Adjuvant bisphosphonate treatment in early breast cancer: meta-analyses of individual patient data from randomised trials

A icles www. helance .com Vol 386 Oc obe 3, 2015 1353 Adju an bisphosphona e ea men in ea ly b eas cance : me a-analyses o indi idual pa ien da a om andomised ials Ea ly B eas Cance T ialis s’ Collabo a i e G oup (EBCTCG)* Summa y Backg ound Bisphosphona es ha e p o ound eff ec s on bone physiology, and could modi y he p ocess o me as asis. We unde ook collabo a i e me a-analyses o cla i y he isks and benefi s o adju an bisphosphona e ea men in b eas cance . Me hods We sough indi idual pa ien da a om all uncon ounded ials in ea ly b eas cance ha andomised be ween bisphosphona e and con ol. P ima y ou comes we e ecu ence, dis an ecu ence, and b eas cance mo ali y. P ima y subg oup in es iga ions we e si e o fi s dis an ecu ence (bone o o he ), menopausal s a us (pos menopausal [combining na u al and a ifi cial] o no ), and bisphosphona e class (aminobisphosphona e [eg, zoled onic acid, iband ona e, pamid ona e] o o he [ie, clod ona e]). In en ion- o- ea log- ank me hods yielded bisphosphona e e sus con ol fi s -e en a e a ios (RRs). Findings We ecei ed da a on 18 766 women (18 206 [97%] in ials o 2–5 yea s o bisphosphona e) wi h median ollow-up 5·6 woman-yea s, 3453 fi s ecu ences, and 2106 subsequen dea hs. O e all, he educ ions in ecu ence (RR 0·94, 95% CI 0·87–1·01; 2p=0·08), dis an ecu ence (0·92, 0·85–0·99; 2p=0·03), and b eas cance mo ali y (0·91, 0·83–0·99; 2p=0·04) we e o only bo de line signifi cance, bu he educ ion in bone ecu ence was mo e defi ni e (0·83, 0·73–0·94; 2p=0·004). Among p emenopausal women, ea men had no appa en eff ec on any ou come, bu among 11 767 pos menopausal women i p oduced highly signifi can educ ions in ecu ence (RR 0·86, 95% CI 0·78–0·94; 2p=0·002), dis an ecu ence (0·82, 0·74–0·92; 2p=0·0003), bone ecu ence (0·72, 0·60–0·86; 2p=0·0002), and b eas cance mo ali y (0·82, 0·73–0·93; 2p=0·002). E en o bone ecu ence, howe e , he he e ogenei y o benefi was ba ely signifi can by menopausal s a us (2p=0·06 o end wi h menopausal s a us) o age (2p=0·03), and i was non-signifi can by bisphosphona e class, ea men schedule, oes ogen ecep o s a us, nodes, umou g ade, o concomi an chemo he apy. No diff e ences we e seen in non-b eas cance mo ali y. Bone ac u es we e educed (RR 0·85, 95% CI 0·75–0·97; 2p=0·02). In e p e a ion Adju an bisphosphona es educe he a e o b eas cance ecu ence in he bone and imp o e b eas cance su i al, bu he e is defi ni e benefi only in women who we e pos menopausal when ea men began. Funding Cance Resea ch UK, Medical Resea ch Council. Copy igh © Ea ly B eas Cance T ialis s’ Collabo a i e G oup (EBCTCG). Open Access a icle dis ibu ed unde he e ms o CC BY. In oduc ion Ci cula ing umou cells can be a ac ed o su aces wi hin he bone whe e hey can displace haemopoie ic s em cells and bind o he os eoblas ic niche.1 These dissemina ed malignan cells can emain quiescen o yea s. Then, o easons ha a e no well unde s ood, hey can exi his do man s a e, s a o p oli e a e, and es ablish mac o- me as ases in he bone o elsewhe e.2,3 Bisphosphona es ha e p o ound eff ec s on os eoclas s, and aff ec T-cell unc ion, so could also be eff ec i e as adju an ea men s, pa icula ly in p e en ing o delaying bone ecu ence.4–6 Fo his eason, and because bisphosphona es can be added o he a oma ase inhibi o ea men o pos menopausal b eas cance o es ic ad e se skele al eff ec s o oes ogen dep i a ion, eliable e idence is needed abou he eff ec s o bisphosphona es on b eas cance ou comes. Imp o emen s in bone-me as asis- ee su i al, disease- ee su i al, and o e all su i al in women wi h ea ly b eas cance ha e been epo ed in some adju an ials o o al clod ona e7,8 o o in a enous zoled onic acid.9,10 Howe e , in o he ials o adju an bisphos phona es no signifi can benefi s we e seen in analyses ha included all andomised pa ien s, al hough bo h planned and explo a o y subse analyses sugges ed benefi s ei he in pos menopausal women11 o in olde women.12,13 This led o he hypo hesis11–13 ha ea men is o benefi only in pa ien s wi h low concen a ions o ep oduc i e ho mones (ie, hose who a e pos menopausal o unde going o a ian supp ession he apy).14–16 To help cla i y whe he adju an bisphosphona es educe he isk o bone and o he me as ases, and whe he meno- pausal s a us aff ec s effi cacy, we unde ook collab o a i e Lance 2015; 386: 1353–61 This online publica ion has been co ec ed. The co ec ed e sion fi s appea ed a helance .com on Dec 31, 2015 Published Online July 24, 2015 h p://dx.doi.o g/10.1016/ S0140-6736(15)60908-4 See Commen page 1319 *Full lis o membe s a ailable a h p://www.c su.ox.ac.uk/ esea ch/me a- ials/ebc cg/ ebc cg-page Co espondence o: EBCTCG Sec e a ia , Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Richa d Doll Building, Ox o d OX3 7LF, UK [email p o ec ed] A icles 1354 www. helance .com Vol 386 Oc obe 3, 2015 me a-analyses o all uncon ounded andomised ials ha compa ed b eas cance ou comes in hose alloca ed adju an bisphosphona e e sus hose who we e no . Me hods Iden ifi ca ion o s udies and collec ion o da a The me hods o iden i ying ials, seeking collabo a ion, da a collec ion, colla ion, checking, and p es en a ion a e as in p e ious Ea ly B eas Cance T ialis s’ Collabo a i e G oup (EBCTCG) epo s.17–19 T ials we e eligible i hey began be o e 2008 and andomly assigned women be- ween a bisphosphona e o any ype, dose, and schedule e sus a con ol g oup (open label o placebo) wi h no bisphosphona e, all o he ea men s being simila in bo h g oups. In o ma ion was sough du ing 2012–14 o each indi idual pa ien on da e o andomisa ion, alloca ed ea - men , age, menopausal s a us, umou diame e , g ade, sp ead o loco egional lymph nodes, HER2 and oes ogen and p oges e one ecep o (ER/PR) s a us, da es and si es o any b eas cance ecu ence, o he second p i ma y cance , bone ac u e, and he da e and cause o dea h. The main defi ni ions and analysis me hods a e hose used in p e ious EBCTCG epo s,17–19 