Cholesterol-lowering treatment in chronic kidney disease: multistage pairwise and network meta-analyses

1 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s Choles e ol-Lowe ing ea men in Ch onic Kidney Disease: Mul is age pai wise and Ne wo k Me a- Analyses F ancisco He e a-Gómez 1,2, M. Mon se a Chimeno3, Débo a Ma ín-Ga cía4, F ank Liza aso-so o5, Ál a o Mau ua-B iseño-Meiggs6, Jesús G ande-Villo ia2, Juan Bus aman e-Mungui a7, E ic Alama ine8, Miquel Vila dell9, Ca los ochoa-sang ado 10 & F. Ja ie Ál a ez 1,11 Pai wise and ne wo k me a-analyses on he ela ionship be ween he e icacy o he use o s a ins wi h o wi hou eze imibe and educ ions in low-densi y lipop o ein choles e ol (LDLc) and C- eac i e p o ein (CRP) in pa ien s wi h ch onic kidney disease (CKD) a e p esen ed. In he pai wise me a- analysis, s a ins wi h o wi hou eze imibe we e shown o be e icacious in educing majo ad e se ca dio ascula e en s (MACE) in pa ien s wi h CKD and an es ima ed glome ula il a ion a e (eGFR) o less han 60 ml/min/1.73 m2, in he con ex o bo h p ima y p e en ion [odds a io (OR)/95% con idence in e al (95% CI)/I2/numbe o s udies (n): 0.50/0.40–0.64/0%/6] and p ima y/seconda y p e en ion (0.66/0.57–0.76/57%/18). Howe e , in he Bayesian ne wo k me a-analysis, compa ed o he placebo, only a o as a in 80 mg daily and a o as a in and osu as a in a doses equi alen o sim as a in 20 mg daily educed he odds o MACEs in his pa ien popula ion. The ne wo k me a-analysis o LDLc and CRP ea men objec i es also showed ha , ega dless o eGFR and excluding dialysis pa ien s, he numbe o MACEs dec eased in pa ien s wi h CKD, wi h educ ions in bo h LDLc and CRP o less han 50% (su ace unde he cumula i e anking (SUCRA)/he e ogenei y ( ague)/n: 0.77/0.14/3). The e alua ion o he bene i s o d ugs may lead o indi idualized he apy o CKD pa ien s: Choles e ol- lowe ing ea men o CKD pa ien s wi h high le els o bo h LDLc and CRP is sugges ed. Ca dio ascula disease (CVD) occu s ea ly in indi iduals wi h ch onic kidney disease (CKD)1: In lamma ion is cen al o he de elopmen o he i s and subsequen CVD e en s in CKD pa ien s (as well as in non-CKD pa ien s)2 and plays an essen ial ole in linking CVD and kidney unc ion decline1. Howe e , classic p edic- o s o co ona y and ce eb o ascula e en s seem o pe o m be e in he ea ly s ages o he CKD spec um3. Impo an ly, in CKD pa ien s, po en ial new ma ke s o CVD e en s ha a e de i ed om analyses o he com- plex pa hophysiology o CKD ac as independen p ognos ic ac o s4,5. Despi e ad ances in he quali ica ion o bioma ke s o he use o medicines6, he ansla ion o eal ea men e ec modi ie s o CVD in he CKD pop- ula ion is needed o e alua e he e ec s o he a ailable he apies7. 1Pha macological Big Da a Labo a o y, Pha macology and he apeu ics, acul y o Medicine, Uni e si y o Valladolid, Valladolid, Spain. 2neph ology, Hospi al Vi gen de la concha – Sanidad de cas illa y León, Zamo a, Spain. 3in e nal Medicine, Hospi al Vi gen de la concha – Sanidad de cas illa y León, Zamo a, Spain. 4ca dio ascula isk uni , Hospi al clínico Uni e si a io de Valladolid – Sanidad de cas illa y León, Valladolid, Spain. 5ins i u o de in es igación de la acul ad de Medicina Humana, Uni e sidad de San Ma ín de Po es, Lima, Pe u. 6Woodland Medical P ac ice – nHS, Lincolnshi e, Uni ed Kingdom. 7ca diac Su ge y, Hospi al clínico Uni e si a io de Valladolid – Sanidad de cas illa y León, Valladolid, Spain. 8neph ology, Dialysis and ansplan a ion, cen e Hospi alie Uni e si ai e de Sain -E ienne, Sain -P ies -en-Ja ez, F ance. 9Medicine, acul y o Medicine, Uni e sidad Au ónoma de Ba celona, Ba celona, Spain. 10Resea ch uni , Hospi al Vi gen de la concha – Sanidad de cas illa y León, Zamo a, Spain. 11ceim Á ea de Salud Valladolid es e, Hospi al clínico Uni e si a io de Valladolid, Valladolid, Spain. Co espondence and eques s o ma e ials should be add essed o F.H.-G. (email: [email p o ec ed]) Recei ed: 25 Oc obe 2018 Accep ed: 5 June 2019 Published: xx xx xxxx opeN 2 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ The isk o co ona y e en s in pa ien s wi h CKD is high8, and he ela ed e idence is in a o o educing CVD isk using choles e ol-lowe ing ea men s a egies9. Howe e , he ela ionship be ween se um le els o low-densi y lipop o ein choles e ol (LDLc) and CVD is mo e di icul o unde s and in he con ex o CKD; malnu i ion and ch onic in lamma ion a e addi ional ac o s o conside as CKD p og esses, and hese ac o s inc ease he isk o dea h in CKD pa ien s9. Cu en con ex and s udy objec i e. The Kidney Disease–Imp o ing Global Ou comes (KDIGO) guidelines ecommend he use o s a ins wi h o wi hou eze imibe o adul s 50 yea s o age o olde wi h