808–814 (2012). 15. Ba os-Dios, J. M., Ba ei o, M. A., Ruano-Ra ina, A. & Figuei as, A. Exposu e o esiden ial adon and lung cance in Spain: a popula ion-based case-con ol s udy. Am. J. Epidemiol. 156, 548–555 (2002). 16. Ba os-Dios, J. M., Ruano-Ra ina, A., Gas elu-I u i, J. & Figuei as, A. Fac o s unde lying esiden ial adon concen a ion: esul s om Galicia, Spain. En i on. Res. 103, 185–190 (2007). 17. Ba os-Dios, J. M. e al. Residen ial adon exposu e, his ologic ypes, and lung cance isk. A case-con ol s udy in Galicia, Spain. Cance Epidemiol. Bioma k. P e . Publ. Am. Assoc. Cance Res. Cosponso ed Am. Soc. P e . Oncol. 21, 951–958 (2012). 18. Ruano-Ra ina, A., A agonés, N., Pé ez-Ríos, M., López-Aben e, G. & Ba os-Dios, J. M. Residen ial adon exposu e and esophageal cance . An ecological s udy om an a ea wi h high indoo adon concen a ion (Galicia, Spain). In . J. Radia . Biol. 90, 299–305 (2014). 19. To es-Du án, M. e al. Lung cance in ne e -smoke s: a case-con ol s udy in a adon-p one a ea (Galicia, Spain). Eu . Respi . J. 44, 994–1001 (2014). 20. Va gas, A. & O ega, X. In luence o en i onmen al changes on in eg a ing adon de ec o s: esul s o an in e compa ison exe cise. Radia . P o . Dosime y 123, 529–536 (2007). 21. Gu ie ez-Villanue a, J. L. e al. In e compa ison exe cise on ex e nal gamma dose a e unde ield condi ions a he labo a o y o na u al adia ion (Saelices el Chico, Spain). Radia . P o . Dosime y 155, 459–466 (2013). 22. Besag, J., Yo k, J. Mollie, A. Bayesian image es o a ion, wi h wo applica ions in spa ial s a is ics. 43, 1–59 (1991). 23. Rue, H., Ma ino, S. & Chopin, N. App oxima e Bayesian in e ence o la en Gaussian models by using in eg a ed nes ed Laplace app oxima ions. J. R. S a . Soc. Se . B S a . Me hodol. 71, 319–392 (2009). 24. Hess, C. T., Wei enbach, C. V. & No on, S. A. En i onmen al adon and cance co ela ions in Maine. Heal h Phys. 45, 339–348 (1983). 25. Ti ma che, M., Raphalen, A., Allin, F., Chameaud, J. & B edon, P. Mo ali y o a coho o F ench u anium mine s exposed o ela i ely low adon concen a ions. B . J. Cance 67, 1090–1097 (1993). 26. Vacquie , B. e al. Mo ali y isk in he F ench coho o u anium mine s: ex ended ollow-up 1946-1999. Occup. En i on. Med. 65, 597–604 (2008). 27. Vacquie , B. e al. The in luence o mul iple ypes o occupa ional exposu e o adon, gamma ays and long-li ed adionuclides on mo ali y isk in he F ench ‘pos -55’ sub-coho o u anium mine s: 1956–1999. Radia . Res. 176, 796–806 (2011). 28. K euze , M., Walsh, L., Schnelze , M., Tschense, A. & G osche, B. Radon and isk o ex apulmona y cance s: esul s o he Ge man u anium mine s’ coho s udy, 1960–2003. B . J. Cance 99, 1946–1953 (2008). 29. K euze , M. e al. Radon and isk o dea h om cance and ca dio ascula diseases in he Ge man u anium mine s coho s udy: ollow-up 1946–2003. Radia . En i on. Biophys. 49, 177–185 (2010). 30. Schubaue -Be igan, M. K., Daniels, R. D. & Pinke on, L. E. Radon exposu e and mo ali y among whi e and Ame ican Indian u anium mine s: an upda e o he Colo ado Pla eau coho . Am. J. Epidemiol. 169, 718–730 (2009). 31. Da by, S. C. e al. Radon and cance s o he han lung cance in unde g ound mine s: a collabo a i e analysis o 11 s udies. J. Na l. Cance Ins . 87, 378–384 (1995). 32. Da by, S. C., Rad o d, E. P. & Whi ley, E. Radon exposu e and cance s o he han lung cance in Swedish i on mine s. En i on. Heal h Pe spec . 103(Suppl 2), 45–47 (1995). 33. Ma, J., Yoneha a, H., Ikebuchi, M. & Aoyama, T. E ec o adon exposu e on supe oxide dismu ase (SOD) ac i i y in a s. J. Radia . Res. (Tokyo) 37, 12–19 (1996). 34. San , M. e al. Su i al o Eu opean pa ien s wi h cen al ne ous sys em umo s. In . J. Cance 131, 173–185 (2012). 35. Ro hman, K. J. Mode n epidemiology. (Li le, B own 1986). 36. B eslow, N. E. & Day, N. E. S a is ical me hods in cance esea ch. Bd. 2: The design and analysis o coho s udies. (In e na ional Agency o Resea ch on Cance 1987). 37. Rosenbaum, P. R. & Rubin, D. B. Di icul ies wi h eg ession analyses o age-adjus ed a es. Biome ics 40, 437–443 (1984). Acknowledgemen s Pa o his wo k has been pe o med du ing a Fulb igh g an o Senio Resea che s awa ded o Albe o Ruano- Ra ina a B own Uni e si y (P o idence, Rhode Island, USA) (REF PRX14/00365). This g an is awa ded by he US Depa men o S a e in collabo a ion wi h he Spanish Minis y o Educa ion. Au ho Con ibu ions A.R.R. and J.M.B.D. w o e he i s d a s o he manusc ip . M.P.L., M.P.R. and G.L.A. execu ed he s a is ical analysis. K.T.K., N.A. and G.L.A. ook pa in he li e a u e sea ch and help in e p e ing he esul s. All au ho s ha e app o ed he inal e sion o he manusc ip and ake esponsibili y o i s con en . All au ho s con ibu ed in ellec ually o he manusc ip . Addi ional In o ma ion Compe ing In e es s: The au ho s decla e ha hey ha e no compe ing in e es s. Publishe 's no 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. Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as 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 e- a i e Commons license, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons license, unless indica ed o he wise in a c edi line o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons license and you in ended use is no pe - mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/. © The Au ho (s) 2017