bu wi h some amendmen s ha efl ec he po en ial eff ec o bisphosphona es on bone me as ases (appendix). Ou comes The p e-defi ned cop ima y endpoin s we e any ecu ence o b eas cance (dis an , loco egional, o new p ima y in he con ala e al b eas ); dis an ecu ence, igno ing any p e ious loco egional o con ala e al ecu ence; and b eas cance mo ali y (es ima ed by log- ank sub ac ion, as in p e ious EBCTCG epo s18,19). Seconda y ou comes we e all-cause mo ali y; dea h wi hou ecu ence; bone ecu ence as he fi s dis an ecu ence (wi h o wi hou concu en o he ecu ence); o he fi s (ex askele al) dis an ecu ence (wi h all analyses o dis an ecu ence igno ing any p e ious loco egional o con ala e al ecu ence); loco egional ecu ence as fi s e en (ipsila e al b eas , ches wall, o loco egional lymph nodes); con ala e al new p ima y b eas cance as fi s e en ; and any bone ac u es. S a is ical analyses Time- o-e en analyses we e s a ifi ed by age, ER s a us, nodal s a us, and ial. Wi hin each s a um, hey compa ed all hose alloca ed bisphosphona e e sus all hose alloca ed con ol, ega dless o ea men compliance (yielding in en ion- o- ea analyses). Log- ank s a is ics we e used o assess he eff ec s (bisphosphona e s con ol) on a ious ou comes, and, o each, o es ima e fi s -e en a e a ios (RRs) and hei CIs. We did s a is ical analyses using EBCTCG in-house Fo an p og ams. P e-specifi ed p ima y subg oup in es iga ions we e o si e o fi s dis an ecu ence (bone, o he ), menopausal s a us (p emenopausal, pe imenopausal, pos menopausal [na u al o induced, ei he po en ially e e sibly, using lu einising ho mone- eleasing ho mone analogues, o pe manen ly by oopho ec omy] o , i menopausal s a us was una ailable, yea s o age, g ouped as <45, 45–54, ≥55 yea s), and class o bisphosphona e (amino bisphos- phona e [zoled onic acid, iband ona e, pamid ona e, ised ona e, alend ona e], o he [clod ona e]). Explo a o y in es iga ions we e unde aken o po en ial in e ac ions be ween ea men effi cacy and ER s a us, nodal s a us, his ological g ade, use o no o adju an chemo he apy, and ollow-up pe iod. I app op ia e, es s compa ing eff ec s in diff e en subg oups we e o end a he han he e ogenei y. We p e-specifi ed ha compa isons o ea men effi cacy wi hin subg oups would exclude local and con ala e al ecu ence i he p io hypo hesis ha bisphosphona es would educe dis an bu no local o con ala e al ecu ence was es ablished om analyses o he o e all esul s in all andomised pa ien s. As bone ecu ence was he only ype o ecu ence signifi can ly educed by bisphosphona es we used his ins ead as he p ima y endpoin o subg oup compa isons, bu he appendix includes subg oup analyses o any dis an ecu ence. Because he ABCSG-129 and AZURE11,14 ials had helped gene a e he hypo hesis o he ele ance o menopausal s a us o he eff ec s o ea men , we p o ide sensi i i y analyses o his hypo hesis ha ea ed hese ials as hypo hesis-gene a ing, wi h he emaining ials hypo hesis- es ing. The policy on da a sha ing om his s udy is a ailable online. See Online o appendix S udies iden ifi ed S udies wi h da a ecei ed T ials (n) Pa ien s (n) T ials (n) Pa ien s (n) %* Yea s† Up o 1 yea o ea men <1 yea clod ona e 2 120 1 72 60% 0·5 <1 yea aminobisphosphona e 2 208 1 40 19% 0·1 1 yea aminobisphosphona e 7 1088 3 448 41% 1·0 To al o ≤1 yea o ea men 11 1416 5 560 40% 0·9 2–5 yea s o ea men 2 yea s clod ona e 4 3978 3 3912 98% 2·0 3–5 yea s clod ona e 1 1069 1 1069 100% 3·0 2 yea s aminobisphosphona e 10 3654 8 3514 96% 2·0 3–5 yea s aminobisphosphona e 12 11 910‡ 9 9711 82%‡ 4·5 To al o 2–5 yea s o ea men 27 20 611‡ 21 18 206 88%‡ 3·5 Any clod ona e egimen 7 5167 5 5053 98% 2·6 Any aminobisphosphona e§ 31 16 860‡ 21 13 713 81%‡ 3·8 To al, all egimens 38 22 027‡ 26 18 766 85%‡ 3·4 *Numbe o pa ien s wi h da a ecei ed as a pe cen age o all andomised pa ien s in iden ifi ed s udies. †Mean scheduled ea men du a ion (weigh ed in p opo ion o numbe s o pa ien s wi h da a ecei ed). ‡Includes wo ials (2116 pa ien s) s ill in p og ess; excluding hese, he o al wi h da a ecei ed is 94%. §The aminobisphosphona es in hese ials we e zoled onic acid (9290 pa ien s wi h da a ecei ed, 1582 ecu ences [46% o all ecu ences]), iband ona e (3072 pa ien s, 380 ecu ences [11%]), pamid ona e (953 pa ien s, 473 ecu ences [14%]), ised ona e (398 pa ien s, 13 ecu ences [0·4%]), and alend ona e (no ials wi h da a ecei ed); he only non-aminobisphosphona e in hese ials was clod ona e (5053 pa ien s, 1005 [29%] ecu ences). Table: Numbe s o uncon ounded andomised ials o an adju an bisphosphona e iden ifi ed, and numbe s wi h da a ecei ed, by du a ion and ype o bisphosphona e ea men Fo he CTSU policy on da a sha ing see h p://www.c su.ox. ac.uk/ esea ch/da a-access- policies/da a-access-and- sha ing-policy/ iew A icles www. helance .com Vol 386 Oc obe 3, 2015 1355 Role o he unding sou ce The unde s o he s udy had no ole in s udy design, da a collec ion, da a analysis, da a in e p e a ion, o w i ing o he epo . The w i ing commi ee had ull access o all he da a in he s udy and had fi nal esponsibili y o he decision o submi o publica ion. Resul s Indi idual pa ien da ase s we e p o ided o 26 ials wi h 18 766 pa icipan s, 97% o all 19 291 women in he 32 comple ed ials ha eco ded ecu ence da a ( able, appendix). In ou o he ials (620 women) ecu ence was no eco ded, and om he wo ongoing ials (2116 women) ou come da a canno ye be p o ided. Mean scheduled ea men du a ion was 3·4 yea s; 18 206 (97%) o 18 766 pa icipan s we e in ials o 2–5 yea s o ea men . Median ollow-up was 5·6 woman-yea s (IQR 3·7–8·0). 