CKD and an es ima ed glome ula il a ion a e (eGFR) o lowe han 60 ml/min/1.73 m2 (KDIGO GFR ca ego ies G3a o G5) who a e no being ea ed wi h ch onic dialysis o kidney ansplan a ion and no aiming o achie e speci ic LDLc a ge s10. S a ins a e also ecommended o indi iduals olde han 50 yea s when he eGFR is 60 ml/ min/1.73 m2 o highe (KDIGO GFR ca ego ies G1 o G2). No ably, in he e alua ion o he impac o choles e ol-lowe ing ea men along he wide spec um o CKD, an unanswe ed ques ion emains: a e s a ins wi h o wi hou eze imibe simila ly e icacious in pa ien s wi h an eGFR lowe han 60 ml/min/1.73 m2 and in pa ien s wi h an eGFR o 60 ml/min/1.73 m2 o highe ? I no , which choles e ol-lowe ing ea men s a egies a e he bes o lowe ing CVD isk in pa ien s wi h an eGFR lowe han 60 ml/min/1.73 m2? Fu he mo e, is he ca dio ascula e icacy o s a ins wi h o wi hou eze imibe ela ed o educ ions in LDLc and C- eac i e p o ein (CRP) in CKD pa ien s as is he case o LDLc in non-CKD pa ien s? Con en ional pai wise me a-analysis may be insu icien o answe hese ques ions, as he e icacy be ween mul i- ple in e en ions ha compa e choles e ol-lowe ing ea men s and placebo/no ea men canno be s udied sep- a a ely. Only one possibili y exis s based on his echnique: o compa e choles e ol-lowe ing ea men s as a whole wi h placebo/no ea men . The same concep applies o he educ ion in LDLc and CRP wi h espec o he sole educ ion in LDLc when me a-analysis on exposu es is pe o med. Ne wo k me a-analysis p o ides a iable solu ion e en hough some in e en ions/exposu es ha e ne e been compa ed head- o-head in andomized con- olled ials ha e alua e he ca dio ascula e icacy o s a ins wi h o wi hou eze imibe in he con ex o CKD. These mul is age pai wise and ne wo k me a-analyses p esen s (1) a summa y o he e icacy o s a ins wi h o wi hou eze imibe along he wide spec um o CKD and (2) he ela ion o his ea men o educ ions in LDLc and CRP wi h he aim o e alua ing he eal impac o choles e ol lowe ing- ea men on CVD in CKD pa ien s. Resul s Li e a u e sea ch. Da a om 106 050 pa icipan s om 22 andomized con olled ials (RCT) a e p e- sen ed he e11–35. Figu e1 p esen s he P e e ed Repo ing I ems o Sys ema ic e iews and Me a-Analyses (PRISMA) lowcha co esponding o he wo sys ema ic e iews36. I ele an ci a ions we e mos ly in e en- ional o obse a ional s udies o pa ien s wi h ea ly CKD (KDIGO GFR ca ego ies G1 o G2) ha did no assess CVD- ela ed ou comes o s udies in ol ing pa ien s wi hou CKD bu a ec ed by o he pa hologies. All o he ollowing RCTs ha we e eligible a he ea men objec i es e alua ion phase had al eady been included a he choles e ol-lowe ing ea men e icacy e alua ion phase: The S udy o Hea and Renal P o ec ion (SHARP)17, he Ai Fo ce/Texas Co ona y A he oscle osis P e en ion s udy (AFCAPS/TexCAPS)23,37, he Collabo a i e A o as a in Diabe es S udy (CARDS)26,38, and he Jus i ica ion o he Use o S a ins in P ima y P e en ion–an In e en ion T ial E alua ing Rosu as a in (JUPITER) s udy28,39. All hese pha maceu ical-indus y-sponso ed ials we e published in pee - e iewed jou nals and co e- sponded o he phase 3 o 4 e alua ion o a ious s a ins wi h o wi hou eze imibe, wi h bo h adminis e ed pe os. Nine een o hese ials sepa a ely assessed pe sons wi h no mal kidney unc ion (NKF)/CKD KDIGO GFR ca ego ies G1–G2 and pa ien s wi h CKD KDIGO GFR ca ego ies G3a–G5 bu no pa ien s ea ed wi h ch onic dialysis16–35; 12 o he abo emen ioned s udies we e pos hoc analyses o o he RCTs concei ed o e alua e pa - icipan s om he gene al popula ion o pa ien s wi h o he pa hologies22–33. Fi e RCTs s udied dialysis pa ien s (hemodialysis and pe i oneal dialysis modali ies)13–18, and 1 RCT e alua ed pa ien s who unde wen kidney ansplan a ion11,12. The eligible RCTs a e p esen ed in TableS1. Choles e ol-lowe ing ea men e icacy e alua ion. Mode a e- o high-quali y ials (TableS2) indica e ha a o as a in 80 mg daily, sim as a in 20 mg daily combined wi h eze imibe 10 mg daily, sim as a in 20 mg daily and millig am-equi alen doses o lu as a in, a o as a in, osu as a in, p a as a in and lo as a in all educed he incidence o majo ad e se ca dio ascula e en s (MACEs) in pa ien s wi h CKD KDIGO GFR ca ego ies G3a–G5 (excluding dialysis pa ien s)16–35. Figu e2a,b shows he pooled odds a ios (OR) and 95% con idence in e als (95% CI) co esponding o pai wise me a-analysis o all ials conduc ed in his popula ion (p ima y/seconda y p e en ion; 0.66, 0.57 o 0.76) and hose ha en olled only pa ien s wi hou known CVD (p ima y p e en ion; 0.50, 0.40 o 0.64). The he e ogenei y (I2) was 57% and 0%, espec i ely. Howe e , Bayesian ne wo k me a-analysis con i med ha compa ed o placebo, only a o as a in 80 mg