A PHARMACOME T RI C-BASE D APPROACH T O F ACI L I T AT E - PowerPoint PPT Presentation

A PHARMACOME T RI C-BASE D APPROACH T O F ACI L I T AT E CRI T I CAL L Y NE E DE D NE W ANT I BI OT I C DE VE L OPME NT Paul G. Ambr ose , Phar m.D., F IDSA I nstitute fo r Clinic a l Pha rma c o dyna mic s L a tha m, Ne w Yo rk Ho no ra ry Re se a rc h F e llo w, I nfe c tio us Dise a se s Nuffie ld De pa rtme nt o f Me dic ine Unive rsity o f Oxfo rd Oxfo rd, E ng la nd, Unite d K ing do m

T HE PROBL E M Re gulato r y Unc e r tainty • Antimic ro b ia l drug de ve lo pme nt ha s g re a tly diminishe d due to the impa c t o f re g ula to ry unc e rta inty re g a rding T he magnitude o f dr ug e ffe c t a nd o T he a b ility o f c o nte mpo ra ry c linic al tr ial e ndpo ints to c aptur e o dr ug be ne fit • T he se e le me nts a re c ritic a l to sta tistic a lly po we r no n- infe rio rity studie s T he la c k o f this info rma tio n unde rmine s the via b ility o f o c o nte mpo ra ry c linic a l tria l de sig ns

HI ST ORI CAL DAT A E XAMI NAT I ON One So lutio n • One a ppro a c h to g a in kno wle dg e a b o ut the ma g nitude o f tre a tme nt e ffe c t is thro ug h the e xa mina tio n o f histo ric a l da ta • Ho we ve r, the re a re o bvio us limitatio ns to histo ric a l da ta tha t c a n intro duc e po te ntia l b ia s No pla c e b o c o ntro l, o L a c k o f b linding a nd ra ndo miza tio n, o Use o f me dic a lly dub io us inte rve ntio ns, a nd o I na b ility to a uthe ntic a te the so urc e da ta o • So urc e da ta ve rific a tio n is a n e sse ntia l c o mpo ne nt in Go o d Clinic a l Pra c tic e g uide line s

HI ST ORI CAL DAT A E XAMI NAT I ON One So lutio n • Using histo ric a l da ta , the US F DA a nd o the rs ha ve prima rily e mplo ye d a sing le sta tistic a l a ppro a c h F re q ue ntist infe re nc e o • Alte rna tive sta tistic a l a ppro a c he s ha ve no t b e e n fully (a nic e wa y o f sa ying “no t a t a ll”) c o nside re d b y the US F DA o r o the rs F re q ue ntist infe re nc e c o mb ine d with pha rma c o me tric o me tho ds Ba ye sia n infe re nc e c o mb ine d with pha rma c o me tric me tho ds o

BAYE SI AN VE RSUS F RE QUE NT I ST S T he Ke y Diffe r e nc e s • F re q ue ntist a na lyse s do no t inc o rpo ra te prio r info rma tio n o r b e lie fs I n a fre q ue ntist a na lysis, me a n is c o nside re d to b e a re a l o va lue a nd o nly the study da ta de te rmine s nume ric a l re sults • I n a Ba ye sia n a na lysis, b o th prio r info rma tio n a nd the study da ta influe nc e the re sults a nd c o nc lusio n Sinc e the po pula tio n me a n is no t c o nside re d fixe d in a o Ba ye sia n a na lyse s, prio r info rma tio n c o nc e rning the like liho o d o f the me a n to fa ll within spe c ifie d inte rva ls c a n b e q ua ntifie d a nd inc o rpo ra te d

A BAYE SI AN APPROACH T hr e e Re aso ns this is Ratio nal • F irst, tig e c yc line ha s de mo nstra te d in vitro a c tivity a g a inst b a c te ria o f c linic a l inte re st • Se c o nd, tig e c yc line wa s e va lua te d in a nima l infe c tio n mo de ls, whe re it wa s e ffe c tive in b a c te ria l killing T ig e c yc line e xpo sure , a s me a sure d o b y the AUC 0-24 :MI C ra tio , a nd o utc o me we re we ll-de sc rib e d b y e xpo sure -re spo nse mo de ls va n Og tro p ML , Ande s D, Sta msta d T J, Co nklin B, We iss WJ, Cra ig WA, a nd Ve sg a O. I n vivo pha rma c o dyna mic a c tivitie s o f two g lyc ylc yc line s a g a inst Gra m-po sitive a nd Gra m-ne g a tive b a c te ria . Antimic ro b. Ag e nts Che mo the r. 2000 44: 943-949.

A BAYE SI AN APPROACH T hr e e Re aso ns this is Ratio nal • T hird, tig e c yc line wa s studie d in pa tie nts fo r who m e ffic a c y wa s de mo nstra te d a nd re la te d to AUC 0-24 :MI C ra tio using e xpo sure -re spo nse mo de ls • I n o the r wo rds, antimic r o bial age nts tha t e nte r la te - sta g e c linic a l de ve lo pme nt ha ve b e e n e ffe c tive ly pr e -sc r e e ne d fo r e ffic ac y As suc h, it is hig hly like ly tha t suc h a g e nts wo uld b e o e ffic a c io us if a ppro pria te do se re g ime ns a re se le c te d fo r c linic a l study

OBJE CT I VE S A Ne w Appr o ac h • Our o b je c tive wa s to de mo nstra te the utility o f fre q ue ntist- a nd Ba ye sia n-pha rma c o me tric -b a se d lo g istic re g re ssio n a na lyse s • De te rmine the ma g nitude o f tre a tme nt e ffe c t • De te rmine the a b ility o f c linic a l tria l e ndpo ints to c a pture drug b e ne fit • A HAP/ VAP c linic a l tria l da ta se t wa s se le c te d fo r a na lysis Co nte mpo ra ry Pha se 3 tria l invo lving tig e c yc line o L a b e ling indic a tio ns ha ve g e ne ra te d c o nside ra b le o disc ussio ns a t US F DA a nd I DSA c o -spo nso re d wo rksho ps a nd in Guida nc e Do c ume nts

