Originl Article Cittion: CPT: Phrmcometrics & Systems Phrmcology (13), e73; doi:1.138/psp.13.7 13 ASCPT All rights reserved 163-836/1 www.nture.com/psp An Integrted Popultion Phrmcokinetic Met-Anlysis of Propofol in Morbidly Obese nd Nonobese Adults, Adolescents, nd Children J Diepstrten 1, V Chidmbrn,3, S Sdhsivm,3, HJ Blussé vn Oud-Albls, T Inge 5, B vn Rmshorst 6, EPA vn Dongen 7, AA Vinks 3,8 nd CAJ Knibbe 1,9 This study describes popultion phrmcokinetic met-nlysis of propofol to chrcterize the influence of body size mesures nd ge in morbidly obese nd nonobese dults, dolescents, nd children. Sixty morbidly obese nd nonobese dult ptients (55 167 kg; 1 79 yers) nd 3 morbidly obese nd nonobese dolescents nd children (37 18 kg; 9 yers) were included. The results show tht clernce incresed with totl body weight in n llometric function while ge ws found to influence clernce in biliner fshion with two distinct slopes, reflecting n initil increse nd subsequent decrese s result of ging. Using these two functions, the influence of both (over)weight nd ge on propofol clernce ws well chrcterized, which my provide bsis for dosing cross this diverse group of ptients. CPT: Phrmcometrics & Systems Phrmcology (13), e73; doi:1.138/psp.13.7; published online 11 September 13 BACKGROUND Although totl body weight (TBW) of children nd dolescents increses due to growth-relted processes cross childhood, obesity my lso substntilly contribute to increses in body weight. 1 As result, morbidly obese children nd dolescents my be s hevy s dults, even though growth-relted processes hve not yet been completed. The question then rises whether TBW, which is commonly used to djust dosing in children nd dolescents, is the pproprite mesure to djust doses of drugs in obese children nd dolescents. Similrly for dults, there is lively discussion bout the best size descriptor for chnges in phrmcokinetics due to obesity.,3 As little is known on how key phrmcokinetic prmeters such s clernce chnge in morbidly obese children, dolescents, nd dults s compred with their nonobese controls, studies re needed nlyzing wide rnge of ges nd relted TBWs. Propofol is widely used for induction nd mintennce of nesthesi in both nonobese nd (morbidly) obese dults, dolescents, nd children. Recently, the phrmcokinetics of propofol hve been compred in premture neontes nd dults, in morbidly obese nd obese dults, 5,6 nd in (morbidly) obese children nd dolescents. 7 In ll these studies, TBW proved the most predictive covrite for clernce, either by using stndrd llometric function 5 7 or TBWdependent exponent llometric function. However, metnlysis on the bsis of ll dt sets in morbidly obese dults, dolescents, nd children together with their nonobese controls in which the influence of obesity nd geing is disentngled hs not been performed. Therefore, the im of this study ws to perform popultion phrmcokinetic met-nlysis of propofol combining dt from morbidly obese nd nonobese dults, dolescents, nd children. To study how obesity nd ge influence phrmcokinetic prmeter estimtes in this diverse ptient group, specific emphsis ws plced on the evlution of the influence of TBW, body mss index, idel body weight, 8 len body weight (LBW), 9,1 nd/or ge on the different phrmcokinetic prmeters. RESULTS Subjects Ninety-four dults, dolescents, nd children with men TBW of 9 kg (rnge: 37 18 kg) were included from which 1,65 concentrtion mesurements were vilble. Demogrphic chrcteristics of the morbidly obese nd nonobese ptients re summrized in Tble 1. Phrmcokinetics A three-comprtment phrmcokinetic model dequtely described the time course of the propofol whole-blood concentrtions in ll morbidly obese nd nonobese dults, dolescents, nd children. Explortory plots of the tested covrites TBW, body mss index, idel body weight, LBW, nd ge ginst individul post hoc prmeter estimtes of the simple model without covrites (model A) showed potentil reltion between clernce nd TBW, with lower vlues for children nd dolescents cross the entire body weight rnge (Figure 1, model A). In ddition, potentil reltionships were observed between centrl volume of distribution (V1) nd 1 Deprtment of Clinicl Phrmcy, St. Antonius Hospitl, Nieuwegein, The Netherlnds; Deprtment of Anesthesiology, Cincinnti Children s Hospitl Medicl Center, Cincinnti, Ohio, USA; 3 Deprtment of Peditrics, College of Medicine, University of Cincinnti, Cincinnti, Ohio, USA; Deprtment of Anesthesiology, Ersmus