3453 women had a ecu ence, a e which 2106 died. Recu ence a es we e sligh ly lowe wi h han wi hou bisphosphona es, bu his was no signifi can in analyses ha included all 18 766 women (RR 0·94, 95% CI 0·87–1·01; 2p=0·08; fi gu e 1). Howe e , he e was a bo de line signifi can educ ion in he isk o dis an ecu ence, igno ing any p e ious local o con ala e al Figu e 1: Recu ence by si e and b eas cance mo ali y in 24 ials o bisphosphona e e sus no bisphosphona e (con ol) Kaplan-Meie g aphs showing eff ec s o ea men alloca ion on 10-yea ou comes in all 18 766 pa ien s. (A) Any ecu ence. (B) Dis an ecu ence. (C) Bone ecu ence. (D) B eas cance mo ali y. O–E=obse ed minus expec ed. V= a iance o O–E. RR= a e a io (exp[{O–E}/V]). E o ba s a e SE. Recu ence a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 18 766 women RR 0·94 (95% CI 0·87–1·01) Log- ank 2p=0·08 10-yea gain 1·1% (95% CI –0·7 o 2·9) Con ol 25·9% Bisphosphona e 24·9% Yea s 0–4 3·63 (1415/38 979) 3·85 (1384/35 984) 0·93 (0·85–1·01) –41·9/593·7 Yea s 5–9 2·34 (308/13 171) 2·36 (314/13 322) 0·98 (0·81–1·14) –3·1/139·4 Yea s ≥10 0·87 (15/1724) 0·95 (17/1790) 0·72 (CI 0·01–1·43) –1·8/5·5 17·4% 16·3% 0 10 20 30 40 50 Recu ence (%) A Dis an ecu ence a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 18 766 women RR 0·92 (95% CI 0·85–0·99) Log- ank 2p=0·03 10-yea gain 1·4% (95% CI –0·3 o 3·1) Con ol 21·8% Bisphosphona e 20·4% Yea s 0–4 2·97 (1173/39 559) 3·20 (1170/36 571) 0·91 (0·83–0·99) –47·5/499·9 Yea s 5–9 1·84 (253/13 746) 1·82 (253/13 931) 0·99 (0·81–1·18) –0·7/114·3 Yea s ≥10 0·67 (13/1932) 1·08 (21/1941) 0·47 (0·21–1·03) –4·5/5·9 14·8% 13·5% 0 10 20 30 40 50 Dis an ecu ence (%) B Bone ecu ence a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 18 766 women RR 0·83 (95% CI 0·73–0·94) Log- ank 2p=0·004 10-yea gain 1·1% (95% CI –0·1 o 2·3) Con ol 9·0% Bisphosphona e 7·8% 0 5 10 Yea s 0–4 0·99 (391/39 559) 1·21 (441/36 571) 0·79 (0·66–0·92) –45·0/189·5 Yea s 5–9 0·76 (104/13 746) 0·71 (99/13 031) 1·02 (0·73–1·31) 0·8/46·8 Yea s ≥10 0·10 (2/1932) 0·10 (2/1941) 0·61 (0·08–2·25) –0·4/0·9 Yea s 5·9% 4·7% 0 10 20 30 40 50 Bone ecu ence (%) C Dea h a es (%/yea : o al a e minus a e in women wi hou ecu ences) and log- ank s a is i Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 18 766 women RR 0·91 (95% CI 0·83–0·99) Log- ank 2p=0·04 10-yea gain 1·7% (95% CI 0·0 o 3·5) Con ol 18·4% Bisphosphona e 16·6% 0 5 10 Yea s 0–4 1·83 (1·70–1·97) 1·98 (1·84–2·12) 0·91 (0·81–1·01) –30·5/321·7 Yea s 5–9 1·81 (1·59–2·03) 1·97 (1·75–2·20) 0·92 (0·75–1·10) –9·5/121·0 Yea s ≥10 1·21 (0·72–1·69) 1·69 (1·12–2·25) 0·66 (0·18–1·15) –4·5/10·9 Yea s 9·7% 8·9% 0 10 20 30 40 50 B eas cance mo ali y (%) D A icles 1356 www. helance .com Vol 386 Oc obe 3, 2015 Figu e 2: Mul iple subg oup analyses o eff ec s on bone ecu ence in ials o bisphosphona e e sus no bisphosphona e (con ol) Resul s a e plo ed as black squa es wi h ho izon al lines ha deno e 99% a he han 95% CIs o allow o mul iple hypo hesis es ing. To al is plo ed as a whi e diamond ha deno es 95% CI. ER=oes ogen ecep o . O–E=obse ed minus expec ed. Ca ego y Ra e a io (CI)E en s/women Alloca ed bisphosphona e Alloca ed con ol (a) Age, yea s ( end χ2 1=4·9; 2p=0·03) <45 45–54 55–69 ≥70 Age unknown (b) Menopausal s a us ( end χ2 1=3·5; 2p=0·06) P emenopausal Pe imenopausal Pos menopausal (c) ER s a us (χ2 1=0·6; 2p=0·4) ER nega i e ER unknown ER posi i e (d) Nodal s a us ( end χ2 1=0·5; 2p=0·5) N0/N– N1–3 N4+ N o he /unknown (e) Tumou g ade ( end χ2 1=0·4; 2p=0·5) Well diffe en ia ed Mode a ely diffe en ia ed Poo ly diffe en ia ed G ade unknown ( ) Bisphosphona e ype (χ2 4=0·3; 2p=0·6) Clod ona e Aminobisphosphona e (g) Bisphosphona e (χ2 4=5·9; 2p=0·21) Clod ona e Zoled onic acid Pamid ona e Iband ona e Rised ona e Alend ona e (h) Bisphosphona e dose (χ2 1=0·4; 2p=0·5) Mo e in ensi e Low in ensi y (i) Bisphosphona e du a ion ( end χ2 1=0·2; 2p=0·7) <1 yea 2 yea s >2 yea s (j) Chemo he apy (χ2 1=0·3; 2p=0·6) Absence P esence (k) Follow-up pe iod, yea s ( end χ2 1=2·5; 2p=0·11) 0–1 2–4 5–9 ≥10 To al 164/2475 (6·6%) 152/3532 (4·3%) 168/3314 (5·1%) 13/531 (2·4%) 0/4 (0·0%) 151/2141 (7·1%) 173/3224 (5·4%) 196/3022 (6·5%) 22/521 (4·2%) 0/2 (0·0%) –0·3 –14·2 –25·1 –5·1 71·3 74·3 84·4 7·1 217/3296 (6·6%) 28/461 (6·1%) 252/6099 (4·1%) 212/2875 (7·4%) 19/367 (5·2%) 311/5668 (5·5%) –7·9 2·0 –42·1 96·4 8·8 128·0 107/1964 (5·4%) 42/637 (6·6%) 348/7255 (4·8%) 135/1684 (8·0%) 47/690 (6·8%) 360/6536 (5·5%) –15·7 1·2 –26·9 56·4 20·9 169·1 1·00 (0·79–1·26) 0·83 (0·61–1·11) 0·74 (0·56–0·98) 0·49 (0·19–1·29) 0·92 (0·71–1·20) 0·72 (0·57–0·90) 0·76 (0·54–1·07) 1·06 (0·60–1·86) 0·85 (0·70–1·04) 4/277 (1·4%) 169/3081 (5·5%) 324/6498 (5·0%) 4/283 (1·4%) 154/2091 (1·4%) 384/6536 (5·9%) 0·8 –18·5 –26·9 1·7 68·6 166·9 0·76 (0·56–1·04) 0·85 (0·70–1·04) 39/1616 (2·4%) 458/8240 (5·6%) 53/1616 (3·3%) 489/7294 (6·7%) –6·3 –38·3 21·0 216·2 0·74 (0·48–1·14) 0·84 (0·70–1·00) 0·829 (0·730–0·941) 2p=0·004 70/2638 (2·7%) 225/4323 (5·2%) 160/1205 (13·3%) 42/1690 (2·5%) 68/2631 (2·6%) 231/3352 (6·9%) 183/1190 (15·4%) 60/1737 (3·5%) 0·6 –24·0 –14·3 –6·9 32·3 104·1 76·1 24·6 1·02 (0·72–1·44) 0·79 (0·62–1·02) 0·83 (0·62–1·11) 0·76 (0·45–1·27) 24/877 (2·7%) 196/3667 (5·3%) 156/2801 (5·6%) 121/2511 (4·8%) 26/793 (3·3%) 203/3249 (6·2%) 174/2326 (7·5%) 139/2542 (5·5%) –1·6 –11·7 –17·6 –4·4 10·9 94·4 76·8 61·4 0·88 (0·68–1·15) 0·80 (0·59–1·07) 0·93 (0·67–1·29) 173/9856 (1·8%) 218/8445 (2·6%) 104/5711 (1·8%) 2/706 (0·3%) 497/9856 (5·0%) 204/8910 (2·3%) 237/7609 (3·1%) 99/5614 (1·8%) 2/758 (0·3%) 542/8910 (6·1%) –25·0 –20·0 0·8 –0·4 –44·6 85·0 104·6 46·8 0·9 237·1 0·75 (0·56–0·99) 0·83 (0·64–1·06) 1·02 (0·73–1·31) 139/2514 (5·5%) 200/4642 (4·3%) 80/460 (17·4%) 78/2040 (3·8%) 0/200 (0·0%) (no da a) 165/2539 (6·5%) 250/4648 (5·4%) 76/493 (15·4%) 49/1032 (4·7%) 2/198 (1·0%) –16·7 –24·1 5·1 –8·0 –0·9 67·5 108·7 33·1 27·3 0·5 0·78 (0·57–1·07) 0·80 (0·63–1·03) 1·17 (0·83–1·64) 0·75 (0·46–1·22) 434/7040 (6·2%) 63/2816 (2·2%) 457/6089 (7·5%) 85/2821 (3·0%) –34·5 –10·1 201·7 35·5 0·84 (0·70–1·01) 0·75 (0·49–1·16) 139/2514 (5·5%) 358/7342 (4·9%) 165/2539 (6·5%) 377/6371 (5·9%) –16·7 –27·9 67·5 169·7 0·78 (0·57–1·07) 0·85 (0·70–1·03) Bisphosphona e e en s Log- ank O–E Va iance o O–E 1·00·50 1·5 2·0 Con ol be e Bisphosphona e be e Ra io o annual e en a es bisphosphona e : con ol A icles www. helance .com Vol 386 Oc obe 3, 2015 1357 b eas ecu ence (10-yea isk 20·4% bisphosphona e s 21·8% con ol; RR 0·92, 95% CI 0·85–0·99; 2p=0·03; fi gu e 1), whe eas he e was no signifi can