daily and a o as a in and osu as a in a doses equi alen o sim as a in 20 mg daily educed he odds o inciden MACEs when he eGFR was lowe han 60 ml/min/1.73 m2 (excluding dialysis pa ien s). The ORs and 95% CIs o he abo emen ioned ea men s we e 0.35, 0.16 o 0.74; 0.50, 0.26 o 0.92; and 0.67, 0.40 o 0.92, espec i ely. Figu e3 shows a league able anking all s a ins and he sim as a in/eze imibe combina ion on he basis o he su ace unde he cumu- la i e anking (SUCRA)40. In he gene a ed ne wo k diag am (Fig.4), he he e ogenei y ( ague) was 0.08 (Fig.5), and he e was no ele an inconsis ency (Fig.6). In pe sons wi h NKF/CKD KDIGO GFR ca ego ies G1–G217,22–35 and in dialysis pa ien s13–18, he ORs and 95% CIs o MACEs we e 0.78, 0.71 o 0.85 and 0.95, 0.81 o 1.07, espec i ely (FigsS1 and S2). T ea men objec i es e alua ion. Se um le els o LDLc and/o CRP dec eased in esponse o ea men wi h sim as a in 20 mg daily combined wi h eze imibe 10 mg daily17, lo as a in 40 mg daily37, a o as a in 10 mg 3 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ daily38, and osu as a in 20 mg daily39, ega dless o eGFR and his o y o diabe es o CVD. Mos pa ien s wi h ea ly CKD (eGFR ≥ 60 ml/min/1.73 m2) had high LDLc le els be o e ea men 17,37–39. A educ ion in CRP o mo e han 50% was no obse ed. Reduc ions in LDLc and/o CRP o less han 50% we e no ed in pa ien s wi h CKD in any KDIGO GFR ca ego y37–39. Pai wise me a-analysis showed ha , ega dless o eGFR and excluding dialysis pa ien s, educ ions in bo h LDLc and CRP o less han 50%, and educ ions in LDLc (</≥50%) and CRP (<50%) sepa a ely we e associa ed wi h a low equency o MACEs (FigsS3–S6)17,37–39. Howe e , in he ne wo k me a-analysis, only he combined educ ion in LDLc and CRP appea ed o ha e he mos p onounced impac on his ou come37–39. The SUCRA was 0.77 (Table1)40. In he ne wo k diag am buil by ea men objec i es (Fig.S7), he he e ogenei y ( ague) was 0.14 (Fig.S8) and he e was no inconsis ency (Fig.S9). Me a- eg ession wi h he combined educ ion in LDLc and CRP and choles e ol-lowe ing ea men s was no possible gi en he numbe o s udies included41. Al hough clea de ails we e no ob ained, pos hoc analyses e alua ing he combined educ ion in LDLc and CRP as p edic o s o u u e MACEs in CKD pa ien s ea ed wi h a s a in o s a in/eze imibe combina ion we e conduc ed ollowing a e ospec i e bioma ke -s a i ied design ha allowed o he e alua ion o codependen heal h echnologies (TableS3)37–39. Discussion Key messages. S a ins a doses equi alen o sim as a in 20 mg daily (e en conside ing he addi ion o eze- imibe) and s a ins a highe doses showed simila e icacy in educing MACEs when he eGFR was lowe han 60 ml/min/1.73 m2: Only a o as a in and osu as a in we e associa ed wi h a clea bene i in e ms o MACE educ ion. I espec i e o eGFR, MACEs we e less equen in pa ien s wi h a mode a e educ ion in LDLc and CRP han in pa ien s wi h educ ions in ei he o hese wo pa ame e s sepa a ely. LDLc and/o CRP dec ease as a esul o ea men wi h s a ins alone o in combina ion wi h eze imibe in CKD pa ien s17,37–39. Howe e , i should no be o go en ha malnu i ion and ch onic in lamma ion appea quickly once he eGFR is lowe han 60 ml/min/1.73 m2 9. As a consequence, he le els o LDLc and o he lipop o- eins dec ease, esul ing in an inc eased isk o all-cause and CVD mo ali y42. The e alua ion o he bene i s o s a ins wi h o wi hou eze imibe along he wide spec um o CKD is hus a eal necessi y. This s udy ound ha mos pa ien s wi h ea ly-s age CKD (eGFR ≥ 60 ml/min/1.73 m2) p esen ed wi h high LDLc le els be o e ea men 17,37–39 and ha he educ ions in LDLc and CRP we e mode a e ega dless o he Figu e 1. The PRISMA lowcha s p esen ing he selec ion p ocesses used in he wo sys ema ic e iews. PRISMA, P e e ed Repo ing I ems o Sys ema ic Re iews and Me a-Analyses. 4 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ eGFR o he pa ien s37–39; i is likely ha in he ea ly s ages o CKD, pa ien s de elop high LDLc le els, which hen dec ease as a esul o he ea men (simila o non-CKD pa ien s). Impo an ly, CRP is mo e di icul o dec ease in CKD pa ien s, and his becomes e en mo e challenging as he need o dialysis app oaches43: Ou me a-analy ic calcula ions e ealed ha educ ions in LDLc and CRP, compa ed o he sole educ ion in LDLc, we e associa ed wi h a low equency o MACEs. In ou opinion, bo h LDLc and CRP should be conside ed o CKD pa ien s ini ia ing choles e ol-lowe ing ea men , as is cu en ly he case o LDLc o non-CKD pa ien s. This s udy explo ed which choles e ol-lowe ing ea men s a egies a e he bes o ea ing CVD isk in his pa icula popula ion. CKD cons i u es he common inal mani es a ion o a cons ella ion o pa hologies ha a ec kidneys