Medicl Centre, Rotterdm, The Netherlnds; 5 Deprtment of Surgery, Division of Peditric Generl nd Thorcic Surgery, Cincinnti Children s Hospitl Medicl Center, Cincinnti, Ohio, USA; 6 Deprtment of Surgery, St. Antonius Hospitl, Nieuwegein, The Netherlnds; 7 Deprtment of Anesthesiology nd Intensive Cre, St. Antonius Hospitl, Nieuwegein, The Netherlnds; 8 Division of Clinicl Phrmcology, Cincinnti Children s Hospitl Medicl Center, Cincinnti, Ohio, USA; 9 Division of Phrmcology, Leiden/Amsterdm Center for Drug Reserch, Leiden, the Netherlnds. Correspondence: CAJ Knibbe (c.knibbe@ntoniusziekenhuis.nl) Received 16 April 13; ccepted 9 July 13; dvnce online publiction 11 September 13. doi:1.138/psp.13.7
Phrmcokinetic Met-Anlysis of Propofol TBW or LBW, nd between intercomprtmentl clernce from the centrl to the second peripherl comprtment (Q3) nd TBW (figures not shown). There were no visul trends between the explored covrites nd other phrmcokinetic prmeters in the simple model without covrites (model A). Subsequently, s depicted in Tble, ll body size mesures nd ge were seprtely incorported on clernce, V1, nd Q3 in the model nd tested for significnce (see section Covrite nlysis). The nlysis showed tht TBW ws the most predictive covrite for propofol clernce when implemented using n llometric function (model B, decrese in objective function vlue (OFV) of 8. points; P <.1; Tble ). Figure 1 model B (nd model A) shows tht dolescents with the sme TBW s dults hd lower clernce vlues (gry vs. blck symbols, respectively). Therefore, in model C, seprte vlue for propofol clernce in dolescents vs. dults ws estimted. This resulted in nother reduction in OFV by 3.5 points (P <.1) with individul clernce vlues for n dolescent of 7 kg nd n dult of 7 kg of 1.75 ml/min nd.18 ml/min, respectively (Tble, Model C). The nonliner increse of propofol clernce with TBW proved the sme for both groups nd ws best described using n llometric function with n estimted exponent of.73 (coefficients of vrition percentge: 6.9) (Figure 1, Model C). However, when the simple model without covrites ws evluted for the effect of ge (Figure, model A), it ws found tht clernce incresed until the medin ge of 1 Tble 1 Bseline chrcteristics of ll morbidly obese nd nonobese dults, dolescents, nd children included in the current nlysis All Ptients Adults Adolescents nd children Morbidly obese 6 Nonobese 11,1 Obese 7 Nonobese 13 Number 9 1 Gender (M/F) 3/6 /16 16/ 8/1 /1 Age (yers) 38 () 5 (1) 55 (1) 16 () 1 (3) TBW (kg) 9 (35) 1 () 7 (11) 15 (9) 5 (13) BMI (kg/m ) 33 (1) 3 (6) 6 () 6 (9) 1 (6) IBW (kg) 61 (9) 61 (7) 6 (8) 59 (1) 55 (9) LBW (kg) 5 (1) 6 (9) 5 (1) 63 (1) 37 (8) Dt re presented s men with stndrd devition (SD). BMI, body mss index; F, femle; IBW, idel body weight; 8 LBW, len body weight; 9 M, mle; SD, stndrd devition; TBW, totl body weight. 6 Model A 6 Model B 5 1 Totl body weight (kg) 15 5 1 Totl body weight (kg) 15 6 Model C 6 Model D 1 yers 65 yers 15 yers 5 1 Totl body weight (kg) 15 5 1 Totl body weight (kg) 15 Figure 1 Individul post hoc propofol clernce estimtes vs. totl body weight for the simple model (model A) nd three covrite phrmcokinetic models (B, C, nd D) for morbidly obese dults (blck circles), dolescents nd children (gry circles) nd their nonobese controls (n = 9). In model B, the blck line indictes the popultion clernce vlues for both the dult nd dolescent popultion; in model C, the blck line indictes the popultion clernce vlues for dults nd the gry line the popultion clernce vlues for dolescents; nd in model D, the blck dotted lines indicte the popultion clernce vlues for 15, 1, nd 65 yers. CPT: Phrmcometrics & Systems Phrmcology
Phrmcokinetic Met-Anlysis of Propofol 3 Tble Results of covrite nlysis for the three-comprtment phrmcokinetic model of propofol in the combined dt set of morbidly obese nd nonobese dults, dolescents, nd children Prmeter Model Model description Reltionship of covrite Number of structurl prmeters OFV A Simple model 11 CL Age biliner F ge 13 16.7 CL LBW (9) liner (LBW i /6) 11 5.8 CL TBW liner /7) 11 68.5 CL B TBW llometric /7) z 1 8. CL C TBW llometric, CL dolescents, nd CL dults,dolescents /7) z 13 17.9, dults /7) z CL D TBW llometric nd ge biliner /7) z F ge 1 116.1 V1 TBW liner V1 i = V1 pop /7) 11. V1 LBW 9 liner V1 i = V1 pop (LBW i /6) 11 5.8 Q3 TBW liner Q3 i = Q3 pop /7) 11 18.1 Q3 TBW llometric Q3 i = Q3 pop /7) d 1 18.1 Finl model CL nd Q3 E TBW llometric nd ge