eff ec on he incidence o local ecu ence as fi s e en (RR 1·10, 0·94–1·28; 2p=0·25; appendix) o o con ala e al b eas cance as fi s e en (RR 0·96, 0·74–1·25; 2p=0·79). The g ea e effi cacy o bis phosphona es in p e en ing dis an ecu ence han in p e en ing o he (local o con ala e al) b eas cance ecu ence was signifi can ( es o in e ac ion 2p=0·01). The eff ec on dis an ecu ence was mainly because o a educ ion in bone ecu ence (10-yea isk 7·8% s 9·0%; RR 0·83, 95% CI 0·73–0·94; 2p=0·004; fi gu e 1). The e was signifi can ly (p=0·04) g ea e eff ec on bone ecu ence han on o he fi s dis an ecu ence (RR 0·98, 95% CI 0·89–1·08; 2p=0·69; appendix), al hough his appa en lack o effi cacy could be pa ly because delay o bone ecu ence wi h bisphosphona e in a woman who would o he wise ha e had bo h bone and o he dis an ecu ence allowed he o he ecu ence o be he fi s e en . B eas cance mo ali y was bo de line signifi can ly lowe in pa ien s alloca ed bisphosphona e han con ol (10-yea isk 16·6% s 18·4%; RR 0·91, 95% CI 0·83–0·99; 2p=0·04; fi gu e 1), and all-cause mo ali y was simila ly educed (10-yea isk 20·8% s 22·3%; RR 0·92, 0·85–1·00; 2p=0·06; appendix). O 2607 dea hs om any cause, 501 (19%) we e in ecu ence- ee women; his non-b eas cance mo ali y appea ed o be unaff ec ed by he ea men alloca ion (RR 0·99, 95% CI 0·82–1·19; 2p=0·91). We did many subg oup analyses o in es iga e he eff ec s o bisphosphona es on any ecu ence, dis an ecu ence, bone ecu ence, and b eas cance mo ali y (appendix). In he o e all analyses, among all 18 766 women, he clea es e idence o eff ec o bis phosphona es was, as an icipa ed, on bone ecu ence, so he mos in o ma i e subg oup analyses should ela e o his endpoin (fi gu e 2). The effi cacy o bisphosphona es in educing bone ecu ence appea ed o be g ea e in olde women (2p=0·03 o end wi h age in ea men eff ec ) o , simila ly, in pos menopausal women (2p=0·06 o end wi h menopausal s a us). As menopausal s a us and age a e closely co ela ed, we canno de e mine eliably which is mo e ele an (appendix). Among he 4616 women younge han 45 yea s, bone ecu ence appea ed o be unaff ec ed by he ea men alloca ion (RR 1·00, 95% CI 0·79–1·26; 2p=0·97), bu among he 7388 women 55 yea s o olde he e was a highly signifi can ea men eff ec (RR 0·72, 0·59–0·88; 2p=0·002). Sensi i i y analyses o he possible ele ance o age and menopausal s a us ha omi ed he hypo hesis-gene a ing ABCSG-129 and AZURE11,14 s udies s ill showed signifi can (2p=0·004) benefi only in pos menopausal women (appendix). As was he case o bone ecu ence, he educ ions in any dis an ecu ence wi h bisphosphona e we e also signifi can ly g ea e in olde women (2p=0·003 o end wi h age) and pos menopausal women (2p=0·01). None o he o he subg oup analyses o bone ecu ence in fi gu e 2 e ealed any signifi can e idence o he e ogenei y o benefi by umou ype (o o o he b eas cance ou comes; appendix). Al hough he benefi appea ed somewha la ge in ER-nega i e han ER-posi i e disease and in node-posi i e han node-nega i e umou s, his appa en he e ogenei y o ea men eff ec did no app oach signifi cance and could be a chance fi nding. Likewise, he e was no signifi can he e ogenei y be ween he appa en eff ec s on bone ecu ence o he diff e en bisphosphona e egimens es ed in hese ials. Fo his ou come, he benefi s o he non- amino bisphosphona e (clod ona e, n=5053) and o he wo mos widely es ed aminobisphosphona es (zoled onic acid, n=9290, and iband ona e, n=3072) appea ed simila , bu he e was no appa en benefi in he smalle o al pamid ona e g oup (n=953). Fo bone ecu ence, he benefi s appea ed o be simila in ials o low-in ensi y an i-os eopo osis schedules (eg, 6-mon hly in a enous zoled onic acid) and in ials o mo e in ensi e schedules such as hose app o ed o use in me as a ic bone disease (eg, mon hly zoled onic acid, daily o al iband ona e, o daily o al clod ona e). Likewise, he a e age eff ec appea ed simila in ials ha es ed diff e en du a ions o ea men ( ials o 2 yea s bisphosphona e s none: RR 0·76, 95% CI 0·60–0·97; 2p=0·026; ials o 3–5 yea s bisphosphona e s none: RR 0·85, 0·73–0·99; 2p=0·037; fi gu e 2), and in he p esence o absence o chemo he apy. The e we e signifi can educ ions in bone ecu ence du ing yea s 0–1 and yea s 2–4 a e andomisa ion bu he e appea ed o be no u he educ ion he ea e . Again, hough, his dec ease in ea men eff ec o e ime was no signifi can ( end 2p=0·11), pe haps because he e is hus a only limi ed ollow-up a e he fi s 5 yea s. The 10-yea disease ou comes o p emenopausal and pos menopausal women sepa a ely a e summa ised in fi gu e 3 and he appendix. In p emenopausal women, ea men appea ed o ha e li le eff ec on bone me as ases o b eas cance mo ali y, whe eas in pos menopausal women i p oduced highly signifi can educ ions in ecu ence (RR 0·86, 95% CI 0·78–0·94; 2p=0·002; appendix), dis an ecu ence (RR 0·82, 0·74–0·92; 2p=0·0003; appendix), bone ecu ence (RR 0·72, 0·60–0·86; 2p=0·0002), and b eas cance mo ali y (RR 0·82, 0·73–0·93; 2p=0·002). In bo h menopausal subg oups, a es o fi s dis an ecu ence a si es o he han bone appea ed o be unaff ec ed by ea men . In he pos menopausal subg oup, o bone ecu ence he absolu e gain om ea men was 2·2% (95% CI 0·6–3·8) (10-yea isks 6·6% s 8·8%; RR 0·72, 95% CI 0·60–0·86; 2p=0·0002), whe eas o b eas cance mo ali y he absolu e gain was 3·3% (95% CI 0·8–5·7) (10-yea isks 14·7% s 18·0%; RR 0·82, 0·73–0·93; 2p=0·002). To enhance s a is ical powe , he mul iple subg oup analyses o bone ecu ence (fi gu e 2) and he co esponding analyses o o he ou comes (appendix) can A icles 1358 www. helance .com Vol 386 Oc obe 3, 2015 Figu e 3: Main ou comes in p emenopausal (excluding pe imenopausal) and in pos menopausal women in ials o bisphosphona e e sus no bisphosphona e (con ol) Kaplan-Meie g aphs showing he eff ec s o ea men alloca ion on 10-yea b eas cance ou comes. (A) P emenopausal and (B) pos menopausal bone ecu ence. (C) P emenopausal and (D) pos menopausal dis an ecu ence ou side he bone. (E) P emenopausal and (F) pos menopausal b eas cance