in a ch onic and i e e sible way, so he apies should be adap ed o he complex and in ica e pa hophysiology esul ing om he inc emen al disease bu den; e idence-based ea men s a egies should hus be concei ed as a p agma ic op ion o imp o e decision-making o he daily managemen o pa ien s3. S a ins, in addi ion o lowe ing LDLc le els, exe an i-in lamma o y, an ioxidan , an i-p oli e a i e and immunomodula o y e ec s, such as in luencing plaque s abili y, no malizing sympa he ic ou low, and p e en - ing pla ele agg ega ion, among o he pleio opic bene icial e ec s44–46. The e ec s o s a ins on CVD isk in CKD pa ien s ha e been discussed in he mos ecen sys ema ic e iews and me a-analyses47–50 and o he p e ious e idence summa ies51–53. Ou esul s a e in line wi h he indings o hese s udies. In pa icula , ou indings in he Bayesian ne wo k me a-analysis a e consis en wi h hose ob ained in The Choles e ol T ea men T ialis s (CTT) Collabo a ion me a-analysis o indi idual pa icipan Figu e 2. The e ec o choles e ol-lowe ing ea men on MACEs when he eGFR is lowe han 60 ml/ min/1.73 m2 (excluding dialysis pa ien s). (a) All ials. (b) P ima y p e en ion ials. CI, con idence in e al; eGFR, es ima ed glome ula il a ion a e; EZE, eze imibe; M-H, Man el–Haenszel es ; PBO, placebo; SE, s anda d e o ; STA, s a ins. Figu e 3. SUCRA-based anking o all s a ins and he combina ion o sim as a in plus eze imibe when he eGFR is lowe han 60 ml/min/1.73 m2 (excluding dialysis pa ien s). eGFR, es ima ed glome ula il a ion a e; SUCRA, su ace unde he cumula i e anking. 5 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ da a published in 2016, which concluded ha he e icacy o s a ins is modes in CKD47. Howe e , ou s udy canno con i m he e ec on ce eb o ascula e en s ha was p esen ed in a pai wise me a-analysis conduc ed in 201548, p obably because his ou come was included as a composi e o MACEs in he ials ha we e eligible o ou me a-analysis. Ou s udy also ound ha no e ec o eze imibe alone can be expec ed in indi iduals wi h CKD, and his inding is also consis en wi h he a ailable e idence54. S eng hs and limi a ions. To ou knowledge, his s udy is he i s o p esen a summa y o he e icacy o s a ins wi h o wi hou eze imibe in ela ion o educ ions in LDLc and CRP le els along he wide spec um o CKD. Impo an ly, moni o ing LDLc le els is no longe ecommended by he la es guidelines om he KDIGO lipid managemen wo k g oup, and i should be ese ed o ins ances in which he esul s would al e he man- agemen o pa ien ea men 10. Howe e , he KDIGO lipid managemen wo k g oup ecommends a ull lipid p o ile upon i s p esen a ion (a s a emen made on he basis o low-quali y e idence)10; acco ding o he indings o his me a-analysis, CRP should be measu ed in addi ion o LDLc be o e ini ia ing choles e ol-lowe ing ea - men in CKD pa ien s. Ou app oach gua an ees he p o ision o e idence ha pe inen ly answe s he o mula ed e iew ques ions. The e idence was ob ained om gene al sou ces o ocused sou ces acco ding o a me hod ha esembles he “mixed-c i e ia” quali y app aisal me hod o Wo man55. In addi ion, a mul idisciplina y supe ision mechanism was planned o con ex ualizing he sea ch indings56. Ne e heless, some limi a ions mus be men ioned. Fi s , only 3 pos hoc analyses o RCTs including pa ien s wi h and wi hou CKD p o ided e idence o he combined educ ion in LDLc and CRP in CKD pa ien s37–39. The c i icism agains a ibu ing mo e alue o he ou come o an unplanned analysis is di icul o igno e57. In addi ion, conside ing ha he e was no possibili y o pe o m me a- eg ession41, ou esul s should be conside ed p uden ly, pa icula ly conside ing he cu en pauci y o clinical esea ch on CVD in CKD3, which should highligh he need o mo e esea ch58. Second, publica ion bias is likely o a ec ou esul s. No unpublished s udies we e ound, bu asymme y in he unnel plo s may be a discou aging inding. Repo ing bias occu s because signi ican esul s sugges ing a bene icial e ec o in e en- ions a e mo e likely o be published han nonsigni ican esul s59. Cau ion is hus necessa y when in e p e ing he manusc ip , and common sense is equi ed when applying ou esul s in o e e yday clinical p ac ice. Finally, he powe o he ials e alua ing lo as a in and lu as a in was p obably insu icien , esul ing in an inapp op ia e anking o all s a ins in he ne wo k me a-analysis23,29. Once mo e, cau ion is needed. Fo in la ed associa ions, being ai wi h he in e p e a ion o he esul s is an impo an conside a ion60. Conclusion and s udy con ibu ion. In conclusion, s a ins wi h and wi hou eze imibe ha e a modes e icacy o educing CVD isk in pa ien s wi h CKD and eGFRs lowe han 60 ml/min/1.73 m2. I espec i e o Figu e 4. Ne wo k diag am o he compe ing choles e ol-lowe ing ea men s a egies (Ma ko chain Mon e Ca lo simula ion). 