biliner on CL nd TBW liner on Q3 /7) z F ge 1 13. Q3 i = Q3 pop /7) Age 1 yers: F ge = (1 + b (Age 1)) nd Age > 1 yers: F ge = (1 + c (Age 1)). CL, clernce;, clernce in ith individul; CL pop, popultion men vlue for clernce; F ge, ge fctor for clernce; LBW, len body weight; OFV, delt objective function vlue s compred with simple model; Q3, comprtmentl clernce between V1 nd V3; TBW, totl body weight; V1, centrl volume of distribution; z, llometric scling fctor in = CL 7 kg (TBW/7) z. Interindividul vribility CL Model A Interindividul vribility CL 1..5..5 1. Model E..... Age (yers) Age (yers) 6 8 6 8 Figure Interindividul vribility of propofol clernce vs. ge for the simple model without covrites (model A) nd the finl covrite model including ge nd totl body weight on propofol clernce (model E). yers fter which it decresed. As result, insted of estimtion of two different popultion vlues for dolescents vs. dults s in model C, in model D, ge ws implemented using biliner function which significntly reduced the OFV ( OFV s compred with model C = 8. points; P <.5). On the bsis of the covrites of model D, the interindividul vribility of propofol clernce ws reduced by 5%. Figure model E shows tht fter implementtion of ge in biliner function, interindividul vribility ws rndomly distributed with ge. Figure 1, model D shows the post hoc propofol clernce estimtes for model D vs. TBW with popultion predictions for clernce for three different ges (15, 1, nd 65 yers), illustrting the biliner reltion with ge in model D. The finl eqution for propofol clernce ws given s follows (Eq. 1): CL = CL TBW/7 F i pop i Age.77 ( ) ge ge ( ) ge ( ( )) 1 yers : F = 1+.13 ( Age 1) Age > 1 yers : F = 1.539 Age 1 where represents clernce in the ith individul, CL pop is the popultion men vlue for clernce in n individul of 7 kg nd of 1 yers, TBW i is the TBW of the ith individul, nd 7 is the stndrd body weight in kilogrms. Concerning covrites for V1, Tble shows tht there ws only modest influence of the body size descriptors on V1 with trend towrd n increse in V1 with LBW (P >.5). There ws substntil shrinkge (3%) on V1, which not only renders plots using post hoc prmeter estimtes less relible but lso indictes tht the individul dt in the dt sets re not rich in informtion bout this prmeter. 11 Therefore, no covrite on V1 ws incorported in the finl model. By contrst, TBW s covrite for Q3 significntly improved (1) www.nture.com/psp
Phrmcokinetic Met-Anlysis of Propofol the model ( OFV = 18.1; P <.5; Tble ) nd ws therefore considered the finl model (Tble, model E). There ws no influence of the explored covrites on the other phrmcokinetic prmeters (Q nd V). Tble 3 lists ll prmeter estimtes including their coefficient of vrition vlues nd OFVs of the simple model (model A) nd the finl model (model E). The observed vs. popultion-predicted plots strtified by the different cohorts in Figure 3 confirm tht the finl model describes not only the study popultion s whole but lso the individul study popultions without bis. The stbility of the finl model ws shown by the bootstrp nlysis (Tble 3). Figure shows popultion propofol clernce vlues vs. ge for different TBWs using the finl model (model E). This figure shows both the llometric increse of propofol clernce with TBW s the distnce between the weight clsses decreses with incresing TBW, nd the biliner reltionship of propofol clernce with ge. DISCUSSION To describe the influence of obesity nd ge on the phrmcokinetics of propofol, popultion phrmcokinetic met-nlysis ws performed using dt from morbidly obese dults, dolescents, nd children nd their nonobese controls. In this study, wide rnge of TBW (37 18 kg) nd ge (9 79 yers) ws studied, with dt from (morbidly) obese nd nonobese individuls in ech ge rnge. The results of the systemtic nlysis shows tht combintion of TBW nd ge proved to best cpture chnges in propofol clernce s result of obesity nd geing. Although it is yet unknown how these results should be put in physiologicl perspective, the current model seems to provide the best description of the dt from these lrgely divergent ptient popultions. In recent reports in (morbidly) obese dults, it ws shown tht the increse in propofol clernce ws relted to TBW nd could be best described using n llometric function. 5,6 In ddition, n llometric reltionship between TBW nd propofol clernce ws found in dt set of morbidly obese dolescents. 