mo ali y. O-E=obse ed minus expec ed. V= a iance o O–E. RR= a e a io (exp[{O–E}/V]). E o ba s a e SE . Bone ecu ence a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 6171 women RR 0·92 (95% CI 0·75–1·12) Log- ank 2p=0·42 10-yea gain 0·0% (95% CI –2·1 o 2·2) Con ol 10·3% Bisphosphona e 10·3% Yea s 0–4 1·35 (169/12 510) 1·54 (175/11 390) 0·86 (0·66–1·07) –11·4/77·4 Yea s 5–9 1·00 (47/4710) 0·69 (36/5196) 1·24 (0·73–1·74) 3·9/18·5 Yea s ≥10 0·07 (1/1390) 0·07 (1/1424) 0·37 (0·02–2·33) –0·4/0·4 7·4% 6·5% 0 10 20 30 40 50 Bone ecu ence (%) A Bone ecu ence a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 11 767 women RR 0·72 (95% CI 0·60–0·86) Log- ank 2p=0·0002 10-yea gain 2·2% (95% CI 0·6 o 3·8) Con ol 8·8% Bisphosphona e 6·6% Yea s 0–4 0·78 (197/25 220) 1·06 (251/23 642) 0·68 (0·52–0·84) –39·3/101·2 Yea s 5–9 0·67 (55/8157) 0·76 (60/7870) 0·90 (0·54–1·26) –2·8/26·8 Yea s ≥10 0·0 (0/513) 0·0 (0/484) 5·4% 3·6% 0 10 20 30 40 50 Bone ecu ence (%) B Recu ence ou side bone a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 6171 women RR 1·08 (95% CI 0·92–1·26) Log- ank 2p=0·35 10-yea loss 1·1% (95% CI 1·4 o 3·6) Bisphosphona e 17·0% Con ol 15·9% Yea s 0–4 2·64 (330/12 510) 2·41 (275/11 390) 1·12 (0·93–1·30) 14·2/128·4 Yea s 5–9 1·25 (59/4710) 1·14 (59/5196) 1·03 (0·64–1·42) 0·7/26·1 Yea s ≥10 0·43 (6/1390) 0·77 (11/1424) 0·37 (0·13–1·08) –3·0/3·1 12·2% 11·1% 0 10 20 30 40 50 Dis an ecu ence ou side bone (%) C Recu ence ou side bone a e/yea (%), e en s/woman-yea s and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 11 767 women RR 0·90 (95% CI 0·79–1·02) Log- ank 2p=0·10 10-yea gain 1·6% (95% CI –0·4 o 3·5) Con ol 13·6% Bisphosphona e 12·1% Yea s 0–4 1·67 (420/25 220) 1·98 (427/23 642) 0·90 (0·77–1·04) –18·4/182·4 Yea s 5–9 1·04 (85/8157) 1·17 (92/7870) 0·89 (0·60–1·19) –4·5/39·7 Yea s ≥10 0·78 (4/513) 1·65 (8/484) 0·38 (0·09–1·34) –1·7/1·8 8·7% 7·8% 0 10 20 30 40 50 Dis an ecu ence ou side bone (%) D Dea h a es (%/yea : o al a e minus a e in women wi hou ecu ence) and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 6171 women RR 1·00 (95% CI 0·86–1·15) Log- ank 2p=0·96 10-yea gain 0·1% (95% CI –2·9 o 3·0) Con ol 20·7% Bisphosphona e 20·6% 0 5 10 Yea s 0–4 2·43 (2·16–2·69) 2·50 (2·22–2·79) 0·97 (0·81–1·14) –3·3/130·6 Yea s 5–9 2·26 (1·84–2·67) 2·03 (1·65–2·40) 1·10 (0·81–1·40) 5·0/50·0 Yea s ≥10 1·13 (0·58–1·68) 1·29 (0·71–1·88) 0·71 (0·34–1·50) –2·3/6·9 Yea s 12·1% 11·8% 0 10 20 30 40 50 B eas cance mo ali y (%) E Dea h a es (%/yea : o al a e minus a e in women wi hou ecu ence) and log- ank s a is ics Alloca ion Bisphosphona e Con ol Ra e a io (95% CI) om (O–E)/V 11 767 women RR 0·82 (95% CI 0·73–0·93) Log ank 2p=0·002 10-yea gain 3·3% (95% CI 0·8 o 5·7) Con ol 18·0% Bisphosphona e 14·7% 0 5 10 Yea s 0–4 1·56 (1·41–1·72) 1·74 (1·58–1·91) 0·86 (0·72–0·99) –27·1/174·9 Yea s 5–9 1·57 (1·30–1·84) 2·04 (1·74–2·35) 0·76 (0·55–0·97) –18·0/65·0 Yea s ≥10 1·30 (0·34–2·26) 2·73 (1·30–4·16) 0·52 (0·18–1·44) –2·4/3·6 Yea s 8·7% 7·5% 0 10 20 30 40 50 B eas cance mo ali y (%) F A icles www. helance .com Vol 386 Oc obe 3, 2015 1359 all be es ic ed o pos menopausal women (appendix). In pos menopausal women, he e was signifi can (p=0·01) he e ogenei y be ween agen s in he educ ions in bone ecu ence, explained by he appa en lack o benefi om pamid ona e. The clod ona e esul s did appea somewha mo e p omising han he aminobisphosphona e esul s bu his diff e ence was no signifi can o dis an ecu ence o o bone ecu ence, and was o only bo de line signifi cance o b eas cance mo ali y, e en hough he educ ion in pos menopausal b eas cance mo ali y was signifi can wi h clod ona e bu no wi h he agg ega e o all aminobisphosphona e egimens. In o ma ion on ac u es was a ailable om only 13 341 (71%) o 18 766 women. Among hem, 422 (6·3%) o 6649 bisphosphona e-alloca ed pa ien s had a ac u e epo ed, as agains 487 (7·3%) o 6692 con ol pa ien s (RR 0·85, 95% CI 0·75–0·97; 2p=0·02; appendix), and he 5-yea ac u e isk was educed om 6·3% o 5·1%, wi h li le eff ec in yea s 0–1 and mos o he gain in yea s 2–4. A e yea 5 he e appea ed o be li le u he gain, bu in bo h g oups he absolu e a es a e yea 5 we e lowe han in yea s 0–4, pe haps efl ec ing incomple e asce ainmen . Discussion Taking all women oge he , ega dless o menopausal s a us, his collabo a i e me a-analysis o indi idual pa ien da a om 18 766 women andomised in ials o adju an bisphosphona es ound a highly signifi can educ ion only in bone ecu ence, and no in o he b eas cance ou comes. Subg oup analyses sugges ed benefi jus in pos menopausal women, among whom he e we e highly signifi can educ ions no only in bone ecu ence bu also in any dis an ecu ence (bone o o he ), b eas cance mo ali y, and o e all mo ali y. Nei he in he o e all esul s no in he esul s jus among pos menopausal women, howe e , was he e any signifi can eff ec on dis an ecu ence a ex a-osseous si es, on loco egional ecu ence, o on he incidence o con ala e al b eas cance . The lack o eff ec on new con ala e al b eas cance s is consis en wi h fi ndings o he la ge FIT and HORIZON-PFT ac u e p e en ion ials,20 bu con as s wi h epo s om epidemiological s udies ha b eas cance incidence is educed in pos menopausal women aking bisphosphona es o os eopo osis.21,22 Thus, he andomised e idence p o ides no suppo o he use o bisphosphona es as a b eas cance chemop e en ion s a egy. Though he s a is ical signifi cance o he appa en in e ac ion be ween menopausal s a us and ea men effi cacy is no ex eme, g ea e benefi o pos menopausal women had been hypo hesised o explain he appa en disco dance be ween he ABCSG-129 and AZURE11,14 ial esul s. Sensi i i y analyses ha excluded hese wo hypo hesis-gene a ing da ase s only ma ginally weakened he e idence o an in e ac ion wi h menopausal s a us, and he benefi was s ill signifi can in he emaining pos menopausal women. Mo eo e , he e is some p eclinical e idence ha ep oduc i e ho mones can inhibi bisphosphona e effi cacy agains cance cells in he bone. The eff ec s o zoled onic acid (100 μg/kg weekly) on he g ow h