6 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ eGFR, i LDLc and CRP le els a e high be o e ea men , educ ions in such pa ame e s a e app op ia e o ob ain a bene i om choles e ol-lowe ing ea men . In o he wo ds, he use o s a ins should be ini ia ed when le els o LDLc and CRP a e high o educe he CVD isk o pa ien s. Impo an ly, pa ien s wi h ad anced CKD may p esen wi h low le els o LDLc. Such pa ien s ha e a highe isk o CVD e en s, much like he isk o pa ien s p esen ing wi h high le els o LDLc42. The bene i o choles e ol-lowe ing ea men in CKD pa ien s is e iden , bu he impac on hose wi h ad anced CKD emains unclea . Finally, in acco dance wi h he cu en guidelines, a e an ini ial measu emen , moni o ing le els o LDLc is p obably no equi ed o CKD pa ien s10. Howe e , Figu e 5. Ne wo k o es plo o he ixed and andom e ec s o he compe ing choles e ol-lowe ing ea men s a egies. C I, c edible in e als; OR, odds a io. 7 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ CRP should be assessed in addi ion o LDLc be o e ini ia ing choles e ol-lowe ing ea men . E idence-based ea men s a egies may lead o indi idualized he apies o CKD pa ien s. Me hods S udy design, eligibili y c i e ia and sys ema ic sea ches. These mul is age pai wise and ne wo k me a-analyses we e ca ied ou and epo ed acco ding o he PRISMA guidelines56. Two p o ocols we e egis- e ed in he In e na ional P ospec i e Regis e o Sys ema ic Re iews (PROSPERO) unde he ollowing egis a- ion IDs: CRD42017075166 and CRD42017055787. A wo-s age sys ema ic e iew design cons i u es he basis o ou mul is age app oach58,61. A sys ema ic mapping (choles e ol-lowe ing ea men e icacy e alua ion phase) ollowed by an in-dep h sys ema ic e iew ( ea men objec i es e alua ion phase) was planned, and he o me was complemen ed by an independen pa allel one-s age sys ema ic e iew (Ma e ial S1). An expe ad iso y g oup was o med o con ex ualize he sea ch indings (E.A., F.J.A., J.B.-M. and M.V.). The e iew ques ions co esponding o each o he 2 sys ema ic e iews and he pa icipan s/popula ion, in e en ion(s), and compa a o s a e p esen ed in Table2. The p ima y ou come was inciden MACEs, which included all a al and non a al co ona y e en s, including e ascula iza ion p ocedu es, and all ce eb o ascula e en s, including ansien ischemic a acks (TIAs). The seconda y ou come was ma e ial codependency when combining echnologies ela ed o ea men and he po en ial ea men e ec modi ie s o a low equency o MACEs. The eligible s udies we e all andomized con olled ials (RCTs) wi h ex ended ollow-up pe iods and pos hoc analysis ha we e ca ied ou on o ha included CKD pa ien s. In pa allel, published and unpublished li e a u e sou ces we e sea ched a Pha macological Big Da a Labo a o y, Pha macology and The apeu ics, Facul y o Medicine, Uni e si y o Valladolid, Valladolid, Spain and Cen o de In es igación en Salud Pública, Ins i u o de In es igación de la Facul ad de Medicina Humana, Uni e sidad de San Ma ín de Po es, Lima, Pe ú h ough Sep embe 2018 acco ding o he homogenous PICO elemen -based sea ch s a egies ha a e a ailable online om he sys ema ic e iew p o ocols egis e ed in PROSPERO (Ma e ial S2). Da a collec ion and analyses. Anonymized da ase s desc ibing cha ac e is ics o he s udies and hei pa icipan s, in e en ions, compa a o s, ou comes, and ollow-up pe iods we e cons uc ed. Be o e analysis, he isk o bias in he eligible s udies was assessed using he s anda d ool p oduced by he Coch ane Collabo a ion62. A wo-s age pai wise and ne wo k me a-analysis o agg ega e-le el da a was planned (C.O.-S. and F.H.-G.). Howe e , i he e we e no enough da a o i he e was a doub abou he compa abili y o he da a, a wo-s aged sys ema ic na a i e syn hesis was en isaged63. Figu e 6. Inconsis ency plo o he andom e ec s o he compe ing choles e ol-lowe ing ea men s a egies. T ea men objec i es†SUCRA‡ LDLc educ ion < 50% plus CRP educ ion < 50% 0.7719 LDLc educ ion ≥ 50% 0.6261 LDLc educ ion < 50% 0.5933 CRP educ ion < 50% 0.4995 None 0.0092 Table 1. The anking o he LDLc and CRP ea men objec i es based on SUCRA. †In a Bayesian con ex (Ma ko chain Mon e Ca lo simula ion), he LDLc and/o CRP ea men objec i es we e he pa ame e s used o anking acco ding o p obabili ies o being he bes , he second bes , he hi d bes , and so on =Po b() , =…b a1, , . ‡SUCRA o each ea men objec i e o ou o he a compe ing ea men objec i es equi es calcula ion o he a ec o o he cumula i e p obabili ies cumob, o be among he b bes ea men objec i es, =…b a1, , . Abb e ia ions: CRP, C- eac i e p o ein; LDL-c; low-densi y lipop o ein choles e ol; SUCRA, su ace unde he cumula i e anking. 