7 Allometric scling fctors of.7 (ref. 6) nd.8 (ref. 7) were estimted for morbidly obese dults nd children nd dolescents, respectively. As these fctors re close to the fctor of.75 predicted by llometry theory, 1 this implies tht obese individuls cn be viewed s lrge individuls ( different body size) Tble 3 Popultion phrmcokinetic prmeter estimtes for the simple nd the finl phrmcokinetic model for propofol in nonobese nd (morbidly) obese children, dolescents, nd dults including the bootstrp vlues of the finl model Prmeter Simple model Finl model Bootstrp finl model Men (CV%) Men (CV%) Men (CV%) Model A E Number of ptients 9 9 5 CL (l/min).37 (3.8) CL 7 kg, 1 yers (l/min).3 (.3).31 (.6) z.77 (6.9).77 (7.) b b.13 (13.5).9 (18.5) c b.539 ( 33.8).85 ( 39.7) V1 (L) 3.33 (11.5) 3.17 (11.3) 3.16 (1.) V (L) 6.55 (1.5) 5.89 (15.) 5.78 (15.3) V3 (L) 118 (9.1) 116 (7.5) 117 (6.7) Q (l/min) 1.7 (9.7) 1.6 (11.7) 1.57 (1.) Q3 (l/min).(7.6) Q3 7 kg (l/min) c 1.5 (6.) 1.6 (5.) Interindividul vribility (%) CL 35.1 (15.5) 17.5 (13.9) 17.7 (1.6) V1 6.6 (.) 5.6 (1.3) 51.1 (38.8) V3. (.) 36. (3.9) 35. (7.8) Q3 8.1 (.3). (37.5) 35.1 (3.) Proportionl intrindividul error (%).3 (1.).3 (1.3).1 (11.1) OFV,331,7,5 = CL 7 kg (TBW/7) z F ge. b Age 1 y: F ge = (1 + b (ge 1)) nd Age > 1 y: F ge = (1 + c (ge 1)). c Q3 i = Q3 7 kg (TBW/7). b, ge fctor for clernce ge 1 yers; c, ge fctor for clernce ge > 1 yers; CL, clernce; CL 7 kg, popultion men vlue for clernce in n individul of 7 kg; CL 7 kg, 1 yers, popultion men vlue for clernce in n individul of 7 kg nd 1 yers; CV, coefficient of vrition of the prmeter vlues; F ge, ge fctor for clernce; OFV, objective function vlue; Q, comprtmentl clernce between V1 nd V; Q3, comprtmentl clernce between V1 nd V3; Q3 7 kg, popultion men vlue for comprtmentl clernce between V1 nd V3 in n individul of 7 kg; V1, centrl volume of distribution; V, peripherl volume 1; V3, peripherl volume ; z, llometric scling fctor in = CL 7 kg (TBW/7) z. CPT: Phrmcometrics & Systems Phrmcology
Phrmcokinetic Met-Anlysis of Propofol 5 Morbidly obese dults Nonobese dults Observed Observed 6 Popultion predicted 6 Popultion predicted Morbidly obese children nd dolescents Nonobese children nd dolescents Observed Observed 6 Popultion predicted 6 Popultion predicted Figure 3 Observed vs. popultion-predicted ln propofol concentrtions of the finl model (model E). Pnels represent dt of morbidly obese dults, nonobese dults, morbidly obese children nd dolescents, nd nonobese children nd dolescents. The solid gry line represents the line of identity, x = y. insted of individuls hving excess body ft ( different body composition). Although these results were confirmed in the current met-phrmcokinetic nlysis, we lso showed tht morbidly obese dolescents cnnot be viewed s dults s their propofol clernce proved lower thn tht of morbidly obese dults with the sme TBW (Figure 1, model A). This difference in propofol clernce could be described with two seprte functions for propofol clernce; i.e., one eqution for children nd dolescents nd one eqution for dults (model C). Alterntively nd significntly better, ge ws incorported s covrite on propofol clernce using biliner function (models D nd E). Therefore, in the finl model, the influence of ge nd obesity on propofol clernce ws described using both TBW nd ge s covrites for propofol clernce. This finl eqution (Eq. 1) is independent of the definitions for ge (e.g., dolescents nd dults) nd obesity (e.g., obese nd morbidly obese) ctegories nd might prove useful for clinicl prctice. In this study, there ws no significnt reltionship between body size mesures nd volumes of distribution. Previously, ge nd TBW hve been identified s covrites for volumes of distribution of propofol in nonobese nd obese ptients. 5,13,1 As result of the finding tht LBW correlted with V1, Ingrnde et l. 15 suggested to use LBW for the induction of nesthesi with propofol. The lck of significnt influence of LBW on volume of distribution in our nlysis my be explined by the fct tht the studies included in the current nlysis minly contined observtions following propofol mintennce infusions. As such, 5 3 1 Age (yers) 18 kg 16 kg 1 kg 1 kg 1 kg 8 kg 6 kg 6 8 Figure Model-bsed predictions of popultion clernce estimtes of propofol vs. ge for ptients with different totl body weights. these dt sets did not contin sufficient observtions just fter the induction bolus dose of propofol to dequtely describe erly (re-)distribution nd the influence of covrites on volumes of distribution. It therefore seems tht dditionl reserch is needed to chrcterize covrites predictive of volume of distribution tht will llow estimtion of propofol-loding doses in morbidly obese dults nd children. This study hd few limittions. We investigted cohort of children nd dolescents with lower ge limit of 9 yers. In clinicl prctice, propofol is dministrted to even www.nture.com/psp