o dissemina ed MDA-231 b eas cance cells in bone we e compa ed in o a iec omised mice (modelling he pos menopausal se ing) and in sham-ope a ed mice (modelling he p emenopausal se ing). Zoled onic acid dec eased he numbe o de ec able umou s in bone only in he o a iec omised animals.23 Likewise, in a p os a e cance mouse model he abili y o dissemina ed umou cells in he bone o o m de ec able umou s was inhibi ed by zoled onic acid only in cas a ed mice, no in sham- ope a ed mice.24 The eff ec s on bone ecu ence emphasise he po en ial impo ance o hos mic oen i onmen ac o s o me as asis. Fu he s udies a e needed o cla i y why menopausal s a us should impo an ly aff ec he esponse o bisphosphona es. The complex in e ac ions be ween ep oduc i e ho mones, umou biology, bone cell unc ion, and bone ma ow s em cells could well change as pa ien s p og ess om he p emenopausal se ing, whe e oes adiol and inhibin a e o majo impo ance in bones, o he pos menopausal se ing, whe e ac i in and o he membe s o he TGF-β supe amily become he main egula o s o bone cell me abolism.25 A clea e unde s anding o some o he o he mechanisms in ol ed in he de elopmen o bone me as asis is now eme ging, al hough how hese ela e o menopausal s a us and ep oduc i e ho mones emains unknown.26 O he han he appa en eff ec o menopausal s a us o , simila ly, age on ea men effi cacy, he p opo ional educ ions in bone ecu ence and b eas cance mo ali y wi h ea men did no depend signifi can ly on o he pa ien o clinico pa hological p ima y umou cha ac e - is ics, including ER s a us, axilla y lymph node in ol e- men , and umou g ade. Simila educ ions we e seen in he p esence and absence o chemo he apy, sugges ing ha he benefi s o bisphosphona es a e app oxima ely addi i e o hose o chemo he apy, and ice e sa. As subg oup analyses can yield e a ic esul s, i is diffi cul o de e mine om hem whe he diff e en bisphosphona e egimens ha e diff e en eff ec s. The endpoin ha should yield he mos eliable subg oup analyses is bone ecu ence. Bo h o all women and o pos menopausal women, subg oup analyses o bone ecu ence sugges ed simila eff ec s o o al clod ona e and o he agg ega e o all aminobisphosphona e egimens (mainly in a enous zoled onic acid). Likewise, hey sugges ed no signifi can he e ogenei y in effi cacy be ween he diff e en aminobisphosphona es, hough no benefi was seen wi h o al pamid ona e (which could be eal, as o al pamid ona e is poo ly abso bed, has li le eff ec on bone eso p ion bioma ke s o he unde lying me as a ic bone disease, and ailed o show effi cacy in myeloma27,28). Numbe s we e insuffi cien o assess he effi cacy o he s anda d ea men s o os eopo osis, o al A icles 1360 www. helance .com Vol 386 Oc obe 3, 2015 ised ona e o alend ona e, as he apy o ea ly b eas cance . Subg oup analyses based ins ead on b eas cance mo ali y sugges ed a g ea e eff ec wi h clod ona e han wi h aminobisphosphona es. Howe e , as he wo d ugs appea ed o ha e simila eff ec s on bone ecu ence, hei appa en ly diff e en eff ec s on b eas cance mo ali y could be a chance fi nding. Much mo e eliable compa isons o diff e en bis- phosphona e egimens will eme ge om ongoing ials ha compa e hem di ec ly. The SWOG0307 ial (NCT00127205) compa ing clod ona e e sus zoled onic acid e sus iband ona e in 5400 pa ien s has comple ed ec ui men and add esses he choice o agen ; he SUCCESS ial (NCT02181101) compa ing 5 yea s e sus 2 yea s o zoled onic acid in 3800 pa ien s has also comple ed ec ui men and add esses du a ion. Simila ly, esul s om wo ongoing ials (HOBOE-p emenopausal [NCT00412022] and TEAM-IIb [ISRCTN17633610]), plus longe ollow-up o he ials included in his me a- analysis, will e en ually p o ide be e e idence on any eff ec o bisphosphona es in p emenopausal women, and will p o ide mo e s able es ima es o he 10-yea ou comes in pos menopausal women. Consis en wi h he known eff ec s on bone mine al densi y and quali y, he use o adju an bisphosphona es was associa ed wi h a small educ ion in ac u e incidence. Al hough no highly signifi can , i can be accep ed as eal because o e idence o ac u e educ ion in o he ypes o pa ien . The e was no appa en eff ec o adju an bisphosphona es on non-b eas cance mo ali y. Majo ad e se e en s wi h bisphosphona es a e uncommon, bu can include impai ed enal unc ion and os eonec osis o he jaw. F om he da a p o ided, we we e unable o assess he incidence o os eonec osis o he jaw, bu p e ious epo s sugges i anges om unde 1% wi h clod ona e, iband ona e, o 6-mon hly zoled onic acid12,13,29 o abou 2% wi h mo e in ensi e zoled onic acid schedules.30 These ials ha e shown ha some yea s o adju an bisphosphona e ea men can educe b eas cance ecu ence a es in bone and imp o e b eas cance su i al, bu ha e p o ided clea e idence o benefi only in women who a e pos menopausal (na u al o induced) a he ime bisphosphona es a e s a ed. The use o bisphosphona es in b eas cance is mainly o educe bone loss and isk o ac u e in pos menopausal women wi h ER-posi i e disease ea ed wi h a oma ase inhibi o s. Ou esul s show ha such bisphosphona e ea men can, in addi ion, p o ide oncological benefi , and sugges ha adju an bis phosphona es should be conside ed in a b oade ange o pos menopausal women. Con ibu o s Analyses we e planned by R Coleman, R G ay, T Powles, A Pa e son, M Gnan , J Be gh, K I P i cha d, J Bliss, and D Came on, and unde aken by R B adley, R G ay, H Pan, and R Pe o in Ox o d. R Coleman, R G ay, T Powles, and A Pa e son d a ed he epo and e ised i wi h ad ice om all w i ing commi ee membe s. The EBCTCG sec e a ia (R B adley, J Bu e , M Cla ke, C Da ies, F Duane, V E ans, L Ge ins, J Godwin, R G ay, H Liu, P McGale, E MacKinnon, T McHugh, S James, P Mo is, H Pan, R Pe o, S Read, C Taylo , Y Wang, Z Wang) iden ifi ed ials, ob ained da ase s, and had ull access o hem. W i ing commi ee R Coleman (Sheffi eld Cance Resea ch Cen e, Wes on Pa k Hospi al, Uni e si y o Sheffi eld, UK); T Powles (Royal Ma sden Hospi al, Su on, Su ey, UK); A Pa e son (Tom Bake Cance Cen e and Uni e si y o Calga