8 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ The o e all odds a ios and hei 95% con idence in e als o he ou come o inciden MACEs co espond- ing o choles e ol-lowe ing ea men (s age 1) and he ea men objec i es o LDLc and/o CRP (s age 2) we e ob ained (Man el-Haenszel andom-e ec me hod) h ough con en ional pai wise me a-analysis. He e ogenei y was examined by compu ing he I² s a is ic (inconsis ency) and he χ² s a is ic, and he p esence o epo ing bias was assessed by isual inspec ion o unnel plo s o he es ima es agains hei s anda d e o s. A calcula ion o he eg ession coe icien co esponding o choles e ol-lowe ing ea men and he ea men objec i es o LDLc and/ o CRP (s age 2) was planned ( andom-e ec s me a- eg ession). Re iew Manage so wa e (Re Man) e sion 5.3 (Coch ane Collabo a ion) and he ‘me a eg’ mac o om S a a e sion 12.1 (S a aCo p) we e used o conduc he pai wise me a-analy ic calcula ions and me a- eg ession, espec i ely. Pooled ORs and he co esponding 95% c edible in e als o inciden MACEs (conside ed a ‘bad’ ou come) we e calcula ed ( andom-e ec s model using ague p io s wi h co ec ion o ze o alues) ia Bayesian ne wo k me a-analysis (Ma ko chain Mon e Ca lo sim- ula ion); he compe ing s a egies (s age 1) and ea men objec i es o LDLc and/o CRP (s age 2) we e anked based on SUCRA (su ace unde he cumula i e anking) a e e i ying con e gence (B ooks-Gelman-Rubin me hod) and inconsis ency. Ne wo k me a-analysis was pe o med using Ne Me aXL so wa e (Canadian Agency o D ugs and Technologies in Heal h and Co ne s one Resea ch G oup)64. Analysis o subg oups was planned a s age 1 (NKF/CKD KDIGO GFR ca ego ies G1–G2, CKD KDIGO GFR ca ego ies G3a–G5, ch onic dialysis, ansplan a ion) and s age 2 (LDLc educ ion < 50% combined wi h CRP educ ion < 50%, LDLc educ ion < 50%, LDLc educ ion ≥ 50%, CRP educ ion < 50%). Codependency when combining echnologies ela ed o he ea men and he po en ial ea men e ec mod- i ie s o a low equency o MACEs (LDLc combined wi h CRP) was assessed using an adap a ion o Me lin’s ool included in he guidelines o p epa ing a submission o he Pha maceu ical Bene i s Ad iso y Commi ee (PBAC) om he Depa men o Heal h o Aus alia (F.H.-G. and A.M.-B.-M.)65. The ool sec ions o he eco- nomic e alua ion and use o he medicine in p ac ice we e no conside ed. Da a A ailabili y All da a gene a ed and analyzed du ing his s udy a e included in his published a icle and he supplemen a y ma e ials. Re e ences 1. S en inkel, P. Ch onic kidney disease: a public heal h p io i y and ha binge o p ema u e ca dio ascula disease. J. In e n. Med. 268, 456–467, h ps://doi.o g/10.1111/j.1365-2796.2010.02269.x (2010). 2. Yousu , O. e al. High-sensi i i y C- eac i e p o ein and ca dio ascula disease: a esolu e belie o an elusi e link? J. Am. Coll. Ca diol. 62, 397–408, h ps://doi.o g/10.1016/j.jacc.2013.05.016 (2013). 3. Fo men ini, I. e al. Cu en d ug de elopmen challenges in ch onic kidney disease (CKD)–iden i ica ion o indi idualized de e minan s o enal p og ession and p ema u e ca dio ascula disease (CVD). Neph ol. Dial. T ansplan . 27, iii81–iii88, h ps:// doi.o g/10.1093/nd /g s270 (2012). 4. Joshi, S. & Viljoen, A. Renal bioma ke s o he p edic ion o ca dio ascula disease. Cu . Opin. Ca diol. 30, 454–460, h ps://doi. o g/10.1097/HCO.0000000000000177 (2015). 5. Zoccali, C. T adi ional and eme ging ca dio ascula and enal isk ac o s: an epidemiologic pe spec i e. Kidney In . 70, 26–33, h ps://doi.o g/10.1038/sj.ki.5000417 (2006). 6. Bai, J. P. e al. T ansla ional bioma ke s: om p eclinical o clinical a epo o 2009 AAPS/ACCP Bioma ke Wo kshop. AAPS J. 13, 274–283, h ps://doi.o g/10.1208/s12248-011-9265-x (2011). 7. Me lin, T. e al. Assessing pe sonalized medicines in Aus alia: a na ional amewo k o e iewing codependen echnologies. Med. Decis. Making. 33, 333–342, h ps://doi.o g/10.1177/0272989X12452341 (2013). 8. Tonelli, M. e al. Risk o co ona y e en s in people wi h ch onic kidney disease compa ed wi h hose wi h diabe es: a popula ion- le el coho s udy. Lance 380, 807–814, h ps://doi.o g/10.1016/S0140-6736(12)60572-8 (2012). 9. Massy, Z. A. & de Zeeuw, D. LDL choles e ol in CKD– o ea o no o ea ? Kidney In . 84, 451–456, h ps://doi.o g/10.1038/ ki.2013.181 (2013). Sys ema ic mapping (s age 1)/ sys ema ic e iew suppo In-dep h me a-analysis (s age 2) Re iew ques ion A e s a ins wi h o wi hou eze imibe e icacious in educing CVD isk in pa ien s wi h CKD?§ A e se um le els o LDLc and CRP ela ed o CVD e en s in pa ien s wi h CKD ecei ing ea men wi h s a ins alone o in combina ion wi h eze imibe? Pa icipan s/popula ion Adul indi iduals wi h NKF/CKD KDIGO GFR ca ego ies G1–G2, pa ien s wi h CKD KDIGO GFR ca ego ies G3a– G5,# and pa ien s ea ed wi h ch onic dialysis o kidney ansplan a ion.