6 Phrmcokinetic Met-Anlysis of Propofol younger children, nd therefore, more reserch is needed to investigte if the current findings re lso pplicble for children younger thn 9 yers. In ddition, the phrmcodynmics of propofol were not included in the current nlysis. Obesity influenced the phrmcodynmics of propofol in children nd dolescents using propofol. 16 Underlying diseses such s dibetes nd chnged (ptho)physiology in (morbidly) obese ptients my lso influence the phrmcodynmics of propofol. 17 To develop dosing lgorithm for propofol in morbidly obese dults, children, nd dolescents, n integrted phrmokinetic nd phrmcodynmic met-nlysis is urgently needed. It remins to be speculted how the influence of TBW on propofol clernce tht ws found in our study cn be explined. Studies hve shown tht obese ptients suffer from low-grde inflmmtion, 18 which is probbly the underlying cuse of the high prevlence of nonlcoholic stetoheptitis. 19 It is known tht nonlcoholic stetoheptitis increses ft deposition in the liver cusing sinusoidl nrrowing nd ltered functionl morphology of the liver. By contrst, becuse of incresed blood volume nd crdic output, heptic blood flow is possibly incresed in obese subjects. 1 As result, incresed propofol clernce my be nticipted s propofol is high extrction rtio drug minly metbolized by vrious UDP-glucuronosyltrnsferse enzymes. 3 Dt on other high-extrction drugs nd drugs metbolized by UDP-glucuronosyltrnsferse suggest tht both UDP-glucuronosyltrnsferse ctivity 6 nd liver blood flow 7,8 re incresed in obese dults. Furthermore, UDP-glucuronosyltrnsferse ctivity is incresed in obese dolescents s compred with nonobese dolescents. 9 Even though this cnnot be proven, it cn be hypothesized tht heptic blood flow is even more incresed due to prolonged durtion of obesity in dults s compred with dolescents with the sme TBW. This is supported by the fct tht ge could be incorported s covrite on propofol clernce. As propofol clernce is limited by the blood flow through the liver, the effect of both TBW nd ge on propofol my be explined by chnges in liver blood flow. In this phrmcokinetic met-nlysis, we developed model for scling propofol clernce over wide rnges of TBW nd ge using dt from morbidly obese dults, dolescents, nd children nd their nonobese controls. The results show tht TBW ws the most predictive covrite for propofol clernce mong ptients when implemented s n llometric function. In ddition, ge ws incorported using biliner function with two distinct slopes, reflecting n initil increse nd subsequent decrese in clernce s result of ge. Using these two functions, the influence of both (over) weight nd ge on propofol clernce ws well chrcterized, which my provide bsis for dosing cross this diverse group of ptients. METHODS Ptients. Dt of five previously published studies were used for this nlysis. 6,7,3 3 Ptient chrcteristics of the five different studies re provided in Tble 1. Detils of the studies re briefly summrized when relevnt to the current nlysis. Morbidly obese dults. Twenty morbidly obese dults scheduled for britric surgery with men TBW of 1 kg (rnge: 98 167 kg) received either propofol induction dose of or 35 mg. Mintennce propofol infusion rte ws initited t 1 mg/kg times TBW/h nd djusted to keep Bispectrl index vlues between nd 6. Remifentnil continuous infusion ws dministrted 5 µg/h/kg bsed on idel body weight. Multiple blood smples were collected before the strt of the propofol bolus until 15 min fter the end of the infusion. 6 Nonobese dults. Forty nonobese dults with men TBW of 7 kg (rnge: 55 98 kg) were included. Twenty-four femle ptients received bolus injection of.5 mg/kg of propofol for induction of nesthesi while nesthesi ws mintined with isoflurne. Following onset of unconsciousness, fentnyl (.3.5 mg/kg) ws dministrted nd s required dditionl mounts of 5 1 mg. Blood smples were collected from 1 min until 18 min fter the induction dose of propofol. 