y, Calga y, Albe a, Canada); M Gnan (Depa men o Su ge y and Comp ehensi e Cance Cen e , Medical Uni e si y o Vienna, Aus ian B eas & Colo ec al Cance S udy G oup, Vienna, Aus ia); S Ande son (Depa men o Bios a is ics, Uni e si y o Pi sbu gh G adua e School o Public Heal h, Pi sbu gh, PA, USA); I Diel (Cen e o Gynecological Oncology, Mannheim, Ge many); J G alow (Uni e si y o Washing on School o Medicine, Sea le, WA, USA); G on Minckwi z (Ge man B eas G oup, Neu-Isenbu g, Ge many); V Moebus (Klinikum F ank u Höchs , F ank u , Ge many); J Be gh (Ka olinska Ins i u e and Uni e si y Hospi al, S ockholm, Sweden); K I P i cha d (Sunnyb ook Ode e Cance Cen e and Uni e si y o To on o, To on o, Canada); J Bliss (ICR-CTSU, Di ision o Clinical S udies, The Ins i u e o Cance Resea ch, London, UK); D Came on (Edinbu gh Cance Resea ch Cen e, Uni e si y o Edinbu gh, Edinbu gh, UK); V E ans (Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, Ox o d, UK); H Pan (Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, Ox o d, UK); R Pe o (Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, Ox o d, UK); R B adley (Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, Ox o d, UK); R G ay (Clinical T ial Se ice Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, Ox o d, UK). G oups con ibu ing da a o he adju an bisphosphona es me a-analyses Aus ian B eas Cance S udy G oup, Vienna, Aus ia (R Ba sch, P Dubsky, C Fesl, H Fohle , M Gnan , R G eil, R Jakesz, A Lang, G Luschin-Ebeng eu h, C Ma h, B Mline i sch, H Samonigg, C F Singe , G G S ege , H S öge ); B i ish Columbia Cance Agency, Vancou e , Canada (I Oli o o, J Ragaz); Danish B eas Cance Coope a i e G oup, Copenhagen, Denma k (P Ch is iansen, B Ejle sen, M Ewe z, M-B Jensen, S Mølle , H T Mou idsen); Ge man Adju an B eas G oup, F ank u , Ge many (W Eie mann, J Hil ich, W Jona , M Kau mann, R K eienbe g, M Schumache ); Ge man B eas G oup, Neu-Isenbu g, Ge many (J U Blohme , S D Cos a, H Eid mann, B Ge be , C Jackisch, S Loibl, G on Minckwi z); Hellenic B eas Su geons Socie y, G eece (U Da ni, C Ma kopoulos); Helsinki Uni e si y, Finland (C Blomq is , T Saa o); Ko ean Cance S udy G oup, Seoul, Sou h Ko ea (J-H Ahn, K H Jung); Is i u o Nazionale Tumo i IRCCS Fondazione Pascale, Naples, I aly (F Pe one); Na ional Su gical Adju an B eas and Bowel P ojec , Pi sbu gh, PA, USA (S Ande son, G Bass, A B own [deceased], J B yan [deceased], J Cos an ino, J Dignam, B Fishe , C Geye , E P Mamounas, S Paik, C Redmond, S Swain, L Wicke ham, N Wolma k); No h Cen al Cance T ea men G oup, Mayo Clinic, Roches e , MN, USA (E Pe ez, JN Ingle, V J Suman); P oBONE s udy g oup, Ma bu g, Ge many (P Hadji); Royal Ma sden Hospi al, London and Su on, UK (R A’He n, M Dowse , A Mak is, M Pa on, K Penne , T J Powles, I E Smi h, J R Ya nold); SABRE ial g oup (in e na ional) (G Clack, C Van Poznak); Tel A i Sou asky Medical Cen e , Is ael (T Sa a); Uni e si y o Leeds, UK (R Bell, D Came on, RE Coleman, D Dodwell, S Hinsley, H C Ma shall); Uni e si y o Saa land, Ge many (E Solomaye , T Fehm); Uni e si y o Sheffi eld, UK (R E Coleman, J Les e , M C Win e , J M Ho sman); Washing on Uni e si y, S Louis, Missou i, USA (R A ); Z-FAST, ZO-FAST & E-ZO-FAST s udy g oups (in e na ional) (A M B u sky, R Coleman, H A Llomba on behal o No a is Pha maceu icals). EBCTCG s ee ing commi ee J Be gh (co-chai ), K P i cha d (co-chai ), K Albain, S Ande son, R A iagada, W Ba low, E Be gs en-No ds öm, J Bliss, F Bocca do, R B adley*, M Buyse, D Came on, M Cla ke*, A Coa es, R Coleman, C Co ea, J Cos an ino, J Cuzick, N Da idson, C Da ies*, A Di Leo, M Dowse , M Ewe z, J Fo bes, R Gelbe , C Geye , L Gianni, M Gnan , A Goldhi sch, R G ay*, D Hayes, C Hill, J Ingle, W Janni, E MacKinnon*, M Ma ín, P McGale*, L No on, Y Ohashi, S Paik, H Pan*, E Pe ez, R Pe o*, M Picca , L Pie ce, V Raina, P Ra din, J Robe son, E Ru ge s, J Spa ano, S Swain, C Taylo *, G Viale, G on Minckwi z, X Wang, T Whelan, N Wilcken, E Wine , N Wolma k, W Wood. *Sec e a ia . A icles www. helance .com Vol 386 Oc obe 3, 2015 1361 Decla a ion o in e es s Clinical T ial Se ice Uni (CTSU) s aff policy excludes hono a ia o consul ancy ees o any membe o he Ea ly B eas Cance T ialis s’ Collabo a i e G oup Sec e a ia . EBCTCG is unded by Cance Resea ch UK and UK Medical Resea ch Council g an s o he CTSU. SA epo s g an s om Na ional Ins i u es o Heal h (U10 CA069974 and U10 CA69651), du ing he conduc o he s udy. JBe epo s ha Ka olinska Uni e si y Hospi al and Ka olinska Ins i u e ha e ecei ed paymen s o academic clinical s udies and esea ch g an s o molecula biological s udies and PET s udies om Amgen, As aZeneca, Baye , Me ck, Pfi ze , Roche, and Sanofi -A en is; he was Swedish p incipal in es iga o (PI) o an adju an bisphosphona e s udy o which o al pamid ona e was p o ided ee-o -cha ge ( his o mula ion is no licensed); he is also Swedish PI o he ongoing ABCSG-18 adju an denosumab s udy ( he d ug was p o ided ee-o -cha ge); he epo s no pe sonal paymen s in he pas 3 yea s. DC epo s suppo om No a is o a end he Ame ican Socie y o Clinical Oncology and San An onio B eas Cance Symposium con e ences, ou side he submi ed wo k. RC epo s pe sonal ees om No a is ( o expe es imony), ou side he submi ed wo k. MG epo s g an s and pe sonal ees om No a is and pe sonal ees om Amgen, du ing he conduc o he s udy; ou side he submi ed wo k he has ecei ed g an s and pe sonal ees om No a is, Roche, and GlaxoSmi hKline, g an s om Sanofi -A en is, Pfi ze , and Smi h Medical, and pe sonal ees om As aZeneca, Nanos ing Technologies, and Accelsio s. JG epo s g an s om No a is, Amgen, and Roche, ou side he submi ed wo k. VM epo s g an s and pe sonal ees om Amgen and Roche, g an s om No a is, and pe sonal ees om Celgene, ou side he submi ed wo k. KIP epo s g an s and pe sonal ees om As aZeneca, Pfi ze , Roche, No a is, and Eisai, and pe sonal ees om Amgen and GlaxoSmi hKline, ou side he submi ed wo k. JBl, RB, ID, VE, RG, HP, AP, RP, TP, and G M decla e no compe ing in e es s. Acknowledgmen s We hank he ens o housands o women who ook pa in he ials, he many s aff in ial cen es and pa icipa ing clinics who helped conduc he ials, and he ialis s who sha ed hei da a. Funding o indi idual ials was chiefl y om manu ac u e s (see ial publica ions) bu no comme cial unding was sough o used by he sec e a ia . Funding o he EBCTCG sec e a ia is h ough he di ec suppo om Cance Resea ch UK and he UK Medical Resea ch Council, o he Clinical T ial Se ice Uni and Epidemiological S udies Uni , Nuffi eld Depa men o Popula ion Heal h, Uni e si y o Ox o d, UK. Re e ences 1 Weilbaeche KN, Guise TA, McCauley LK. Cance o bone: a a al a ac ion. Na Re Cance 2011; 11: 411–25. 