§ Adul indi iduals wi h NKF/CKD KDIGO GFR ca ego ies G1–G2, pa ien s wi h CKD KDIGO GFR ca ego ies G3a– G5, and pa ien s ea ed wi h ch onic dialysis o kidney ansplan a ion. In e en ion(s)/exposu es(s) S a ins wi h o wi hou eze imibe.§Se um le els o LDLc and/o CRP in pa ien s ea ed wi h s a ins wi h o wi hou eze imibe. Compa a o s Placebo/usual ca e.@Se um le els o LDLc and/o CRP unde placebo/usual ca e.@ Table 2. Re iew ques ions and s udy eligibili y. §Only s a ins we e s udied, and pa ien s ea ed wi h ch onic dialysis o kidney ansplan a ion we e included in he sys ema ic e iew o suppo sys ema ic mapping. #Indi iduals wi h CKD we e di ided acco ding o eGFR (in ml/min/1.73 m2): ≥60 (KDIGO GFR ca ego ies G1–G2) o < 60 (G3a–G5). @Usual ca e may include a s a in i a s a in a a highe dose o a s a in/eze imibe combina ion was al eady he in e en ion. Abb e ia ions: CKD, ch onic kidney disease; CRP, C- eac i e p o ein; CVD, ca dio ascula disease; eGFR, es ima ed glome ula il a ion a e; KDIGO, he Kidney Disease: Imp o ing Global Ou comes; LDLc, low-densi y lipop o ein choles e ol; NKF, no mal kidney unc ion. 9 Scien i ic RepoR s | (2019) 9:8951 | h ps://doi.o g/10.1038/s41598-019-45431-5 www.na u e.com/scien i ic epo s www.na u e.com/scien i ic epo s/ 10. KDIGO Clinical P ac ice Guideline o Lipid Managemen in CKD: summa y o ecommenda ion s a emen s and clinical app oach o he pa ien . Kidney In . 85, 1303–1309, h ps://doi.o g/10.1038/ki.2014.31 (2014). 11. Holdaas, H. e al. E ec o lu as a in on ca diac ou comes in enal ansplan ecipien s: a mul icen e, andomised, placebo- con olled ial. Lance 361, 2024–2031, h ps://doi.o g/10.1016/S0140-6736(03)13638-0 (2003). 12. Holdaas, H. e al. Long- e m ca diac ou comes in enal ansplan ecipien s ecei ing lu as a in: he ALERT ex ension s udy. Am. J. T ansplan . 5, 2929–2936, h ps://doi.o g/10.1111/j.1600-6143.2005.01105.x (2005). 13. Wanne , C. e al. A o as a in in pa ien s wi h ype 2 diabe es melli us unde going hemodialysis. N. Engl. J. Med. 353, 238–248, h ps://doi.o g/10.1056/NEJMoa043545 (2005). 14. K a ne , V. e al. Long- e m e ec s ollowing 4 yea s o andomized ea men wi h a o as a in in pa ien s wi h ype 2 diabe es melli us on hemodialysis. Kidney In . 89, 1380–1387, h ps://doi.o g/10.1016/j.kin .2015.12.033 (2016). 15. Fells öm, B. C. e al. Rosu as a in and ca dio ascula e en s in pa ien s unde going hemodialysis. N. Engl. J. Med. 360, 1395–1407, h ps://doi.o g/10.1056/NEJMoa0810177 (2009). 16. S egmay , B. G. e al. Low-dose a o as a in in se e e ch onic kidney disease pa ien s: a andomized, con olled endpoin s udy. Scand. J. U ol. Neph ol. 39, 489–497, h ps://doi.o g/10.1080/00365590500329304 (2005). 17. Baigen , C. e al. The e ec s o lowe ing LDL choles e ol wi h sim as a in plus eze imibe in pa ien s wi h ch onic kidney disease (S udy o Hea and Renal P o ec ion): a andomised placebo-con olled ial. Lance 377, 2181–2192, h ps://doi.o g/10.1016/ S0140-6736(11)60739-3 (2011). 18. Land ay, M. e al. The second Uni ed Kingdom Hea and Renal P o ec ion (UK-HARP-II) S udy: a andomized con olled s udy o he biochemical sa e y and e icacy o adding eze imibe o sim as a in as ini ial he apy among pa ien s wi h CKD. Am. J. Kidney Dis. 47, 385–395, h ps://doi.o g/10.1053/j.ajkd.2005.11.018 (2006). 19. Ueshima, K. e al. E ec s o a o as a in on enal unc ion in pa ien s wi h dyslipidemia and ch onic kidney disease: a ionale and design o he ASsessmen o clinical Use ulness in CKD pa ien s wi h A o as a in (ASUCA) ial. Clin. Exp. Neph ol. 17, 211–217, h ps://doi.o g/10.1007/s10157-012-0676-5 (2013). 20. Nanayakka a, P. W. e al. E ec o a ea men s a egy consis ing o p a as a in, i amin E, and homocys eine lowe ing on ca o id in ima-media hickness, endo helial unc ion, and enal unc ion in pa ien s wi h mild o mode a e ch onic kidney disease: esul s om he An i-Oxidan The apy in Ch onic Renal Insu iciency (ATIC) S udy. A ch. In e n. Med. 167, 1262–1270, h ps://doi. o g/10.1001/a chin e.167.12.1262 (2007). 21. Fasse , R. G., Robe son, I. K., Ball, M. J., Ge agh y, D. P. & Coombes, J. S. E ec o a o as a in on kidney unc ion in ch onic kidney disease: a andomised double-blind placebo-con olled ial. A he oscle osis 213, 218–224, h ps://doi.o g/10.1016/j. a he oscle osis.2010.07.053 (2010). 22. Chonchol, M., Cook, T., Kjekshus, J., Pede sen, T. R. & Linden eld, J. Sim as a in o seconda y p e en ion o all-cause mo ali y and majo co ona y e en s in pa ien s wi h mild ch onic enal insu iciency. Am. J. Kidney Dis. 49, 373–382, h ps://doi.o g/10.1053/j. ajkd.2006.11.043 (2007). 