7 replies to this topic

[Jim McDowell]

Dear All,

I'm asked to compare the results from an NLME 2-cmpt analyis of steady-state data with a 19-43 day half-life. I used the reparamiterization from my 1994 NONMEM User's Guide, appendix 2, which indicated I should use K10=Cl/V from the cerntral compartment. I followed this with the standard ln(2)/K10=t1/2. My results differ from 21 days for the NLME to 30 days for the NCA. Since the CL values are dead on, that suggests that the V term is inadequately being estimated by my reparmeterization.

 

What is the proper reparmeterization to estimate the secondary paramter K10?

 

4-Week Schedule: NLME Results

    V Cl K10 t12 t12-Days

    L L/hr 1/hr hr day

Mean: 3.43 0.00483 0.001409 512.045 21.335

SD: 0.37 0.00111 0.000283 109 5

CV(%): 10.7% 23.0% 20.1% 21.4% 21.4%

Median: 3.49 0.00487 0.001377 504 21

Minimum: 2.77 0.00281 0.000812 335 14

Maximum: 3.90 0.00737 0.002068 854 36

 

4-Week Schedule: NCA Results

    Vz CLss Lambda_z HLLambda_z HLLambda_z

    L L/hr 1/hr hr day

Mean: 4.95 0.00484 0.000999 720.023 30.001

SD: 1.29 0.00110 0.000201 140 6

CV(%): 26.0% 22.8% 20.1% 19.5% 19.5%

Median: 4.84 0.00476 0.000986 703 29

Minimum: 3.30 0.00290 0.000674 445 19

Maximum: 8.74 0.00738 0.001556 1028 43

[Jim McDowell]

Sorry, my table lost all of it's structure.

[file name=Phoenix_NLME_lambda_z_post.xlsx size=10571]http://pharsight.com/extranet/media/kunena/attachments/legacy/files/Phoenix_NLME_lambda_z_post.xlsx[/file]

Attached Thumbnails

• 

Attached Files

•  Phoenix_NLME_lambda_z_post.xlsx   10.32KB   303 downloads

[Simon Davis]

Jim, you're using a 'built-in' 2 compartment model, right? So you can use the useful trick of Set WNL model to set up the secondary parameters for you.

 

I "think" this illustrates the wrong assumption in your calculation, to have a slope and therefore half-life analogus to the NCA model you should be instead compare HL_lambda_z with the ...

 

    secondary(Beta_hl = log(2)/Beta)

 

 But I haven't had a chance to check this with an example set - does the reasoning sound reasonable to you since the NCA thalf comes only from the last slope?

 

  Simon

[Jim McDowell]

Dear Simon,

 

Yes, it seems most reasonable. I did not know that beta was a value I could access in defining secondary parameters. Is there a list of these some where so I can scout them out for myself?

 

Thanks, and kind regards, Jim

[Simon Davis]

Secondary parameters can be defined by users in the model, the Set WNL model will write a whole bunch of them for 'library' models

you can see them for this model here;

I post the fill code for this example 2-com Extravascular Dosing

Simon
test(){
cfMicro(A1, Cl / V, Cl2 / V, Cl2 / V2, first = (Aa = Ka))
dosepoint(Aa, idosevar = AaDose)
C = A1 / V
error(CEps = 1)
observe(CObs = C + CEps)
stparm(Ka = tvKa * exp(nKa))
stparm(V = tvV * exp(nV))
stparm(V2 = tvV2 * exp(nV2))
stparm(Cl = tvCl * exp(nCl))
stparm(Cl2 = tvCl2 * exp(nCl2))
fixef(tvKa = c(, 1, ))
fixef(tvV = c(, 1, ))
fixef(tvV2 = c(, 1, ))
fixef(tvCl = c(, 1, ))
fixef(tvCl2 = c(, 1, ))
secondary(Ke = tvCl/tvV)
secondary(K12 = tvCl2/tvV)
secondary(K21 = tvCl2/tvV2)
secondary(Ktot = K12+K21+Ke)
secondary(r1 = (Ktot^2-4*K21*Ke)^0.5)
secondary(Alpha = (Ktot+r1)/2)
secondary(Beta = (Ktot-r1)/2)
secondary(Gamma = tvKa)
secondary(a1 = (K21-Alpha)/(Beta-Alpha))
secondary(b1 = (K21-Beta)/(Alpha-Beta))
secondary(KaAlpha = tvKa/(tvKa-Alpha))
secondary(KaBeta = tvKa/(tvKa-Beta))
secondary(aa = KaAlpha*a1)
secondary(bb = KaBeta*b1)
secondary(cc = -(aa+bb))
secondary(DoseOverVol = AaDose/tvV)
secondary(A = DoseOverVol*aa)
secondary(B = DoseOverVol*bb)
secondary(C = DoseOverVol*cc)
secondary(AUC = DoseOverVol/Ke)
secondary(Alpha_hl = log(2)/Alpha)
secondary(Beta_hl = log(2)/Beta)
secondary(Tmax = CalcTMax(A,Alpha, B, Beta, C, Gamma))
secondary(Cmax = A*exp(-Alpha*Tmax)+B*exp(-Beta*Tmax)+C*exp(-Gamma*Tmax))
secondary(Ke_hl = log(2)/Ke)
secondary(Ka_hl = log(2)/tvKa)
ranef(diag(nV, nCl, nKa, nV2, nCl2) = c(1, 1, 1, 1, 1))
}

Attached Thumbnails

• 
• 

[Jim McDowell]

Whoa! I did not realize that option was available. Set WNL is a check box I never examined!

 

Major help with something I've been struggling with for the past couple of years. I figured it had to be documented somewhere but I could not find it anywhere.

 

Let me buy you a pint!

[Simon Davis]

yes it saves me having to pull out reference books and re-arrange formulas all the time. Any good Pale Ales in Texas? or are you in Raleigh now?

[Jim McDowell]

Yes, I now hale from the Republic of Texas and located in Austin. I wouldn't know about the pale ales but my son seems to have found some he likes. I'm more into the Bock lagers.

 

One last question on the secondary parameters. I'm getting a full complement of statistics on all of the parameters excet for Tinf. Here I am missing the stdeff and CV% but getting all of the others. Any ideas why this might me happening.

 

Individual figures confirm the program is recognizing the infusion length corrects as I have a sample at the start and end of the infusion.

 

Kind regards, Jim