3 Of these ptients, only ptients were in this study becuse height mesure of ptients ws not vilble. Another nonobese intensive cre ptients received continuous propofol infusions for 5 dys with propofol doses bsed on the Rmsy six-point-scle morphine dministrtion s n nlgesic ws given when the ptient ws considered to be in pin (seven ptients; ech ptient received morphine s continuous infusion of 5 mg/dy). Blood smples were collected four times dily during propofol mintennce infusion for 5 dys. 31 Morbidly obese children nd dolescents. In morbidly obese dolescents nd children scheduled for britric surgery with men TBW of 15 kg (rnge: 7 18 kg) nd men ge of 16 yers (rnge: 9 18 yers), propofol ws dministered using dosing weight clculted ccording to the method of Servin et l. Fentnyl 1 µg ws dministered just fter the induction, nd 5 µg doses were dministered in cse of indequte nlgesi. Blood smples were collected before the strt of the propofol infusion nd from 15 min fter the strt of infusion until 1 min fter the end of the infusion. 7 Nonobese children nd dolescents. In 1 nonobese dolescents nd children with men TBW of 5 kg (rnge: 37 8 kg) nd men ge of 1 yers (rnge: 9 yers), nesthesi ws induced with bolus dose of propofol ( mg/kg) nd mintined with propofol by continuous infusion ( 1 mg/kg/h) nd remifentnil (. 1 µg/ kg/min) for scoliosis surgery. Blood smples were tken before induction of nesthesi, nd t ~15 or 3 min fter the strt of propofol infusion until 1 min fter the end of the infusion. 3 Dt nlysis nd internl vlidtion. The nlysis ws performed by mens of nonliner mixed-effects modeling using NONMEM (version VI, relese 1.1; GloboMx LLC, Hnover, MD) 33 with S-plus (version 6.; Insightful Softwre, Settle, CPT: Phrmcometrics & Systems Phrmcology
Phrmcokinetic Met-Anlysis of Propofol 7 WA) to visulize the dt. Discrimintion between different models ws mde by comprison of the OFV ( log likelihood (OVF)). A vlue of P <.5, representing decrese of 3.8 in the OVF, ws considered sttisticlly significnt. In ddition, goodness-of-fit plots (observed vs. individully predicted, observed vs. popultion predicted, conditionl weighted residuls vs. time, nd conditionl weighted residuls vs. popultion predictions) were used for dignostic purposes. Furthermore, the confidence intervl of the prmeter estimtes, the correltion mtrix, nd visul improvement of the individul plots were used to evlute the model. η-shrinkge s defined by Krlsson et l. 3 ws clculted for ll model prmeters for which interindividul vribility ws estimted. The internl vlidity of the popultion phrmcokinetic models ws ssessed by per-study strtified bootstrp resmpling method using 5 replictes. 33 Vlidtion using externl dt sets ws not prt of the current nlysis. Phrmcokinetic model. Log-trnsformed propofol concentrtion dt were described by three-comprtment model (NONMEM VI, ADVAN11, nd TRANS) prmeterized in terms of volume of distribution of the centrl (V1), first (V), nd second peripherl comprtments (V3), intercomprtmentl clernce from the centrl to the first peripherl comprtment (Q) nd from the centrl to the second peripherl comprtment (Q3), nd clernce from the centrl comprtment (CL). The interindividul vlue (post hoc vlue) of the prmeters of the ith subject ws modeled by the following formul: θ = θ e i men where θ men is the popultion men nd η i is rndom vrible with men zero nd vrince ω, ssuming lognorml distribution in the popultion. The intrindividul vribility, resulting from ssy errors, model misspecifictions, nd other unexplined sources, ws best described with proportionl error model. This mens for the jth observed log-trnsformed propofol concentrtion of the ith individul, the reltion (Y ij ): Y where c pred is the predicted propofol concentrtion nd ε ij is the rndom vrible with men zero nd vrince σ. Covrite nlysis. Covrites were plotted independently ginst the individul post hoc prmeter estimtes of ll phrmcokinetic prmeters nd the conditioned weighted residuls to visulize potentil reltions. The following covrites were tested: TBW, body mss index, idel body weight 8 nd LBW, 9,1 gender, nd ge. Covrites were tested using liner nd power equtions: where P i nd P p represent individul nd popultion prmeter estimtes, respectively, Cov represents the ηi = log cpred, + ε ij ij ij Cov Pi = Pp Covstndrd z () (3) () covrite, nd Cov stndrd represents stndrdized (i.e., 7 kg for TBW) or medin vlue of the covrite for the popultion. z represents the scling fctor, which ws fixed to 1 for liner function or n estimted vlue for power eqution. The