23. Kend ick, J. e al. E ec o lo as a in on p ima y p e en ion o ca dio ascula e en s in mild CKD and kidney unc ion loss: a pos hoc analysis o he Ai Fo ce/Texas Co ona y A he oscle osis P e en ion S udy. Am. J. Kidney Dis. 55, 42–49, h ps://doi. o g/10.1053/j.ajkd.2009.09.020 (2010). 24. Rahman, M. e al. P a as a in and ca dio ascula ou comes s a i ied by baseline eGFR in he lipid- lowe ing componen o ALLHAT. Clin. Neph ol. 80, 235–248, h ps://doi.o g/10.5414/CN107922 (2013). 25. Ko en, M. J. e al. Focused a o as a in he apy in managed-ca e pa ien s wi h co ona y hea disease and CKD. Am. J. Kidney Dis. 53, 741–750, h ps://doi.o g/10.1053/j.ajkd.2008.11.025 (2009). 26. Colhoun, H. M. e al. E ec s o a o as a in on kidney ou comes and ca dio ascula disease in pa ien s wi h diabe es: an analysis om he Collabo a i e A o as a in Diabe es S udy (CARDS). Am. J. Kidney Dis. 54, 810–819, h ps://doi.o g/10.1053/j. ajkd.2009.03.022 (2009). 27. Holme, I. e al. Ca dio ascula ou comes and hei ela ionships o lipop o ein componen s in pa ien s wi h and wi hou ch onic kidney disease: esul s om he IDEAL ial. J. In e n. Med. 267, 567–575, h ps://doi.o g/10.1111/j.1365-2796.2009.02176.x (2010). 28. Ridke , P. M., MacFadyen, J., C essman, M. & Glynn, R. J. E icacy o osu as a in among men and women wi h mode a e ch onic kidney disease and ele a ed high-sensi i i y C- eac i e p o ein: a seconda y analysis om he JUPITER (Jus i ica ion o he Use o S a ins in P ima y P e en ion–an In e en ion T ial E alua ing Rosu as a in) ial. J. Am. Coll. Ca diol. 55, 1266–1273, h ps://doi. o g/10.1016/j.jacc.2010.01.020 (2010). 29. Lemos, P. A. e al. Long- e m lu as a in educes he haza dous e ec o enal impai men on ou -yea a he oscle o ic ou comes (a LIPS subs udy). Am. J. Ca diol. 95, 445–451, h ps://doi.o g/10.1016/j.amjca d.2004.10.008 (2005). 30. Nakamu a, H. e al. P a as a in and ca dio ascula isk in mode a e ch onic kidney disease. A he oscle osis 206, 512–517, h ps:// doi.o g/10.1016/j.a he oscle osis.2009.03.031 (2009). 31. Deedwania, P. C., S one, P. H., Fayyad, R. S., Laskey, R. E. & Wilson, D. J. Imp o emen in Renal Func ion and Reduc ion in Se um U ic Acid wi h In ensi e S a in The apy in Olde Pa ien s: A Pos Hoc Analysis o he SAGE T ial. D ugs Aging. 32, 1055–1065, h ps://doi.o g/10.1007/s40266-015-0328-z (2015). 32. Shephe d, J. e al. In ensi e lipid lowe ing wi h a o as a in in pa ien s wi h co ona y hea disease and ch onic kidney disease: he TNT (T ea ing o New Ta ge s) s udy. J. Am. Coll. Ca diol. 51, 1448–1454, h ps://doi.o g/10.1016/j.jacc.2007.11.072 (2008). 33. Tonelli, M. e al. E ec o p a as a in on ca dio ascula e en s in people wi h ch onic kidney disease. Ci cula ion 110, 1557–1563, h ps://doi.o g/10.1161/01.CIR.0000143892.84582.60 (2004). 34. Asselbe gs, F. W. e al. E ec s o osinop il and p a as a in on ca dio ascula e en s in subjec s wi h mic oalbuminu ia. Ci cula ion 110, 2809–2816, h ps://doi.o g/10.1161/01.CIR.0000146378.65439.7A (2004). 35. B ouwe s, F. P. e al. Long- e m e ec s o osinop il and p a as a in on ca dio ascula e en s in subjec s wi h mic oalbuminu ia: Ten yea s o ollow-up o P e en ion o Renal and Vascula End-s age Disease In e en ion T ial (PREVEND IT). Am. Hea J. 161, 1171–1178, h ps://doi.o g/10.1016/j.ahj.2011.03.028 (2011). 36. Mohe , D. e al. P e e ed epo ing i ems o sys ema ic e iews and me a-analyses: he PRISMA s a emen . PLoS Med 6, e1000097, h ps://doi.o g/10.1371/jou nal.pmed.1000097 (2009). 37. Ridke , P. M. e al. Measu emen o C- eac i e p o ein o he a ge ing o s a in he apy in he p ima y p e en ion o acu e co ona y e en s. N. Engl. J. Med. 344, 1959–1965, h ps://doi.o g/10.1056/NEJM200106283442601 (2001). 38. Soedamah-Mu hu, S. S. e al. E ec o a o as a in on C- eac i e p o ein and bene i s o ca dio ascula disease in pa ien s wi h ype 2 diabe es: analyses om he Collabo a i e A o as a in Diabe es T ial. Diabe ologia 58, 1494–1502, h ps://doi.o g/10.1007/s00125- 015-3586-8 (2015). 39. Ridke , P. M. e al. Reduc ion in C- eac i e p o ein and LDL choles e ol and ca dio ascula e en a es a e ini ia ion o osu as a in: a p ospec i e s udy o he JUPITER ial. Lance 373, 1175–1182, h ps://doi.o g/10.1016/S0140-6736(09)60447-5 (2009). 40. Salan i, G., Ades, A. E. & Ioannidis, J. P. G aphical me hods and nume ical summa ies o p esen ing esul s om mul iple- ea men me a-analysis: an o e iew and u o ial. J. Clin. Epidemiol. 64, 163–171, h ps://doi.o g/10.1016/j.jclinepi.2010.03.016 (2010). 41. Bake , W. L. e al. Unde s anding he e ogenei y in me a-analysis: he ole o me a- eg ession. In . J. Clin. P ac . 63, 1426–1434, h ps://doi.o g/10.1111/j.1742-1241.2009.02168.x (2009).