2 Roodman GD. Mechanisms o bone me as asis. N Engl J Med 2004; 350: 1655–64. 3 Guo W. Concise e iew: b eas cance s em cells: egula o y ne wo ks, s em cell niches, and disease ele ance. S em Cells T ansl Med 2014; 3: 942–48. 4 Mundy GR. Me as asis o bone: causes, consequences and he apeu ic oppo uni ies. Na Re Cance 2003; 2: 584–93. 5 Thompson K, Roelo s AJ, Jauhiainen M, Mönkkönen H, Mönkkönen J, Roge s MJ. Ac i a ion o γδ T cells by bisphosphona es. Ad Exp Med Biol 2010; 658: 11–20. 6 Kanis J, Powles T, Pa e son A, e al. Clod ona e dec eases he equency o skele al me as ases in women wi h b eas cance . Bone 1996; 19: 663–67. 7 Powles TJ, Pa e son AE, McCloskey E, e al. Reduc ion in bone elapse and imp o ed su i al wi h o al clod ona e o adju an ea men o ope able b eas cance . B eas Cance Res T ea 2006; 8: R13, 1–7. 8 Diel IJ, Jaschke A, Solomaye EF, e al. Adju an o al clod ona e imp o es he o e all su i al o p ima y b eas cance pa ien s wi h mic ome as ases o he bone ma ow—a long e m ollow up. Ann Oncol 2008; 19: 2007–11. 9 Gnan M, Mline i sch B, Schippinge W, e al. Endoc ine he apy plus zoled onic acid in p emenopausal b eas cance . N Engl J Med 2009; 360: 679–91. 10 Coleman RE, de Boe R, Eid mann H, e al. Zoled onic acid (zoled ona e) o pos menopausal women wi h ea ly b eas cance ecei ing adju an le ozole (ZO- as s udy): fi nal 60-mon h esul s. Ann Oncol 2013; 24: 398–405. 11 Coleman RE, Came on D, Dodwell D, e al, on behal o he AZURE in es iga o s. Adju an zoled onic acid in pa ien s wi h ea ly b eas cance : fi nal effi cacy analysis o he AZURE (BIG 01/04) andomised open-label phase 3 ial. Lance Oncol 2014; 15: 997–1006. 12 Pa e son AHG, Ande son SJ, Lembe sky BC, e al. O al clod ona e o adju an ea men o ope able b eas cance (Na ional Su gical Adju an B eas and Bowel P ojec p o ocol B-34): a mul icen e, placebo-con olled, andomised ial. Lance Oncol 2012; 13: 734–42. 13 on Minckwi z G, Möbus V, Schneeweiss A, e al. Ge man adju an in e g oup node-posi i e s udy: a phase III ial o compa e o al iband ona e e sus obse a ion in pa ien s wi h high- isk ea ly b eas cance . J Clin Oncol 2013; 31: 3531–39. 14 Coleman RE, Ma shall H, Came on D, e al, on behal o he AZURE in es iga o s. B eas cance adju an he apy wi h zoled onic acid. N Engl J Med 2011; 365: 1396–405. 15 Coleman R, Gnan M, Mo gan G, Cleza din P. Eff ec s o bone- a ge ed agen s on cance p og ession and mo ali y. J Na l Cance Ins 2012; 104: 1059–67. 16 Hadji P, Coleman R, Gnan M, G een J. The impac o menopause on bone, zoled onic acid, and implica ions o b eas cance g ow h and me as asis. Ann Oncol 2012; 23: 2782–90. 17 Ea ly B eas Cance T ialis s’ Collabo a i e G oup. T ea men o ea ly b eas cance : wo ldwide e idence, 1985–1990. In oduc ion and me hods. Ox o d: Ox o d Uni e si y P ess, 1990. h p://www.c su. ox.ac.uk/ epo s/ebc cg-1990/index_h ml (accessed July 7, 2015). 18 Ea ly B eas Cance T ialis s’ Collabo a i e G oup (EBCTCG). Eff ec s o chemo he apy and ho monal he apy o ea ly b eas cance on ecu ence and 15-yea su i al: an o e iew o he andomised ials. Lance 2005; 365: 1687–717. 19 Ea ly B eas Cance T ialis s’ Collabo a i e G oup. Rele ance o b eas cance ho mone ecep o s and o he ac o s o he effi cacy o adju an amoxi en: pa ien -le el me a-analysis o andomised ials. Lance 2011; 378: 771–84. 20 Hue TF, Cummings SR, Cauley JA, e al. Eff ec o bisphosphona e use on isk o pos menopausal b eas cance : esul s om he andomized clinical ials o alend ona e and zoled onic acid. JAMA In e n Med 2014; 174: 1550–57. 21 Chlebowski RT, Chen Z, Cauley JA, e al. O al bisphosphona e use and b eas cance incidence in pos menopausal women. J Clin Oncol 2010; 28: 3582–90. 22 Renne G, Pinche M, Renne HS. Use o bisphosphona es and isk o pos menopausal b eas cance . J Clin Oncol 2010; 28: 3577–81. 23 O ewell PD, Wang N, Meek J, e al. Zoled onic acid has diff e en ial an i umo ac i i y in he p e- and pos menopausal bone mic oen i onmen in i o. Clin Can Res 2014; 20: 2922–32. 24 O ewell PD, Wang N, B own HK, e al. Cas a ion-induced bone loss igge s g ow h o dissemina ed p os a e cance cells in bone. Endoc ine-Rela ed Cance 2014; 21: 769–81. 25 Nicks KM, Fowle TW, Akel NS, e al. Bone u no e ac oss he menopausal ansi ion, he ole o gonadal inhibins. Ann NY Acad Sci 2010; 1192: 153–60. 26 Cox TR, Rumney RMH, Schoo EM, e al. The hypoxic cance sec e ome induces p e-me as a ic bone lesions h ough lysyl oxidase. Na u e 2015; 522: 106–10. 27 K is ensen B, Ejle sen B, Mou idsen H, e al. Bisphosphona e ea men in p ima y b eas cance : esul s om a andomised compa ison o o al pamid ona e e sus no pamid ona e in pa ien s wi h p ima y b eas cance . Ac a Oncol 2008; 47: 740–46. 28 B incke H, Wes in J, Abildgaa d N, e al. Failu e o o al pamid ona e o educe skele al mo bidi y in mul iple myeloma—a double blind placebo con olled ial. B J Haema ol 1998; 101: 280–86. 29 Gnan M, Mline i sch B, S oege H, e al. Zoled onic acid combined wi h adju an endoc ine he apy o amoxi en e sus anas ozole plus o a ian unc ion supp ession in p emenopausal ea ly b eas cance : fi nal analysis o he Aus ian B eas and Colo ec al Cance S udy G oup T ial 12. Ann Oncol 2014; 26: 313–20. 30 Ra hbone EJ, B own JE, Ma shall HC, e al. Os eonec osis o he jaw and o al heal h- ela ed quali y o li e a e adju an zoled onic acid: an Adju an Zoled onic Acid o Reduce Recu ence T ial subp o ocol (BIG1/04). J Clin Oncol 2013; 31: 2685–92.