ME T HODS Statistic s • L o g istic re g re ssio n mo de ls T he de pe nde nt va ria b le s we re mic ro b io lo g ic a l a nd c linic a l o suc c e ss T he inde pe nde nt va ria b le in e a c h a na lysis wa s the lo g 10 o tra nsfo rm o f the AUC 0-24 :MI C ra tio • F re q ue ntisit lo g istic re g re ssio n a na lyse s we re c a rrie d o ut in the sta nda rd fa shio n b a se d o n ma ximum like liho o d e stima tio n

ME T HODS Statistic s • Ba ye sia n lo g istic re g re ssio n a na lyse s we re pe rfo rme d e mplo ying a no n-info rma tive unifo rm prio r distrib utio n fo r the inte rc e pt pa ra me te r • Ba ye sia n lo g istic re g re ssio n a na lyse s we re pe rfo rme d e mplo ying a no rma l prio r distrib utio n fo r the slo pe pa ra me te r c o rre spo nding to drug e xpo sure No rma l prio r e stima te s fo r the Ba ye sia n a na lyse s we re o se le c te d to re fle c t a 99% prio r like liho o d o f tre a tme nt e ffe c t within a n inte rva l o f -0.25 to 0.75  F a vo rs a po sitive tre a tme nt e ffe c t (drug he lps pa tie nts)  Allo ws the po ssib ility o f a ne g a tive tre a tme nt e ffe c t (drug hurts pa tie nts)

ME T HODS Statistic s • T re a tme nt e ffe c t wa s de fine d a s the diffe re nc e in like liho o d o f suc c e ss a t lo w a nd hig h ma g nitude s o f drug e xpo sure L o w a nd hig h e xpo sure s we re o fixe d a t fre e -drug AUC 0-24 :MI C ra tio s o f 0.01 a nd 25, re spe c tive ly • 95% lo we r b o unds fo r tre a tme nt e ffe c t we re o b ta ine d using 1,000 b o o tstra p sa mple s a nd the b ia s-c o rre c ting a c c e le ra tio n me tho d va n Og tro p ML , Ande s D, Sta msta d T J, Co nklin B, We iss WJ, Cra ig WA, a nd Ve sg a O. I n vivo pha rma c o dyna mic a c tivitie s o f two g lyc ylc yc line s a g a inst Gra m-po sitive a nd Gra m-ne g a tive b a c te ria . Antimic ro b. Ag e nts Che mo the r. 2000 44: 943-949.

RE SUL T S Analysis Po pulatio n • 61 pa tie nts who we re b o th c linic a lly e va lua b le a nd mic ro b io lo g ic a lly e va lua b le a nd ha d suffic ie nt PK da ta we re inc lude d in the se a na lyse s • Clinic a l a nd mic ro b io lo g ic a l re spo nse we re hig hly c o nc o rda nt (95%, 58/ 61) 42 pa tie nts we re c a te g o rize d b o th a s c linic a l a nd o mic ro b io lo g ic a l suc c e sse s 16 pa tie nts we re c a te g o rize d b o th a s c linic a l a nd o mic ro b io lo g ic a l fa ilure s 2 pa tie nts we re c a te g o rize d mic ro b io lo g ic a l b ut no t c linic a l o suc c e sse s 1 pa tie nt wa s c a te g o rize d a s a c linic a l b ut no t mic ro b io lo g ic a l o suc c e ss Bha vna ni SM, Rub ino CM, Ha mme l JP, F o rre st A, Da rto is N, Co o pe r A, K o rth-Bra dle y J, Amb ro se PG. Pha rma c o lo g ic a l a nd Pa tie nt-Spe c ific Re spo nse De te rmina nts in Pa tie nts with Ho spita l-Ac q uire d Pne umo nia T re a te d with T ig e c yc line . Antimic ro b Age nts Che mo the r. Awa iting Dispo sitio n .

RE SUL T S F r e que ntist L o gistic Re gr e ssio n, Clinic al Endpo int 1.0 1.0 Probability of Clinical Success 0.8 0.8 Relative Frequency 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.01 0.1 1 10 25 Free-Drug AUC 0  24 :MIC Ratio Amb ro se PG, Ha mme l JP, Bha vna ni SM, Rub ino CM, E llis-Gro sse E J, Drusa no GL . A Ba ye sia n Pha rma c o me tric -Ba se d Appro a c h to F a c ilita te Critic a lly Ne e de d Ne w Antib io tic De ve lo pme nt: Ove rc o ming L ie s, Da mn L ie s a nd Sta tistic s. Antimic ro b Age nts Che mo the r. Awa iting Dispo sitio n .

RE SUL T S Baye sian L o gistic Re gr e ssio n, Clinic al Endpo int 1.0 1.0 Probability of Clinical Success 0.8 0.8 Relative Frequency 0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 0.01 0.1 1 10 25 Free-Drug AUC 0  24 :MIC Ratio Amb ro se PG, Ha mme l JP, Bha vna ni SM, Rub ino CM, E llis-Gro sse E J, Drusa no GL . A Ba ye sia n Pha rma c o me tric -Ba se d Appro a c h to F a c ilita te Critic a lly Ne e de d Ne w Antib io tic De ve lo pme nt: Ove rc o ming L ie s, Da mn L ie s a nd Sta tistic s. Antimic ro b Age nts Che mo the r. Awa iting Dispo sitio n .