influence of the covrite ge on clernce ws lso tested using biliner function with two distinct slopes, i.e., liner increse in clernce for ge vlues below the medin ge nd liner decrese in clernce for ge vlues higher thn the medin ge 35 (Eq. 5). CL = CL F ge F i pop ge ( ge medin ge) = 1+ b F ge ( ge > medin ge ) = 1+ c ( ) ge in yers medin ge ( in yers) ( ) ge in yers medin ge ( in yers) Potentil covrites were seprtely entered into the model nd sttisticlly tested by use of the OFV nd if pplicble the 95% confidence intervl of the dditionl prmeter. A P <.5 ws pplied to evlute the covrites in the forwrd inclusion (OFV decrese > 7.9), wheres the bckwrd deletion procedure used stricter criterion (OFV decrese > 1.8; P <.1). When more thn one significnt covrite for the simple model ws found, the covrite-djusted model with the lrgest decrese in objection function ws chosen s bsis to sequentilly explore the influence of dditionl covrites using the sme criteri. Finlly, fter forwrd inclusion, bckwrd exclusion procedure ws pplied to justify the covrite. The choice of the covrite model ws further evluted s discussed under the section Dt nlysis nd internl vlidtion. Acknowledgments. The uthors cknowledge the Teen Longitudinl Assessment of Britric Surgery tem t Cincinnti Children s Hospitl Medicl Center (www.teen- LABS.org) for their support during this study. This study ws funded by the Trnsltionl Reserch Inititive grnt from Cincinnti Children s Reserch Foundtion, Cincinnti Children s Hospitl Medicl Center, Cincinnti, OH, USA. The uthors lso cknowledge finncil support from Ntionl Institutes of Helth grnt 1KHD5387 (A.A.V.). The uthors thnk Simone vn Krlingen for executing the study tht ws used in this met-nlysis nd Uppsl Phrmcometric Summer School 1 for helping to nlyze the dt. Author Contributions. J.D., V.C., S.S., B.v.R., E.P.A.v.D., A.A.V., nd C.A.J.K. wrote the mnuscript. J.D., V.C., S.S., H.J.B.v.O-A., T.I., B.v.R., E.P.A.v.D., A.A.V., nd C.A.J.K. designed the reserch. J.D., V.C., S.S., H.J.B.v.O-A., T.I., nd E.P.A.v.D. performed the reserch. J.D., A.A.V., nd C.A.J.K. nlyzed the dt. Conflict of Interest. The uthors declred no conflict of interest. (5) www.nture.com/psp
8 Phrmcokinetic Met-Anlysis of Propofol Study Highlights WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC? 33 There is very limited knowledge on how key phrmcokinetic prmeters such s clernce chnge in morbidly obese children, dolescents, nd dults s compred with their nonobese counterprts. WHAT QUESTION DID THIS STUDY ADDRESS? 33 This study chrcterizes the influence of both morbid obesity nd ge on the phrmcokinetics of propofol using dt from ptients vrying in TBW from 37 to 18 kg nd in ge from 9 to 79 yers. WHAT THIS STUDY ADDS TO OUR KNOWLEDGE 33 Propofol clernce incresed with TBW in n llometric function, wheres ge ws found to influence clernce in biliner fshion with two distinct slopes, reflecting n initil increse nd subsequent decrese s result of ging. HOW THIS MIGHT CHANGE CLINICAL PHARMACOLOGY AND THERAPEUTICS 33 While the proposed phrmcokinetic model ws the best description of the vilble dt, the results cn be used to further explore physiologicl explntions for the findings. 1. Ogden, C.L., Crroll, M.D., Curtin, L.R., Lmb, M.M. & Flegl, K.M. Prevlence of high body mss index in US children nd dolescents, 7-8. JAMA 33, 9 (1).. Eleveld, D.J., Proost, J.H., Abslom, A.R. & Struys, M.M. Obesity nd llometric scling of phrmcokinetics. Clin. Phrmcokinet. 5, 751 753; discussion 755 (11). 3. Hn, P.Y., Duffull, S.B., Kirkptrick, C.M. & Green, B. Dosing in obesity: simple solution to big problem. Clin. Phrmcol. Ther. 8, 55 58 (7).. Wng, C. et l. A bodyweight-dependent llometric exponent for scling clernce cross the humn life-spn. Phrm. Res. 9, 157 1581 (1). 5. Cortínez, L.I. et l. Influence of obesity on propofol phrmcokinetics: derivtion of phrmcokinetic model. Br. J. Anesth. 15, 8 56 (1). 6. vn Krlingen, S. et l. Popultion phrmcokinetics nd phrmcodynmics of propofol in morbidly obese ptients. Clin. Phrmcokinet. 5, 739 75 (11). 7. Diepstrten, J. et l. Propofol clernce in morbidly obese children nd dolescents: influence of ge nd body size. Clin. Phrmcokinet. 51, 53 551 (1). 8. Pi, M.P. & Ploucek, F.P. The origin of the idel body weight equtions. Ann. Phrmcother. 3, 166 169 (). 9. Jnmhstin, S., Duffull, S.B., Ash, S., Wrd, L.C., Byrne, N.M. & Green, B. Quntifiction of len bodyweight. Clin. Phrmcokinet., 151 165 (5). 1. Peters, A.M., Snelling, H.L., Glss, D.M. & Bird, N.J. Estimtion of len body mss in children. Br. J. Anesth. 16, 719 73 (11). 11. Svic, R.M. & Krlsson, M.O. Importnce of shrinkge in empiricl byes estimtes for dignostics: problems nd solutions. AAPS J. 11, 558 569 (9). 1. Anderson, B.J. & Holford, N.H. Mechnism-bsed concepts of size nd mturity in phrmcokinetics. Annu. Rev. Phrmcol. Toxicol. 8, 33 33 (8). 13. Schüttler, J. & Ihmsen, H. Popultion phrmcokinetics of propofol: multicenter study. Anesthesiology 9, 77 738 (). 1. Servin, F., Frinotti, R., Hberer, J.P. & Desmonts, J.M. Propofol infusion for mintennce of nesthesi in morbidly obese ptients receiving nitrous oxide. A clinicl nd phrmcokinetic study. Anesthesiology 78, 657 665 (1993). 15. Ingrnde, J., Brodsky, J.B. & Lemmens, H.J. Len body weight sclr for the nesthetic induction dose of propofol in morbidly obese subjects. Anesth. Anlg. 113, 57 6 (11). 16. Olutoye, O.A. et l. The effect of obesity on the ED(95) of propofol for loss of consciousness in children nd dolescents. Anesth. Anlg. 115, 17 153 (1). 17. Adms, J.P. & Murphy, P.G. Obesity in nesthesi nd intensive cre. Br. J. Anesth. 85, 91 18 (). 18. Wellen, K.E. & Hotmisligil, G.S. Inflmmtion, stress, nd dibetes. J. Clin. Invest. 115, 1111 1119 (5). 19. Guzzloni, G., Grugni, G., Minocci, A., Moro, D. & Morbito, F. Liver stetosis in juvenile obesity: correltions with lipid profile, heptic biochemicl prmeters nd glycemic nd insulinemic responses to n orl glucose tolernce test. Int. J. Obes. Relt. Metb. Disord., 77 776 ().. Frrell, G.C., Teoh, N.C. & McCuskey, R.S. Heptic microcircultion in ftty liver disese. Ant. Rec. (Hoboken). 91, 68 69 (8). 1. Csti, A. & Putzu, M. Anesthesi in the obese ptient: phrmcokinetic considertions. J. Clin. Anesth. 17, 13 15 (5).. Al-Jhdri, W.S., Ymmoto, K., Hirok, H., Nkmur, K., Goto, F. & Horiuchi, R. Prediction of totl propofol clernce bsed on enzyme ctivities in microsomes from humn kidney nd liver. Eur. J. Clin. Phrmcol. 6, 57 533 (6). 3. King, T.K., Ensom, M.H. & Chng, T.K. UDP-glucuronosyltrnsferses nd clinicl drugdrug interctions. Phrmcol. Ther. 16, 97 13 (5).. Abernethy, D.R., Divoll, M., Greenbltt, D.J. & Ameer, B. Obesity, sex, nd cetminophen disposition. Clin. Phrmcol. Ther. 31, 783 79 (198). 5. Vn Wrt, S. et l. Popultion phrmcokinetics nd phrmcodynmics of grenoxcin in ptients with community-cquired respirtory trct infections. Antimicrob. Agents Chemother. 8, 766 777 (). 6. Abernethy, D.R., Greenbltt, D.J., Divoll, M. & Shder, R.I. Enhnced glucuronide conjugtion of drugs in obesity: studies of lorzepm, oxzepm, nd cetminophen. J. Lb. Clin. Med. 11, 873 88 (1983). 7. Sprreboom, A. et l. Evlution of lternte size descriptors for dose clcultion of nticncer drugs in the obese. J. Clin. Oncol. 5, 77 713 (7). 8. Schwrtz, A.E., Mtteo, R.S., Ornstein, E., Young, W.L. & Myers, K.J. Phrmcokinetics of sufentnil in obese ptients. Anesth. Anlg. 73, 79 793 (1991). 9. Brshop, N.J., Cpprelli, E.V., Sirlin, C.B., Schwimmer, J.B. & Lvine, J.E. Acetminophen phrmcokinetics in children with nonlcoholic ftty liver disese. J. Peditr. Gstroenterol. Nutr. 5, 198 (11). 3. Knibbe, C.A. et l. Phrmcokinetics, induction of nesthesi nd sfety chrcteristics of propofol 6% SAZN vs propofol 1% SAZN nd Diprivn-1 fter bolus injection. Br. J. Clin. Phrmcol. 7, 653 66 (1999). 31. Knibbe, C.A., Zuideveld, K.P., DeJongh, J., Kuks, P.F., Arts, L.P. & Dnhof, M. Popultion phrmcokinetic nd phrmcodynmic modeling of propofol for long-term sedtion in criticlly ill ptients: comprison between propofol 6% nd propofol 1%. Clin. Phrmcol. Ther. 7, 67 68 (). 3. Peeters, M.Y. et l. Prediction of propofol clernce in children from n llometric model developed in rts, children nd dults versus.75 fixed-exponent llometric model. Clin. Phrmcokinet. 9, 69 75 (1). 33. Bel, S.L., Sheiner, L.B., Boeckmnn, A. NONMEM User s Guide. (University of Cliforni, Sn Frncisco, 1999). 3. Krlsson, M.O. & Svic, R.M. Dignosing model dignostics. Clin. Phrmcol. Ther. 8, 17 (7). 35. Jonsson, E.N. & Krlsson, M.O. Automted covrite model building within NONMEM. Phrm. Res. 15, 163 168 (1998). CPT: Phrmcometrics & Systems Phrmcology is n open-ccess journl published by Nture Publishing Group. This work is licensed under Cretive Commons Attribution- NonCommercil-NoDerivtive Works 3. License. To view copy of this license, visit http://cretivecommons.org/licenses/by-nc-nd/3./ CPT: Phrmcometrics & Systems Phrmcology